├── README.md ├── chatgpt_module ├── annualreport.pdf ├── app.py └── requirements.txt └── custom_module ├── .gitignore ├── PDFSummary.txt ├── WebSummary.txt ├── api.py ├── base.py ├── crawl_news.ipynb ├── crawl_papers_with_code.ipynb ├── pdf ├── 1910.10536v3.pdf └── 1911.03607v1.pdf ├── requirements.txt ├── templates ├── base.html ├── create.html └── index.html └── utils.py /README.md: -------------------------------------------------------------------------------- 1 | # Text Summarization 2 | 3 | ## About The Project 4 | 5 | This is an implementation of Text Summarization. 6 | 7 | ## Installation 8 | 9 | 1. Clone the repo 10 | 11 | ```sh 12 | git clone https://github.com/phkhanhtrinh23/text_summarization.git 13 | ``` 14 | 15 | 2. Use any code editor to open the folder **text_summarization**. 16 | 17 | ## Run 18 | 1. Create conda virtual environment: `conda create -n summarizer python=3.9`, activate it: `conda activate summzarizer` 19 | - In order to run `custom_module`, you should install `tesseract` by command: `sudo apt-get install tesseract-ocr`. 20 | 21 | 2. You have the following choices: 22 | - `chatgpt_module`: 23 | * Install the required packages: `pip install -r requirements.txt`. 24 | * Start the app: `streamlit run app.py`. 25 | - `custom_module`: 26 | * Install the required packages: `pip install -r requirements.txt`. 27 | * Run the file: `python main.py`: 28 | - `summarize_custom()` is heuristic summarization. 29 | - `summarize_model()` is summarization using LLMs. 30 | 31 | ## Contribution 32 | 33 | Contributions are what make GitHub such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated. 34 | 35 | 1. Fork the project 36 | 2. Create your Contribute branch: `git checkout -b contribute/Contribute` 37 | 3. Commit your changes: `git commit -m 'add your messages'` 38 | 4. Push to the branch: `git push origin contribute/Contribute` 39 | 5. Open a pull request 40 | 41 | ## Contact 42 | 43 | Email: phkhanhtrinh23@gmail.com -------------------------------------------------------------------------------- /chatgpt_module/annualreport.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/phkhanhtrinh23/text_summarization/4efd25dfedd13ecb6b88dd3b1a1668a0fe7d96e2/chatgpt_module/annualreport.pdf -------------------------------------------------------------------------------- /chatgpt_module/app.py: -------------------------------------------------------------------------------- 1 | # Import os to set API key 2 | import os 3 | # Import OpenAI as main LLM service 4 | from langchain.llms import OpenAI 5 | from langchain.embeddings import OpenAIEmbeddings 6 | # Bring in streamlit for UI/app interface 7 | import streamlit as st 8 | 9 | # Import PDF document loaders...there's other ones as well! 10 | from langchain.document_loaders import PyPDFLoader 11 | # Import chroma as the vector store 12 | from langchain.vectorstores import Chroma 13 | 14 | # Import vector store stuff 15 | from langchain.agents.agent_toolkits import ( 16 | create_vectorstore_agent, 17 | VectorStoreToolkit, 18 | VectorStoreInfo 19 | ) 20 | 21 | # Set APIkey for OpenAI Service 22 | # Can sub this out for other LLM providers 23 | os.environ['OPENAI_API_KEY'] = "The OpenAI API Key" 24 | 25 | # Create instance of OpenAI LLM 26 | llm = OpenAI(temperature=0.1, verbose=True) 27 | embeddings = OpenAIEmbeddings() 28 | 29 | # Create and load PDF Loader 30 | loader = PyPDFLoader('annualreport.pdf') 31 | # Split pages from pdf 32 | pages = loader.load_and_split() 33 | # Load documents into vector database aka ChromaDB 34 | store = Chroma.from_documents(pages, embeddings, collection_name='annualreport') 35 | 36 | # Create vectorstore info object - metadata repo? 37 | vectorstore_info = VectorStoreInfo( 38 | name="annual_report", 39 | description="a banking annual report as a pdf", 40 | vectorstore=store 41 | ) 42 | # Convert the document store into a langchain toolkit 43 | toolkit = VectorStoreToolkit(vectorstore_info=vectorstore_info) 44 | 45 | # Add the toolkit to an end-to-end LC 46 | agent_executor = create_vectorstore_agent( 47 | llm=llm, 48 | toolkit=toolkit, 49 | verbose=True 50 | ) 51 | st.title('🦜🔗 ChatGPT + LangChain') 52 | # Create a text input box for the user 53 | prompt = st.text_input('Input your prompt here') 54 | 55 | # If the user hits enter 56 | if prompt: 57 | # Then pass the prompt to the LLM 58 | response = agent_executor.run(prompt) 59 | # ...and write it out to the screen 60 | st.write(response) 61 | 62 | # With a streamlit expander 63 | with st.expander('Document Similarity Search'): 64 | # Find the relevant pages 65 | search = store.similarity_search_with_score(prompt) 66 | # Write out the first 67 | st.write(search[0][0].page_content) 68 | -------------------------------------------------------------------------------- /chatgpt_module/requirements.txt: -------------------------------------------------------------------------------- 1 | aiohttp==3.8.4 2 | aiosignal==1.3.1 3 | altair==4.2.2 4 | anyio==3.6.2 5 | asttokens==2.2.1 6 | async-timeout==4.0.2 7 | attrs==23.1.0 8 | backcall==0.2.0 9 | backoff==2.2.1 10 | blinker==1.6.2 11 | cachetools==5.3.0 12 | certifi==2022.12.7 13 | charset-normalizer==3.1.0 14 | chromadb==0.3.21 15 | click==8.1.3 16 | clickhouse-connect==0.5.23 17 | colorama==0.4.6 18 | comm==0.1.3 19 | dataclasses-json==0.5.7 20 | debugpy==1.6.7 21 | decorator==5.1.1 22 | duckdb==0.7.1 23 | entrypoints==0.4 24 | executing==1.2.0 25 | fastapi==0.95.1 26 | filelock==3.12.0 27 | frozenlist==1.3.3 28 | fsspec==2023.4.0 29 | gitdb==4.0.10 30 | GitPython==3.1.31 31 | greenlet==2.0.2 32 | h11==0.14.0 33 | hnswlib==0.7.0 34 | httptools==0.5.0 35 | huggingface-hub==0.14.1 36 | idna==3.4 37 | importlib-metadata==6.6.0 38 | ipykernel==6.22.0 39 | ipython==8.13.2 40 | jedi==0.18.2 41 | Jinja2==3.1.2 42 | joblib==1.2.0 43 | jsonschema==4.17.3 44 | jupyter_client==8.2.0 45 | jupyter_core==5.3.0 46 | langchain==0.0.150 47 | lz4==4.3.2 48 | markdown-it-py==2.2.0 49 | MarkupSafe==2.1.2 50 | marshmallow==3.19.0 51 | marshmallow-enum==1.5.1 52 | matplotlib-inline==0.1.6 53 | mdurl==0.1.2 54 | monotonic==1.6 55 | mpmath==1.3.0 56 | multidict==6.0.4 57 | mypy-extensions==1.0.0 58 | nest-asyncio==1.5.6 59 | networkx==3.1 60 | nltk==3.8.1 61 | numexpr==2.8.4 62 | numpy==1.24.3 63 | openai==0.27.6 64 | openapi-schema-pydantic==1.2.4 65 | packaging==23.1 66 | pandas==2.0.1 67 | parso==0.8.3 68 | pickleshare==0.7.5 69 | Pillow==9.5.0 70 | platformdirs==3.5.0 71 | posthog==3.0.1 72 | prompt-toolkit==3.0.38 73 | protobuf==3.20.3 74 | psutil==5.9.5 75 | pure-eval==0.2.2 76 | pyarrow==12.0.0 77 | pycryptodome==3.17 78 | pydantic==1.10.7 79 | pydeck==0.8.1b0 80 | Pygments==2.15.1 81 | Pympler==1.0.1 82 | pypdf==3.8.1 83 | pyrsistent==0.19.3 84 | python-dateutil==2.8.2 85 | python-dotenv==1.0.0 86 | pytz==2023.3 87 | pytz-deprecation-shim==0.1.0.post0 88 | PyYAML==6.0 89 | pyzmq==25.0.2 90 | regex==2023.5.5 91 | requests==2.30.0 92 | rich==13.3.5 93 | scikit-learn==1.2.2 94 | scipy==1.10.1 95 | sentence-transformers==2.2.2 96 | sentencepiece==0.1.99 97 | six==1.16.0 98 | smmap==5.0.0 99 | sniffio==1.3.0 100 | SQLAlchemy==2.0.12 101 | stack-data==0.6.2 102 | starlette==0.26.1 103 | streamlit==1.22.0 104 | sympy==1.11.1 105 | tenacity==8.2.2 106 | threadpoolctl==3.1.0 107 | tiktoken==0.3.3 108 | tokenizers==0.13.3 109 | toml==0.10.2 110 | toolz==0.12.0 111 | torch==2.0.0 112 | torchvision==0.15.1 113 | tornado==6.3.1 114 | tqdm==4.65.0 115 | traitlets==5.9.0 116 | transformers==4.28.1 117 | typing-inspect==0.8.0 118 | typing_extensions==4.5.0 119 | tzdata==2023.3 120 | tzlocal==4.3 121 | urllib3==2.0.2 122 | uvicorn==0.22.0 123 | validators==0.20.0 124 | watchdog==3.0.0 125 | watchfiles==0.19.0 126 | wcwidth==0.2.6 127 | websockets==11.0.2 128 | yarl==1.9.2 129 | zipp==3.15.0 130 | zstandard==0.21.0 131 | -------------------------------------------------------------------------------- /custom_module/.gitignore: -------------------------------------------------------------------------------- 1 | __pycache__/ -------------------------------------------------------------------------------- /custom_module/PDFSummary.txt: -------------------------------------------------------------------------------- 1 | DeepMask: an algorithm for cloud and cloud shadow detection in optical satellite remote sensing images using deep residual network 2 | deepmask algorithm cloud cloud shadow detection inoptical satellite remote sensing images using deep residual network. cloud mask generation algorithm istrained evaluated landsat cloud cover assessment validation dataset distributedacross different land types compared cfmask widely used cloud detectionalgorithm landtypespecific deepmask models achieve higher accuracy across land typest. We share motivation using convolution neural networks cnn cloud detection assome previous works pick resnet one widelyused cnn architecturesas backbone develop new cloud mask algorithm called deepmask different theprevious cnnbased approaches model singlestage require preprocessingor postprocessing steps compared semantic segmentation cnnrnn networksour algorithm aims network parsimoniousness efficiency maintaining levelof precision. we present quantitative qualitative results section provide detailed discussion ofthe results future directions section conclusion given section materials methods landsat cloud cover assessment validation dataset. Convolutional neural networks cnns led advances many computer vision applications such as image classification object detection semantic segmentation. we pick resnet deep cnn backbone resnet first introduced ilsvrc competition become one widelyused deep cnnns recent years. Deepmask algorithm is a pipeline consisting local region extractor modulea resnet backbone module b zoomedin view module c typical residual blockis also giventhis step reduces computational cost training resnet classifier compared using allthe local regions image training set testing since would like predictthe class label. Land type experiments explainedin section account satellites like planetscope planet lab also perform anablation experiment basic spectral bands red green blue nir similar alllandtype experiments dropsingleband keepbands experiments repeatthe experiment five times take average performance evaluation metrics following tradition cloud mask machine learning. Landtypespecificmodels visually assess deepmasks performance shown figure figure eachland cover type select one typical scene image location cloud ratio relevantinformation test scenes provided table display rgb scene image groundtruth label mask cfmask cloud mask deepmaskcloud mask rgb image generated combining red green blue bands imagegroundtruth labelmask. cf mask qualitative results in addition numerical results provide visualizations. This paper presents deepmask algorithm developed based resnet cloud cloudshadow mask generation landsat imagery potentially widely applicable different sortsof optical satellite imagery compared thresholdbased methods. we acknowledge support doe cabbi guan also acknowledge support nasaterrestrial ecology program nasa carbon monitoring system program. Remote sensing ofenvironment k x zhang ren j sun deep residual learning image recognition inproceedings ieee conference computer vision pattern recognition pages e h helmer lefsky roberts biomass accumulation rates amazoniansecondary forest biomass oldgrowth forests landsat time series geosciencelaser altimeter system journal applied remote sensing. hill improvement fmask algorithm forsentinel images separating clouds bright surfaces based parallax effects remotesensing environment. conference proceedings march portland proceedings rmrspfort collins co us department agriculture forest service rocky mountain researchstation p volume ronneberger p fischer 't ' brox unet convolutional networks biomedical imagesegmentation international conference medical image computing computerassisteduntervention pages springer w b rossow r schiffer advances understanding clouds. 3 | 4 | Self-attention for raw optical Satellite Time Series Classification 5 | The amount available earth observation data increased dramatically recent years efficiently making use entire body information current challenge remote sensing demands lightweight problemagnostic models requiren region problemspecific expert knowledge endtoend trained deep learning models make use raw sensory data byco learning. Landsat ccdc modis bfastin contrast artificial neural networks aim learning feature extraction classification single dynamicmodel solely provided data minimal supervision. the stateoftheart remote sensing remains focused convolutional recurrent architectures selfattention combination pretraining started dominating thestateoftheart. Time series classification aims learningthe mapping foex input time series x x rix individual measurementsz r features one c classes classes are represented onehot target vector wherey indicates boolean class membership denote approximation target vector aneural network nonlinear differentiable function fo fo flofs ' et h consists linear transformationo ts l cascaded layers fg layer mapping input representation h ' rxd corresponding hidden features h '' r encode increasingly higherlevel features time series classification one ofthe following mechanisms used implement nonlinearmapping fully connected layers. Natural language processing models experimentally analyzed later sectionsand softattentionsoftattention implemented duplo model special case eq values v x queriesq tanx obtained linear transformationfrom input tensor x contrast selfattention keysk fixed weight matrix learned gradientdescent. following definitions modified fromeq softattention yieldsh froxx softmax tan x ox x alwhere nonlinearity implemented tangents tan withoutscaling factor compare selfatt attention simplifiedsoftattention dedicated experiment section modelsthe previous section provided overview neural networklayers used temporal feature extraction introduced temporal convolution recurrence selfatt attention. Original implementation solely tune number convolutionalkernels architecture tempcnn. three sequential convolutional layers followedby batch normalization relu activation function dropout. feature extraction usually required goodperformance hence added features normalized difference vegetation index. We focus on three spatially separate regions bavaria shownin fig regions. we focus on implicit bias focusing on the common agricultural policy. we use theraytune framework to reduce tuning time class imbalance data. we focused on the temporal signal key sourceof relevant features. Study providedby stmelf follow longtailed class distribution with more distinct categories common categories cover fieldparcels respectively. satellite datawe utilized data optical sentinel satellite constellation consists two satellites orbiting earth sunsynchronous orbit opposite tracks satellites observethe spot earth 's surface every two five days depending latitude data acquired linescannercapturing spectral bands ranging ultraviolet wavelengths. We compared performance deep learning models lstmrnn transformer msresnet duplomodel tempcnn well rf classifieras shallow baseline determined optimal hyperparameters separately preprocessed raw sentinel time seriesdatasets class class categorizations asdescribed section experiment trained andevaluated three different models best secondbest and thirdbest hyperparameter configuration random seeds forparameter initialization composition training batches in tablei. Rf baseline achieved competitive results comtable comparison models preprocessed pre raw datasets class land use class land covercategorizations values reported mean standard deviation three models best secondbest thirdbesthyperparameter sets trained training validation partitions tested evaluation partitiona. Deep learning frameworks utilize automatic differentiationand require ground truth labels for experiment estimated influence inputtime step classification prediction evaluated networks ie lstmrnn recurrence transformerselfattention well msresnet tempcnn convolution figurejillustrates means two separate examples corn parcel summer barley parcel top figuresin show input time series x sequence raw sentinel reflectances year. The transformer models realize multiheaded selfattention asmultiple attention mechanisms parallel recall every selfatt attention mechanism calculates attention matrix using softmax operation following eq attentionscores define influence input time feature higherlevel output time feature visualize values matrix figs c three attention heads firstselfattention layer three matrix seen anadjacency matrix input nodes output nodes infigs b alternatively show matrix bipartite graphs former matrix elements shown asweighted directed edges input output nodes. Deep learning models general extract features increasingcomplexity throughout cascaded layer architectures inthis experiment analyzed property visualizing thehidden features varying deeper layers transformer architecture shown fig analysis extractd dimensional feature vectors encoding time series samples test dataset project twodimensional embedding space. Transformer lstmrnn architectures achievedbetter accuracies compared convolutional models weinvestigated feature importance analysis sectionusing gradients observed selfattentionand recurrence mechanisms helped suppress information notrelevant classification time series. usa remote sensing j pefiabarragan k ngugi r e plant j six objectbased cropidentification using multiple vegetation indices textural features cropphenology remote sensing environment l zhong l hu h zhou deep learning based multitemporal crop classification. Computer vision k simonyan zisserman deep convolutional networks largescale image recognition. database ieee conference oncomputer vision pattern recognition ieee pp russakovsky j deng h su j krause satheesh z huanga karpathy khosla bernstein et al. International geoscience andremote sensing symposium igarss pp tgarssa sharma x liu x yang land cover classification multitemporal multispectral remotely sensed imagery using patchbased recurrent neural networks neural networks. satellite image time series remote sensing v fe garnot l landrieu giordano n chehata satellite imagetime series classification pixelset encoders temporal selfattention. Research nov l mcinnes j healy j melville umap uniform manifold approximation projection dimension reduction arxiv preprintarxiv vf mohammadimanesh b salehi mahdianpari e gill moliniera new fully convolutional neural network semantic segmentation. 6 | 7 | -------------------------------------------------------------------------------- /custom_module/WebSummary.txt: -------------------------------------------------------------------------------- 1 | Biden nominates Space Force generals for promotion 2 | WASHINGTON — President Biden on July 11 submitted to the U.S. Senate the nominations of two threestar Space Force leaders for promotion to fourstar generals. Lt. Gen. Stephen Whiting, commander of the Space Force's Space Operations Command; and Lt. Gen. Michael Guetlein, commander of the Space Force's Space Systems Command, have been nominated to the rank of general and to receive a fourth star. Whiting is expected to become the next commander of U.S. Space Command, replacing Gen. James Dickinson. The White House also submitted the nomination of Lt. Gen. Philip Garrant, deputy chief of space operations for strategy, plans, programs and requirements, to a new assignment at the same rank. Garrant is expected to replace Guetlein as the commander of the Space Systems Command, based in Los Angeles. Block on military nominations If the Space Force nominations are approved by the Senate Armed Services Committee, it's unclear when they will reach the Senate floor. 3 | 4 | Second Israeli lunar lander faces funding uncertainty 5 | WASHINGTON — As India prepares to launch its second lunar lander mission, the fate of a second Israeli lander is in doubt after the organization developing it lost a major source of funding. India's Chandrayaan3 spacecraft is scheduled to launch July 14 on a Geosynchronous Satellite Launch Vehicle Mark 3, also known as LVM3, from the Satish Dhawan Space Centre. The spacecraft will gradually go from a geostationary transfer orbit to a low lunar orbit, from which Chandrayaan3 will descend to the lunar surface. Chandrayaan3 is similar to India's first lunar lander flown as part of the Chandrayaan2 mission, which crashed attempting a soft landing in September 2019. Chandrayaan3 incorporates several revisions, such as additional fuel, based on the investigation into the failed landing. That crash took place five months after Beresheet, a spacecraft originally developed by Israeli nonprofit organization SpaceIL to compete for the Google Lunar X Prize, crashed attempting its own lunar landing. SpaceIL and Israel Aerospace Industries, which built the spacecraft, later said the lander crashed because one of its inertial measurement units malfunctioned. SpaceIL announced its intent to pursue a second mission, called Beresheet 2. The mission had been slated to launch in 2025, a date confirmed in a January 2023 announcement of a joint statement of intent between the Israel Space Agency and NASA to cooperate on Beresheet 2. Like the first mission, SpaceIL projected using philanthropic donations to fund Beresheet 2. However, in May a group of donors announced they were halting future payments to the project after spending 45 million, nearly half its estimated 100 million cost. In a statement representing the donors, Morris Kahn, a billionaire who also supported the original Beresheet mission, said the decision to halt future payments was not related to any problems with Beresheet 2. SpaceIL said at the time it would seek alternative funding to continue the mission, but has provided no updates on those efforts since then. In a June 27 presentation at the European Lunar Symposium, Dan Blumberg, chairman of the Israel Space Agency, said work on Beresheet 2 was continuing for now. 6 | 7 | China’s Landspace reaches orbit with methane-powered Zhuque-2 rocket 8 | HELSINKI — Chinese private rocket firm Landspace achieved a global first late Tuesday by reaching orbit with a methanefueled rocket. Landspace and Chinese state media announced that the second Zhuque2 reached orbit, making it the first methanefueled globally to reach orbit. The Zhuque2 mission carried no payload and the rocket's first stage was not recovered. The successful launch also makes Landspace the second private Chinese launch firm to reach orbit with a liquid propellant rocket. https:twitter.comCNSAWatcherstatus1679006943698976768 Taken together, the achievements indicate a breakthrough and growing level of maturity in Chinese commercial space launch efforts. Landspace has already begun assembling its third Zhuque2 (Vermillion Bird2), indicating that another launch could come before the end of the year. Space Pioneer says it has multiple orders for launches for the Tianlong2, and aims to launch the Falcon 9class Tianlong3 in the first half of 2024. Future Zhuque2 launches with upgraded second stage engines will be capable of delivering a 6,000kilogram payload capacity to a 200kilometer low Earth orbit (LEO), or 4,000 kilograms to 500kilometer sunsynchronous orbit (SSO), according to Landspace. View from above:https:t.co2pY1F2mgFW pic.twitter.comGDAj7uWqS7— Cosmic Penguin (CosmicPenguin) July 12, 2023 Landspace is one of China's first commercial rocket firms. It is also one of the bestfunded Chinese launch firms, but the company's journey to orbit has not been smooth. Landspace has set up an intelligent manufacturing base in Huzhou, Zhejiang Province and established a 1.5 billion medium and largescale liquid rocket assembly and test plant at Jiaxing, also in Zhejiang. Space Pioneer and other later movers such as OrienSpace are developing larger rockets which are targeting contracts to launch batches of satellites for China's national satellite internet megaconstellation project, named Guowang. The Zhuque2 launch was China's 27th orbital mission of 2023, with a total of more than 70 launches planned from stateowned main space contractor CASC and commercial players. 9 | 10 | -------------------------------------------------------------------------------- /custom_module/api.py: -------------------------------------------------------------------------------- 1 | from flask import Flask, render_template, request, url_for, redirect 2 | from base import TextSummarization 3 | from threading import Thread 4 | import pandas as pd 5 | from utils import crawl_web, crawl_pdf, extract_OCR 6 | import os 7 | 8 | os.environ["TOKENIZERS_PARALLELISM"] = "1" 9 | 10 | app = Flask(__name__) 11 | obj = TextSummarization() 12 | 13 | # Example 14 | messages = [ 15 | { 16 | 'text_web': 'https://urlfromsciencewebsite.com', 17 | 'text_pdf': 'https://urlfrompdfwebsite.com', 18 | 'answer_web': [['This is a title from a website.', 'This is the summarization of the answer.']], 19 | 'answer_pdf': [['This is a title from multiple pdf files.', 'This is the summarization of multiple pdf files.']], 20 | }, 21 | ] 22 | 23 | def process_web(obj, web_url, summary=[]): 24 | metadata = crawl_web(web_url) 25 | df = pd.DataFrame(metadata) 26 | print("Processing web information...") 27 | for title, text in zip(df["title"], df["text"]): 28 | print(f"Processing article: {title}") 29 | res = obj.summarize_custom(text) 30 | summary.append([title, res]) 31 | summaryText = open("WebSummary.txt", "w") 32 | for title, res in summary: 33 | summaryText.write(title + "\n" + res + "\n\n") 34 | summaryText.close() 35 | print("Finish processing web information.") 36 | 37 | def process_pdf(obj, pdf_url, summary=[]): 38 | metadata = crawl_pdf(pdf_url) 39 | df = pd.DataFrame(metadata) 40 | print("Processing multiple pdf file information...") 41 | for title, pdf_url in zip(df["title"], df["pdf_url"]): 42 | pdf_filename = os.path.basename(pdf_url) 43 | print(f"Processing a pdf file: {pdf_filename}") 44 | text = extract_OCR(f"./pdf/{pdf_filename}") 45 | res = obj.summarize_model(text) 46 | summary.append([title, res]) 47 | summaryText = open("PDFSummary.txt", "w") 48 | for title, res in summary: 49 | summaryText.write(title + "\n" + res + "\n\n") 50 | summaryText.close() 51 | print("Finish processing multiple pdf files.") 52 | 53 | @app.route('/') 54 | def index(): 55 | return render_template('index.html', messages=messages) 56 | 57 | @app.route('/create/', methods=['GET', 'POST']) 58 | def create(): 59 | if request.method == 'POST': 60 | web_url = request.form['text_web'] 61 | pdf_url = request.form['text_pdf'] 62 | 63 | summary_web, summary_pdf = [], [] 64 | thread_web = Thread(target=process_web, args=(obj, web_url, summary_web, )) 65 | thread_pdf = Thread(target=process_pdf, args=(obj, pdf_url, summary_pdf, )) 66 | thread_web.start() 67 | thread_pdf.start() 68 | thread_web.join() 69 | thread_pdf.join() 70 | 71 | # summary_web = process_web(obj, web_url) 72 | # summary_pdf = process_pdf(obj, pdf_url) 73 | 74 | # print("After processing:", summary_web) 75 | 76 | if len(summary_web) == 0 or len(summary_pdf) == 0: 77 | return render_template('create.html') 78 | else: 79 | messages[0] = {'text_web': web_url, 'text_pdf': pdf_url, 'answer_web': summary_web, 'answer_pdf': summary_pdf} 80 | return redirect(url_for('index')) 81 | 82 | return render_template('create.html') 83 | 84 | if __name__ == '__main__': 85 | app.run(host='0.0.0.0', port=8016, debug=True) -------------------------------------------------------------------------------- /custom_module/base.py: -------------------------------------------------------------------------------- 1 | import re 2 | from nltk.corpus import stopwords 3 | from nltk.tokenize import word_tokenize, sent_tokenize 4 | from nltk.stem.snowball import SnowballStemmer 5 | from transformers import pipeline 6 | import nltk 7 | nltk.download("stopwords") 8 | nltk.download("punkt") 9 | 10 | 11 | class TextSummarization: 12 | def __init__(self, model_name_or_path: str = "facebook/bart-large-cnn"): 13 | self.model_name_or_path = model_name_or_path 14 | self.summarizer = pipeline("summarization", model=self.model_name_or_path) 15 | 16 | 17 | def summarize_custom(self, text): 18 | # Process text by removing numbers and unrecognized punctuation 19 | processedText = re.sub("’", "'", text) 20 | processedText = re.sub("[^a-zA-Z' ]+", "", processedText) 21 | stopWords = set(stopwords.words("english")) 22 | words = word_tokenize(processedText) 23 | 24 | # Normalize words with Porter stemming and build word frequency table 25 | stemmer = SnowballStemmer("english", ignore_stopwords=True) 26 | freqTable = dict() 27 | for word in words: 28 | word = word.lower() 29 | if word in stopWords: 30 | continue 31 | elif stemmer.stem(word) in freqTable: 32 | freqTable[stemmer.stem(word)] += 1 33 | else: 34 | freqTable[stemmer.stem(word)] = 1 35 | 36 | # Normalize every sentence in the text 37 | sentences = sent_tokenize(text) 38 | stemmedSentences = [] 39 | sentenceValue = dict() 40 | for sentence in sentences: 41 | stemmedSentence = [] 42 | for word in sentence.lower().split(): 43 | stemmedSentence.append(stemmer.stem(word)) 44 | stemmedSentences.append(stemmedSentence) 45 | 46 | # Calculate value of every normalized sentence based on word frequency table 47 | # [:12] helps to save space 48 | for num in range(len(stemmedSentences)): 49 | for wordValue in freqTable: 50 | if wordValue in stemmedSentences[num]: 51 | if sentences[num][:12] in sentenceValue: 52 | sentenceValue[sentences[num][:12]] += freqTable.get(wordValue) 53 | else: 54 | sentenceValue[sentences[num][:12]] = freqTable.get(wordValue) 55 | 56 | # Determine average value of a sentence in the text 57 | sumValues = 0 58 | for sentence in sentenceValue: 59 | sumValues += sentenceValue.get(sentence) 60 | 61 | average = int(sumValues / len(sentenceValue)) 62 | 63 | # Create summary of text using sentences that exceed the average value by some factor 64 | # This factor can be adjusted to reduce/expand the length of the summary 65 | summary = "" 66 | for sentence in sentences: 67 | if sentence[:12] in sentenceValue and sentenceValue[sentence[:12]] > (1.0 * average): 68 | summary += " " + " ".join(sentence.split()) 69 | 70 | # Post-process the text in summary 71 | summary = re.sub("’", "'", summary) 72 | summary = re.sub("[^a-zA-Z0-9'\"():;,.!?— ]+", "", summary) 73 | 74 | return summary 75 | 76 | 77 | def summarize_model(self, text): 78 | # Process text by removing numbers and unrecognized punctuation 79 | processedText = re.sub("’", "'", text) 80 | processedText = re.sub("[^a-zA-Z' ]+", "", processedText) 81 | stopWords = set(stopwords.words("english")) 82 | words = word_tokenize(processedText) 83 | temp = [] 84 | 85 | for word in words: 86 | word = word.lower() 87 | if word in stopWords: 88 | continue 89 | else: 90 | temp += [word] 91 | 92 | # Split text into chunks 93 | n = 500 94 | chunks = [temp[i:i+n] for i in range(0, len(temp), n)] 95 | summary = "" 96 | 97 | for chunk in chunks: 98 | text = " ".join(chunk) 99 | res = self.summarizer(text, max_length=140, min_length=40, do_sample=False) 100 | res = str(res[0]["summary_text"]) 101 | if len(summary) == 0: 102 | summary += res.capitalize() 103 | elif res[-1] == ".": 104 | summary += " " + res.capitalize() 105 | else: 106 | summary += ". " + res.capitalize() 107 | 108 | # Post-process the text in summary 109 | summary = re.sub("’", "'", summary) 110 | summary = re.sub("[^a-zA-Z0-9'\"():;,.!?— ]+", "", summary) 111 | 112 | return summary -------------------------------------------------------------------------------- /custom_module/crawl_news.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 3, 6 | "metadata": { 7 | "id": "hLWIwfHrFNsB" 8 | }, 9 | "outputs": [], 10 | "source": [ 11 | "from bs4 import BeautifulSoup\n", 12 | "from urllib.request import urlopen\n", 13 | "from datetime import datetime\n", 14 | "import pandas as pd\n", 15 | "from urllib.request import Request\n", 16 | "\n", 17 | "crawl_base_url = 'https://paperswithcode.com/search?q_meta=&q_type=&q=satellite'\n", 18 | "\n", 19 | "crawl_from = 'July 5, 2023'\n", 20 | "crawl_to = 'July 12, 2023'" 21 | ] 22 | }, 23 | { 24 | "cell_type": "code", 25 | "execution_count": 4, 26 | "metadata": { 27 | "id": "tydl7CrbH9xS" 28 | }, 29 | "outputs": [], 30 | "source": [ 31 | "# 'July 10, 2023'\n", 32 | "date_format = '%B %d, %Y'\n", 33 | "crawl_from = datetime.strptime(crawl_from, date_format).date()\n", 34 | "crawl_to = date_object = datetime.strptime(crawl_to, date_format).date()\n", 35 | "\n", 36 | "req = Request(\n", 37 | " url=f\"{crawl_base_url}/page/{1}/\",\n", 38 | " headers={'User-Agent': 'Mozilla/5.0'}\n", 39 | ")\n", 40 | "html = urlopen(req).read().decode('utf-8')\n", 41 | "soup = BeautifulSoup(html, features='lxml')" 42 | ] 43 | }, 44 | { 45 | "cell_type": "code", 46 | "execution_count": null, 47 | "metadata": { 48 | "colab": { 49 | "base_uri": "https://localhost:8080/" 50 | }, 51 | "id": "IXEQxg6myAfX", 52 | "outputId": "ea49a4f4-26f2-4d52-a719-29473c2e84f8" 53 | }, 54 | "outputs": [], 55 | "source": [ 56 | "max_page = int(soup.find_all('a', {'class': 'page-numbers'})[-2].get_text().replace(',', ''))\n", 57 | "max_page" 58 | ] 59 | }, 60 | { 61 | "cell_type": "code", 62 | "execution_count": 7, 63 | "metadata": { 64 | "colab": { 65 | "base_uri": "https://localhost:8080/" 66 | }, 67 | "id": "vXMh3RHzW7Sq", 68 | "outputId": "2dd9bbc0-c744-4400-ab5d-81c44ce9e36c" 69 | }, 70 | "outputs": [ 71 | { 72 | "name": "stdout", 73 | "output_type": "stream", 74 | "text": [ 75 | "1\n" 76 | ] 77 | }, 78 | { 79 | "ename": "AttributeError", 80 | "evalue": "'NoneType' object has no attribute 'find_all'", 81 | "output_type": "error", 82 | "traceback": [ 83 | "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", 84 | "\u001b[0;31mAttributeError\u001b[0m Traceback (most recent call last)", 85 | "Cell \u001b[0;32mIn[7], line 13\u001b[0m\n\u001b[1;32m 11\u001b[0m html \u001b[39m=\u001b[39m urlopen(req)\u001b[39m.\u001b[39mread()\u001b[39m.\u001b[39mdecode(\u001b[39m'\u001b[39m\u001b[39mutf-8\u001b[39m\u001b[39m'\u001b[39m)\n\u001b[1;32m 12\u001b[0m soup \u001b[39m=\u001b[39m BeautifulSoup(html, features\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39mlxml\u001b[39m\u001b[39m'\u001b[39m)\n\u001b[0;32m---> 13\u001b[0m articles \u001b[39m=\u001b[39m soup\u001b[39m.\u001b[39;49mfind(\u001b[39m'\u001b[39;49m\u001b[39mmain\u001b[39;49m\u001b[39m'\u001b[39;49m, {\u001b[39m'\u001b[39;49m\u001b[39mid\u001b[39;49m\u001b[39m'\u001b[39;49m: \u001b[39m'\u001b[39;49m\u001b[39mmain\u001b[39;49m\u001b[39m'\u001b[39;49m})\u001b[39m.\u001b[39;49mfind_all(\u001b[39m'\u001b[39m\u001b[39marticle\u001b[39m\u001b[39m'\u001b[39m)\n\u001b[1;32m 14\u001b[0m \u001b[39mfor\u001b[39;00m article \u001b[39min\u001b[39;00m articles:\n\u001b[1;32m 15\u001b[0m info \u001b[39m=\u001b[39m {}\n", 86 | "\u001b[0;31mAttributeError\u001b[0m: 'NoneType' object has no attribute 'find_all'" 87 | ] 88 | } 89 | ], 90 | "source": [ 91 | "data = []\n", 92 | "stop = False\n", 93 | "for page in range(1, 2):\n", 94 | " print(page)\n", 95 | " if stop == True:\n", 96 | " break\n", 97 | " req = Request(\n", 98 | " url=f\"{crawl_base_url}/page/{1}/\",\n", 99 | " headers={'User-Agent': 'Mozilla/5.0'}\n", 100 | " )\n", 101 | " html = urlopen(req).read().decode('utf-8')\n", 102 | " soup = BeautifulSoup(html, features='lxml')\n", 103 | " articles = soup.find('main', {'id': 'main'}).find_all('article')\n", 104 | " for article in articles:\n", 105 | " info = {}\n", 106 | " date_string = article.find('time', {'class': 'entry-date published'}).get_text()\n", 107 | " date_object = datetime.strptime(date_string, date_format).date()\n", 108 | " if date_object > crawl_to:\n", 109 | " continue\n", 110 | " if date_object <= crawl_to and date_object >= crawl_from:\n", 111 | " info['date'] = date_object\n", 112 | " info['title'] = article.find('h2').get_text()\n", 113 | " info['author'] = article.find('span', {'class': 'author vcard'}).get_text()\n", 114 | " info['thumbnail'] = article.find('img')['src']\n", 115 | " url = article.find('a')['href']\n", 116 | " html = urlopen(url).read().decode('utf-8')\n", 117 | " soup = BeautifulSoup(html, features='lxml')\n", 118 | " content = soup.find('div', {'class': 'entry-content'})\n", 119 | " info['text'] = ''\n", 120 | " for el in content.find_all(recursive=False):\n", 121 | " if 'Related' not in el.get_text():\n", 122 | " info['text'] += el.get_text() + '\\n'\n", 123 | " print(\"Crawled: \", info)\n", 124 | " data.append(info)\n", 125 | " if date_object < crawl_from:\n", 126 | " stop = True\n", 127 | " break" 128 | ] 129 | }, 130 | { 131 | "cell_type": "code", 132 | "execution_count": null, 133 | "metadata": { 134 | "colab": { 135 | "base_uri": "https://localhost:8080/", 136 | "height": 1000 137 | }, 138 | "id": "RDQSl2bAlMT6", 139 | "outputId": "c8a3b9e7-f4eb-41bb-87f3-a46a3be9824e" 140 | }, 141 | "outputs": [ 142 | { 143 | "data": { 144 | "text/html": [ 145 | "\n", 146 | "\n", 147 | "
\n", 148 | "
\n", 149 | "
\n", 150 | "\n", 163 | "\n", 164 | " \n", 165 | " \n", 166 | " \n", 167 | " \n", 168 | " \n", 169 | " \n", 170 | " \n", 171 | " \n", 172 | " \n", 173 | " \n", 174 | " \n", 175 | " \n", 176 | " \n", 177 | " \n", 178 | " \n", 179 | " \n", 180 | " \n", 181 | " \n", 182 | " \n", 183 | " \n", 184 | " \n", 185 | " \n", 186 | " \n", 187 | " \n", 188 | " \n", 189 | " \n", 190 | " \n", 191 | " \n", 192 | " \n", 193 | " \n", 194 | " \n", 195 | " \n", 196 | " \n", 197 | " \n", 198 | " \n", 199 | " \n", 200 | " \n", 201 | " \n", 202 | " \n", 203 | " \n", 204 | " \n", 205 | " \n", 206 | " \n", 207 | " \n", 208 | " \n", 209 | " \n", 210 | " \n", 211 | " \n", 212 | " \n", 213 | " \n", 214 | " \n", 215 | " \n", 216 | " \n", 217 | " \n", 218 | " \n", 219 | " \n", 220 | " \n", 221 | " \n", 222 | " \n", 223 | " \n", 224 | " \n", 225 | " \n", 226 | " \n", 227 | " \n", 228 | " \n", 229 | " \n", 230 | " \n", 231 | " \n", 232 | " \n", 233 | " \n", 234 | " \n", 235 | " \n", 236 | " \n", 237 | " \n", 238 | " \n", 239 | " \n", 240 | " \n", 241 | " \n", 242 | " \n", 243 | " \n", 244 | " \n", 245 | " \n", 246 | " \n", 247 | " \n", 248 | " \n", 249 | " \n", 250 | " \n", 251 | " \n", 252 | " \n", 253 | " \n", 254 | " \n", 255 | " \n", 256 | " \n", 257 | " \n", 258 | " \n", 259 | " \n", 260 | " \n", 261 | " \n", 262 | " \n", 263 | " \n", 264 | " \n", 265 | " \n", 266 | " \n", 267 | " \n", 268 | " \n", 269 | " \n", 270 | " \n", 271 | " \n", 272 | " \n", 273 | " \n", 274 | " \n", 275 | " \n", 276 | " \n", 277 | " \n", 278 | " \n", 279 | " \n", 280 | " \n", 281 | " \n", 282 | " \n", 283 | " \n", 284 | " \n", 285 | " \n", 286 | " \n", 287 | " \n", 288 | " \n", 289 | " \n", 290 | " \n", 291 | " \n", 292 | " \n", 293 | " \n", 294 | " \n", 295 | " \n", 296 | " \n", 297 | " \n", 298 | " \n", 299 | " \n", 300 | " \n", 301 | " \n", 302 | " \n", 303 | " \n", 304 | " \n", 305 | " \n", 306 | " \n", 307 | " \n", 308 | " \n", 309 | " \n", 310 | " \n", 311 | " \n", 312 | " \n", 313 | " \n", 314 | " \n", 315 | " \n", 316 | " \n", 317 | " \n", 318 | " \n", 319 | " \n", 320 | " \n", 321 | " \n", 322 | " \n", 323 | " \n", 324 | " \n", 325 | " \n", 326 | " \n", 327 | " \n", 328 | " \n", 329 | " \n", 330 | " \n", 331 | " \n", 332 | " \n", 333 | " \n", 334 | " \n", 335 | " \n", 336 | " \n", 337 | " \n", 338 | " \n", 339 | " \n", 340 | " \n", 341 | " \n", 342 | " \n", 343 | " \n", 344 | " \n", 345 | " \n", 346 | " \n", 347 | " \n", 348 | " \n", 349 | " \n", 350 | " \n", 351 | " \n", 352 | " \n", 353 | " \n", 354 | " \n", 355 | " \n", 356 | " \n", 357 | " \n", 358 | " \n", 359 | " \n", 360 | " \n", 361 | " \n", 362 | " \n", 363 | " \n", 364 | " \n", 365 | " \n", 366 | " \n", 367 | " \n", 368 | " \n", 369 | " \n", 370 | " \n", 371 | " \n", 372 | " \n", 373 | " \n", 374 | " \n", 375 | " \n", 376 | " \n", 377 | " \n", 378 | " \n", 379 | " \n", 380 | " \n", 381 | " \n", 382 | " \n", 383 | " \n", 384 | " \n", 385 | " \n", 386 | " \n", 387 | " \n", 388 | " \n", 389 | " \n", 390 | " \n", 391 | " \n", 392 | "
datetitleauthorthumbnailtext
02023-07-12Biden nominates Space Force generals for promo...Sandra Erwinhttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — President Biden on July 11 submit...
12023-07-12Second Israeli lunar lander faces funding unce...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — As India prepares to launch its s...
22023-07-12China’s Landspace reaches orbit with methane-p...Andrew Joneshttps://i0.wp.com/spacenews.com/wp-content/upl...HELSINKI — Chinese private rocket firm Landspa...
32023-07-11U.S. sharpens plan for military space raceSandra Erwinhttps://i0.wp.com/spacenews.com/wp-content/upl...The latest U.S. military budget goes all-in on...
42023-07-11Astranis to deliver GEO broadband satellite fo...Jason Rainbowhttps://i0.wp.com/spacenews.com/wp-content/upl...TAMPA, Fla. — Astranis has sold a small broadb...
52023-07-11Former NASA astronaut to advise Vast on commer...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — Vast Space has brought in a forme...
62023-07-11Muon wins option to monitor ionosphere for Spa...Debra Wernerhttps://i0.wp.com/spacenews.com/wp-content/upl...SAN FRANCISCO – Muon Space will deliver space ...
72023-07-11Thermal imagery sector heats upDebra Wernerhttps://i0.wp.com/spacenews.com/wp-content/upl...British investment firm Seraphim Space surveye...
82023-07-10NASA awards “crossover” spacesuit task orders ...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — NASA awarded task orders to two c...
92023-07-10Astra to raise up to $65 million in stock saleJeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — Faced with dwindling cash and a s...
102023-07-10Voyager Space deepens India ties for commercia...Jason Rainbowhttps://i0.wp.com/spacenews.com/wp-content/upl...TAMPA, Fla. — Voyager Space is considering usi...
112023-07-10Benchmark raises $33 million in Series B roundDebra Wernerhttps://i0.wp.com/spacenews.com/wp-content/upl...SAN FRANCISCO – Benchmark Space Systems raised...
122023-07-10AI, quantum and nuclear technologies are key t...Miriam Klaczynskahttps://i0.wp.com/spacenews.com/wp-content/upl...LOS ANGELES – Artificial intelligence, quantum...
132023-07-10Can space governance keep up with space sustai...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...It is not uncommon to be faced with a problem ...
142023-07-07Plasmos pivots from rocket engines to VCDebra Wernerhttps://i0.wp.com/spacenews.com/wp-content/upl...SAN FRANCISCO – Plasmos, the Los Angeles-based...
152023-07-07Merger rumors swirl around Dish and EchoStarJason Rainbowhttps://i0.wp.com/spacenews.com/wp-content/upl...TAMPA, Fla. — Satellite TV broadcaster Dish Ne...
162023-07-07Interest grows for human spaceflight in EuropeJeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — As the European Space Agency cont...
172023-07-07Chinese launch firm secures fresh funding for ...Andrew Joneshttps://i0.wp.com/spacenews.com/wp-content/upl...HELSINKI — Chinese rocket firm Space Pioneer h...
182023-07-06Europe leans on SpaceX to bridge launcher gapJeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — Europe, temporarily lacking its o...
192023-07-06Viasat signs deal to commercialize European ai...Jason Rainbowhttps://i0.wp.com/spacenews.com/wp-content/upl...TAMPA, Fla. — A group founded by European air ...
202023-07-06China’s Landspace set for second methalox rock...Andrew Joneshttps://i0.wp.com/spacenews.com/wp-content/upl...HELSINKI — Chinese commercial launch firm Land...
212023-07-06Space Command argues for shift from static to ...Sandra Erwinhttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — In order to better keep tabs on a...
222023-07-06HawkEye 360 satellites to monitor illegal fish...Sandra Erwinhttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON —  HawkEye 360, a commercial operat...
232023-07-05Ariane 5 launches for the final timeJeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — One chapter in European access to...
242023-07-05Rivada gets more breathing room to deploy cons...Jason Rainbowhttps://i0.wp.com/spacenews.com/wp-content/upl...TAMPA, Fla. — International regulators have wa...
252023-07-05Radio noise from satellite constellations coul...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — Large satellite constellations ca...
262023-07-05Regulatory uncertainty as commercial human spa...Jeff Fousthttps://i0.wp.com/spacenews.com/wp-content/upl...WASHINGTON — As two companies prepare to begin...
\n", 393 | "
\n", 394 | " \n", 404 | "\n", 405 | "\n", 406 | "\n", 407 | "
\n", 408 | " \n", 419 | "
\n", 420 | "\n", 421 | "\n", 452 | "\n", 453 | " \n", 460 | "\n", 461 | " \n", 472 | " \n", 509 | "\n", 510 | " \n", 534 | "
\n", 535 | "
\n" 536 | ], 537 | "text/plain": [ 538 | " date title \\\n", 539 | "0 2023-07-12 Biden nominates Space Force generals for promo... \n", 540 | "1 2023-07-12 Second Israeli lunar lander faces funding unce... \n", 541 | "2 2023-07-12 China’s Landspace reaches orbit with methane-p... \n", 542 | "3 2023-07-11 U.S. sharpens plan for military space race \n", 543 | "4 2023-07-11 Astranis to deliver GEO broadband satellite fo... \n", 544 | "5 2023-07-11 Former NASA astronaut to advise Vast on commer... \n", 545 | "6 2023-07-11 Muon wins option to monitor ionosphere for Spa... \n", 546 | "7 2023-07-11 Thermal imagery sector heats up \n", 547 | "8 2023-07-10 NASA awards “crossover” spacesuit task orders ... \n", 548 | "9 2023-07-10 Astra to raise up to $65 million in stock sale \n", 549 | "10 2023-07-10 Voyager Space deepens India ties for commercia... \n", 550 | "11 2023-07-10 Benchmark raises $33 million in Series B round \n", 551 | "12 2023-07-10 AI, quantum and nuclear technologies are key t... \n", 552 | "13 2023-07-10 Can space governance keep up with space sustai... \n", 553 | "14 2023-07-07 Plasmos pivots from rocket engines to VC \n", 554 | "15 2023-07-07 Merger rumors swirl around Dish and EchoStar \n", 555 | "16 2023-07-07 Interest grows for human spaceflight in Europe \n", 556 | "17 2023-07-07 Chinese launch firm secures fresh funding for ... \n", 557 | "18 2023-07-06 Europe leans on SpaceX to bridge launcher gap \n", 558 | "19 2023-07-06 Viasat signs deal to commercialize European ai... \n", 559 | "20 2023-07-06 China’s Landspace set for second methalox rock... \n", 560 | "21 2023-07-06 Space Command argues for shift from static to ... \n", 561 | "22 2023-07-06 HawkEye 360 satellites to monitor illegal fish... \n", 562 | "23 2023-07-05 Ariane 5 launches for the final time \n", 563 | "24 2023-07-05 Rivada gets more breathing room to deploy cons... \n", 564 | "25 2023-07-05 Radio noise from satellite constellations coul... \n", 565 | "26 2023-07-05 Regulatory uncertainty as commercial human spa... \n", 566 | "\n", 567 | " author thumbnail \\\n", 568 | "0 Sandra Erwin https://i0.wp.com/spacenews.com/wp-content/upl... \n", 569 | "1 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 570 | "2 Andrew Jones https://i0.wp.com/spacenews.com/wp-content/upl... \n", 571 | "3 Sandra Erwin https://i0.wp.com/spacenews.com/wp-content/upl... \n", 572 | "4 Jason Rainbow https://i0.wp.com/spacenews.com/wp-content/upl... \n", 573 | "5 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 574 | "6 Debra Werner https://i0.wp.com/spacenews.com/wp-content/upl... \n", 575 | "7 Debra Werner https://i0.wp.com/spacenews.com/wp-content/upl... \n", 576 | "8 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 577 | "9 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 578 | "10 Jason Rainbow https://i0.wp.com/spacenews.com/wp-content/upl... \n", 579 | "11 Debra Werner https://i0.wp.com/spacenews.com/wp-content/upl... \n", 580 | "12 Miriam Klaczynska https://i0.wp.com/spacenews.com/wp-content/upl... \n", 581 | "13 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 582 | "14 Debra Werner https://i0.wp.com/spacenews.com/wp-content/upl... \n", 583 | "15 Jason Rainbow https://i0.wp.com/spacenews.com/wp-content/upl... \n", 584 | "16 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 585 | "17 Andrew Jones https://i0.wp.com/spacenews.com/wp-content/upl... \n", 586 | "18 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 587 | "19 Jason Rainbow https://i0.wp.com/spacenews.com/wp-content/upl... \n", 588 | "20 Andrew Jones https://i0.wp.com/spacenews.com/wp-content/upl... \n", 589 | "21 Sandra Erwin https://i0.wp.com/spacenews.com/wp-content/upl... \n", 590 | "22 Sandra Erwin https://i0.wp.com/spacenews.com/wp-content/upl... \n", 591 | "23 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 592 | "24 Jason Rainbow https://i0.wp.com/spacenews.com/wp-content/upl... \n", 593 | "25 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 594 | "26 Jeff Foust https://i0.wp.com/spacenews.com/wp-content/upl... \n", 595 | "\n", 596 | " text \n", 597 | "0 WASHINGTON — President Biden on July 11 submit... \n", 598 | "1 WASHINGTON — As India prepares to launch its s... \n", 599 | "2 HELSINKI — Chinese private rocket firm Landspa... \n", 600 | "3 The latest U.S. military budget goes all-in on... \n", 601 | "4 TAMPA, Fla. — Astranis has sold a small broadb... \n", 602 | "5 WASHINGTON — Vast Space has brought in a forme... \n", 603 | "6 SAN FRANCISCO – Muon Space will deliver space ... \n", 604 | "7 British investment firm Seraphim Space surveye... \n", 605 | "8 WASHINGTON — NASA awarded task orders to two c... \n", 606 | "9 WASHINGTON — Faced with dwindling cash and a s... \n", 607 | "10 TAMPA, Fla. — Voyager Space is considering usi... \n", 608 | "11 SAN FRANCISCO – Benchmark Space Systems raised... \n", 609 | "12 LOS ANGELES – Artificial intelligence, quantum... \n", 610 | "13 It is not uncommon to be faced with a problem ... \n", 611 | "14 SAN FRANCISCO – Plasmos, the Los Angeles-based... \n", 612 | "15 TAMPA, Fla. — Satellite TV broadcaster Dish Ne... \n", 613 | "16 WASHINGTON — As the European Space Agency cont... \n", 614 | "17 HELSINKI — Chinese rocket firm Space Pioneer h... \n", 615 | "18 WASHINGTON — Europe, temporarily lacking its o... \n", 616 | "19 TAMPA, Fla. — A group founded by European air ... \n", 617 | "20 HELSINKI — Chinese commercial launch firm Land... \n", 618 | "21 WASHINGTON — In order to better keep tabs on a... \n", 619 | "22 WASHINGTON —  HawkEye 360, a commercial operat... \n", 620 | "23 WASHINGTON — One chapter in European access to... \n", 621 | "24 TAMPA, Fla. — International regulators have wa... \n", 622 | "25 WASHINGTON — Large satellite constellations ca... \n", 623 | "26 WASHINGTON — As two companies prepare to begin... " 624 | ] 625 | }, 626 | "execution_count": 10, 627 | "metadata": {}, 628 | "output_type": "execute_result" 629 | } 630 | ], 631 | "source": [ 632 | "df = pd.DataFrame(data)\n", 633 | "df" 634 | ] 635 | }, 636 | { 637 | "cell_type": "code", 638 | "execution_count": null, 639 | "metadata": { 640 | "colab": { 641 | "base_uri": "https://localhost:8080/" 642 | }, 643 | "id": "7_LUhENg8p8r", 644 | "outputId": "dc22360b-b7d9-40cc-bc62-b198bec01f54" 645 | }, 646 | "outputs": [ 647 | { 648 | "name": "stdout", 649 | "output_type": "stream", 650 | "text": [ 651 | "{'date': datetime.date(2023, 7, 12), 'title': 'Biden nominates Space Force generals for promotion', 'author': 'Sandra Erwin', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/230307-F-WA228-1070-scaled.jpg?resize=1200%2C900&ssl=1', 'text': 'WASHINGTON — President Biden on July 11 submitted to the U.S. Senate the nominations of two three-star Space Force leaders for promotion to four-star generals.\\nLt. Gen. Stephen Whiting, commander of the Space Force’s Space Operations Command; and Lt. Gen. Michael Guetlein, commander of the Space Force’s Space Systems Command, have been nominated to the rank of general and to receive a fourth star.\\xa0\\nWhiting is expected to become the next commander of U.S. Space Command, replacing Gen. James Dickinson. Guetlein is expected to become the next vice chief of space operations of the Space Force, replacing Gen. David “DT” Thompson.\\xa0\\nThe nominations will be considered by the Senate Armed Services Committee.\\nThe White House also submitted the nomination of Lt. Gen. Philip Garrant, deputy chief of space operations for strategy, plans, programs and requirements, to a new assignment at the same rank.\\nGarrant is expected to replace Guetlein as the commander of the Space Systems Command, based in Los Angeles.\\xa0\\nBlock on military nominations\\nIf the Space Force nominations are approved by the Senate Armed Services Committee, it’s unclear when they will reach the Senate floor. Sen. Tommy Tuberville (R-Ala.)\\xa0 has placed a hold on general and flag officer nominations to protest a DoD policy that covers certain abortion-related travel expenses for service members.\\nTuberville said he will continue to block military nominations until the Pentagon changes the policy or it is passed through legislation.\\xa0\\nThe Pentagon said as many as 265 general and flag officer nominations have been delayed in the Senate, affecting the smooth transition of leadership.\\n'}\n", 652 | "{'date': datetime.date(2023, 7, 12), 'title': 'Second Israeli lunar lander faces funding uncertainty', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2021/10/beresheet-model.jpg?resize=800%2C495&ssl=1', 'text': 'WASHINGTON — As India prepares to launch its second lunar lander mission, the fate of a second Israeli lander is in doubt after the organization developing it lost a major source of funding.\\nIndia’s Chandrayaan-3 spacecraft is scheduled to launch July 14 on a Geosynchronous Satellite Launch Vehicle Mark 3, also known as LVM-3, from the Satish Dhawan Space Centre. The spacecraft will gradually go from a geostationary transfer orbit to a low lunar orbit, from which Chandrayaan-3 will descend to the lunar surface.\\nChandrayaan-3 is similar to India’s first lunar lander flown as part of the Chandrayaan-2 mission, which crashed attempting a soft landing in September 2019. Chandrayaan-3 incorporates several revisions, such as additional fuel, based on the investigation into the failed landing.\\nThat crash took place five months after Beresheet, a spacecraft originally developed by Israeli non-profit organization SpaceIL to compete for the Google Lunar X Prize, crashed attempting its own lunar landing. SpaceIL and Israel Aerospace Industries, which built the spacecraft, later said the lander crashed because one of its inertial measurement units malfunctioned.\\nSpaceIL announced its intent to pursue a second mission, called Beresheet 2. It would be significantly different from the original mission, with two smaller landers deployed from an orbiter. The mission had been slated to launch in 2025, a date confirmed in a January 2023 announcement of a joint statement of intent between the Israel Space Agency and NASA to cooperate on Beresheet 2. NASA agreed to provide an instrument and communications support for the mission.\\nThat schedule, and the mission itself, is now in question. Like the first mission, SpaceIL projected using philanthropic donations to fund Beresheet 2. However, in May a group of donors announced they were halting future payments to the project after spending $45 million, nearly half its estimated $100 million cost.\\nIn a statement representing the donors, Morris Kahn, a billionaire who also supported the original Beresheet mission, said the decision to halt future payments was not related to any problems with Beresheet 2. “These times obligate us to invest our resources and time in other philanthropic projects,” he stated.\\nSpaceIL said at the time it would seek alternative funding to continue the mission, but has provided no updates on those efforts since then. The organization did not respond to questions July 10 about the status of Beresheet 2.\\nIn a June 27 presentation at the European Lunar Symposium, Dan Blumberg, chairman of the Israel Space Agency, said work on Beresheet 2 was continuing for now. “We want to do more than we did in the previous one,” he said, including enhanced opportunities for international cooperation. In addition to the NASA agreement, SpaceIL has an agreement with the German agency DLR, which will provide a navigation system for the lander.\\nHe also emphasized educational outreach for the mission. Students will have opportunities to control the orbiter during its two-year mission, selecting regions on the surface to photograph, he said.\\nBlumberg hinted at SpaceIL’s financial problems in his brief talk, but did not go into details about funding for the mission. “There is a funding issue that we are still dealing with,” he said. “But we’re getting there.”\\n'}\n", 653 | "{'date': datetime.date(2023, 7, 12), 'title': 'China’s Landspace reaches orbit with methane-powered Zhuque-2 rocket', 'author': 'Andrew Jones', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/zhuque-2-y2-0100july12-2023-ourspace-1.jpg?resize=800%2C600&ssl=1', 'text': 'HELSINKI — Chinese private rocket firm Landspace achieved a global first late Tuesday by reaching orbit with a methane-fueled rocket.\\nThe 49.5-meter-long Zhuque-2 lifted off from Jiuquan Satellite Launch Center in the Gobi Desert at 9:00 p.m. Eastern on July 11.\\xa0\\nLandspace and Chinese state media announced that the second Zhuque-2 reached orbit, making it the first methane-fueled globally to reach orbit. This was later verified by U.S. Space Force space tracking data, showing an object in a 431 by 461-kilometer Sun-synchronous orbit with an inclination of\\xa097.3 degrees.\\nThe Zhuque-2 mission carried no payload and the rocket’s first stage was not recovered. An unconfirmed number of payloads were lost in the launcher’s first flight in December 2022.\\xa0\\nZhuque-2 beats a range of other methalox rockets, including SpaceX’s Starship, the ULA Vulcan, Blue Origin’s New Glenn, Rocket Lab’s Neutron and Terran R from Relativity Space, in reaching orbit. These other launch vehicles will be much larger and feature much greater payload capacity.\\nA methane-liquid oxygen propellant mix offers advantages in performance and reduces issues of soot formation and coking for purposes of reusability.\\nThe successful launch also makes Landspace the second private Chinese launch firm to reach orbit with a liquid propellant rocket. This follows the success of Space Pioneer with its Tianlong-2 rocket in April this year.\\n\\nhttps://twitter.com/CNSAWatcher/status/1679006943698976768\\n\\nTaken together, the achievements indicate a breakthrough and growing level of maturity in Chinese commercial space launch efforts.\\nLandspace has already begun assembling its third Zhuque-2 (“Vermillion Bird-2”), indicating that another launch could come before the end of the year. Space Pioneer says it has multiple orders for launches for the Tianlong-2, and aims to launch the Falcon 9-class Tianlong-3 in the first half of 2024.\\nLandspace CEO Zhang Changwu told Chinese language Global Times tabloid that the company could now begin mass-producing the Zhuque-2, having finalized and verified its design.\\nLandspace’s Zhuque-2 is powered by gas generator engines producing 268 tons of thrust. Future Zhuque-2 launches with upgraded second stage engines will be capable of delivering a 6,000-kilogram payload capacity to a 200-kilometer low Earth orbit (LEO), or 4,000 kilograms to 500-kilometer sun-synchronous orbit (SSO), according to Landspace.\\nThe rocket has a diameter of 3.35 meters—the same as a number of national Long March rockets—and a take-off mass of 219 tons.\\xa0\\nThe rocket is currently expendable but Landspace is working on a restartable version of the 80-ton-thrust TQ-12 engine which powers the Zhuque-2 first stage.\\xa0\\n\\nView from above:https://t.co/2pY1F2mgFW pic.twitter.com/GDAj7uWqS7— Cosmic Penguin (@Cosmic_Penguin) July 12, 2023\\n\\nLandspace is one of China’s first commercial rocket firms. It was established in 2015, soon after the Chinese government opened up parts of the space sector to private capital in late 2014, which is seen to be a reaction to developments in the U.S.\\xa0\\nIt is also one of the best-funded Chinese launch firms, but the company’s journey to orbit has not been smooth. Its first launch, which took place in October 2018, used the smaller and simpler solid-propellant Zhuque-1. The mission ended in failure, and the company announced it would not repeat the attempt and instead focus on its methane-fueled Zhuque-2.\\nThe company has however built up infrastructure during this time. Landspace has set up an intelligent manufacturing base in Huzhou, Zhejiang Province and established a $1.5 billion medium and large-scale liquid rocket assembly and test plant at Jiaxing, also in Zhejiang.\\nThe company is now in a position to secure contracts for the Zhuque-2 but faces a field of competition. Other Chinese firms, including iSpace, Galactic Energy and Deep Blue Aerospace, are working on their own, reusable liquid propellant rockets.\\nSpace Pioneer and other later movers such as OrienSpace are developing larger rockets which are targeting contracts to launch batches of satellites for China’s national satellite internet megaconstellation project, named Guowang.\\nChina recently opened a call for space station commercial cargo proposals, further indicating that commercial firms will have a growing role to play in the country’s space sector.\\nThe Zhuque-2 launch was China’s 27th orbital mission of 2023, with a total of more than 70 launches planned from state-owned main space contractor CASC and commercial players. Landspace, Space Pioneer, Galactic Energy, iSpace and Expace have so far complemented national launches this year.\\n'}\n", 654 | "{'date': datetime.date(2023, 7, 11), 'title': 'U.S. sharpens plan for military space race', 'author': 'Sandra Erwin', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/Next-GenerationPolarisIndispensableinHighlyContestedSpace.jpg?resize=800%2C600&ssl=1', 'text': 'The latest U.S. military budget goes all-in on the notion that resilience will be a core feature of space programs. As evidence, the term surpasses 300 mentions in the Space Force’s 2024 budget documents. \\n“It’s amazing how many times you see the word resiliency in the budget justification materials,” said analyst Sam Wilson of the Aerospace Corporation’s Center for Space Policy and Strategy. \\nThe emphasis on resilience — or adaptability in the face of attacks — reflects the priorities set by the new chief of space operations Gen. Chance Saltzman. The running theme in the budget is the need to ensure U.S. access to space and shore up capabilities to compete with space powers like China and Russia. \\n“China, our pacing challenge, is the most immediate threat in, to, and from space for which the Space Force must maintain technological advantage,” Saltzman said in testimony to the Senate Armed Services Committee earlier this year. \\n“Russia, while less capable, remains an acute threat that is developing asymmetric counter-space systems meant to neutralize American satellites,” said Saltzman. \\n‘Competitive endurance’ \\nA push for resilience is part of a multifaceted strategy Saltzman rolled out in March called “competitive endurance” to guide Space Force plans to deter and combat adversaries. \\nVice Chief of Space Operations Gen. David “DT” Thompson said the Space Force’s central responsibility is ensuring U.S. military forces and allies have access to satellite services. \\nAnd that is why resilience is so critical, he said June 15 at the Defense One Tech Summit. \\n“Ukraine showed that proliferation works. It leads to resilient architectures,” Thompson said, referring to Russia’s repelled attempts to jam SpaceX’s Starlink internet satellite service used by Ukrainian forces. \\nThe U.S. armed forces and allies today rely on a secure space infrastructure — for communications, early warnings of ballistic missile attacks and other services — that the Space Force provides from a relatively small number of geostationary satellites in harder to reach orbits. “And we have to continue to understand how to defend and protect those against counter-space attacks,” he said. \\n“I think that’s probably the biggest concern right now,” Thompson said, “ensuring we have enough resilience in both proliferated constellations, and those small numbers of larger systems to be able to defend them.”\\nA shift to proliferated constellations is already under way. \\nThe Space Development Agency — a procurement organization under the Space Force — is moving forward with a multibillion dollar plan to field a mesh network of satellites in low Earth orbit for missile warning and for data transport. The agency in April launched its first batch of 10 satellites and announced plans to launch its next 13 spacecraft in late July. \\nProliferation and redundancy make the loss of a few satellites tolerable, said Thompson, and bolster deterrence because targeting them imposes increased cost on an adversary. \\nThis is how the Space Force plans to deal with China’s threats in the long term, he said. “We are in a long competition with China. They absolutely believe that.” \\n“We can’t look at a finish line in this competition,” Thompson added. “We have to think what we need to do to compete for the next 10 to 50 years.” \\nLt. Gen. DeAnna Burt, chief operations officer of the U.S. Space Force, is briefed on NATO space operations at the NATO Space Center at Ramstein Air Base, Germany, in May. Credit: U.S. Space Force\\nAvoid surprises \\nSpace Force’s competitive endurance strategy has three tenets: avoid operational surprise, deny first-mover advantage, and conduct responsible counter-space campaigns that don’t create long-lasting debris in orbit. \\n“If we get this right we will deter a crisis or conflict from extending into space,” Lt. Gen. DeAnna Burt, deputy chief of space operations, cyber and nuclear, said in an interview. \\n“But if needed, we will ensure space access for the joint force in a manner that maintains safety in the space domain for all responsible actors.” \\nAvoiding unwelcome surprises in space requires continuous awareness of what adversaries are doing in orbit, she said. \\nThe U.S. wants to prevent a Pearl Harbor-style attack in space, which is why the Space Force is turning more attention to space domain awareness, Burt said. “I have to be able to attribute any actions that are nefarious; I have to track and maintain custody of things I consider threats, and provide indications and warnings to other people.” \\nThe second tenet, denying first mover advantage in space, is where resilient architectures play a central role, she explained. \\n“If the enemy knows that attacking U.S. interests in space will require such a massive effort that it will become impractical or self-defeating, we will be deterring such actions in the first place,” she said. “It would be a case when the juice isn’t worth the squeeze.” \\nThe final tenet, conducting responsible counter-space campaigns, is about the idea that, if American assets were threatened, the U.S. would respond appropriately and would try to minimize the creation of debris in orbit. \\n“Polluting the domain is not what I want to do,” said Burt. There are ongoing discussions with allies about how operations could be conducted in a responsible manner, she said. \\nThe Space Force has been secretive about what technologies it might be developing to attack adversaries’ satellites without creating significant debris. \\n“General Saltzman has publicly talked about deliveries that will come in the 2026 timeframe,” Burt said. “And I think that’s when you’re going to see us start to figure out how we’re going to message that.” \\nAccess to space \\nChina and Russia have adopted strategies based on disabling adversaries’ space communications and navigation systems, said Brig. Gen. Anthony Mastalir, commander of U.S. Space Forces Indo-Pacific, a subordinate unit to U.S. Indo-Pacific Command. \\nIt is stunning how fast China has modernized its space infrastructure to enable military capabilities like precision-guided missiles, Mastalir wrote in a June 14 article published by the Air University’s Journal of Indo-Pacific Affairs. \\nBrig. Gen. Anthony Mastalir, commander of United States Space Forces, Indo-Pacific, speaks at a U.S. Indo-Pacific Command ceremony at Joint Base Pearl Harbor–Hickam, Hawaii. Credit: U.S. Navy photo by Mass Communication Specialist 1st Class Anthony J. Rivera\\n“China has made unprecedented investments in its on-orbit capabilities over the past three years,” Mastalir noted. \\nThe U.S. military has been particularly wary of China’s secretive spaceplane — which has been described as a knockoff of the U.S. Air Force’s X-37B that can remain in orbit for years. China’s spaceplane has flown two long-endurance missions, and conducted proximity and capture maneuvers with a subsatellite, according to data from the commercial space-tracking firm LeoLabs.\\nMastalir pointed out that China deployed about 160 satellites in 2022, many of which will support military operations. By some estimates, China plans to launch 200 spacecraft in 2023. \\n“While many strategists are rightly concerned about China’s and Russia’s fielding of anti-satellite weapons designed to degrade or destroy U.S. satellites in space, it is important to note that much of China’s space investment enables its long-range precision strike capability,” he wrote. \\nSpace Force leaders also worry that a breakdown in U.S.-China communication and an underlying distrust that goes both ways could lead to miscalculations. \\nChina’s lack of transparency about its own space activities makes it difficult to reduce those risks, Burt warned in May at a space policy conference hosted by Arizona State University. \\nDefense Secretary Lloyd Austin cautioned June 1 that China’s reluctance to engage with U.S. defense leaders could result in “an incident that could very, very quickly spiral out of control.” \\nMore recently, Lt. Gen. John Shaw, deputy commander of U.S. Space Command, said the “biggest dynamic right now in our relationship with China with regard to space is a lack of communication and virtually zero transparency.” \\nThe absence of dialogue and interaction creates conditions for “miscommunication, misperception, misinterpretation, and then things could go wrong. And that can happen in any domain,” Shaw said June 14 at the Secure World Foundation’s Summit for Space Sustainability. \\nU.S. Space Command traffic watchers at Vandenberg Space Force Base in California issue warnings of close approaches in orbit or potential collision to satellite operations and national agencies, including the Chinese government. But when warnings are sent to Chinese email addresses, Shaw said, “We never get a response. Never.” \\n“Even the Russians know how to communicate with us,” Shaw said. “We don’t have anything like that with the Chinese and that’s the biggest hindrance to transparent operations.” \\n‘Gray zone’ competition \\nExperts agree that space powers are most likely to attack rival satellites in a conflict through “non kinetic” means. \\nStill, both China and Russia have demonstrated they can blow up satellites by striking them with kinetic weapons such as ground-based missiles. However, the United States needs to be more concerned about activities that fall below the threshold of an act of war but are damaging nonetheless, said Todd Harrison, aerospace industry analyst and managing director of Metrea Strategic Insights. \\nCyber and electronic jamming attacks are in the murky category of “gray zone” provocations that fall somewhere between low-intensity conflict and all-out war. \\nMilitary service members from the 4th Space Control Squadron stand in front of the Counter Communications System Block 10.2 at Peterson Air Force Base, Colorado. The CCS is a transportable space electronic warfare system that reversibly denies adversary satellite communications. Credit: U.S. Air Force Photo / Andrew Bertain\\nThese non-kinetic attacks could still do lasting damage to satellites and their ground systems. A satellite that can’t see, think or communicate is as good as dead. “It’s probably going to be in the ground segment that you’ve got to be the most worried about, but it could be on the space side as well as there are now lasers that can blind sensors,” Harrison said. \\n“An adversary like China or others might think that they can get away with using that,” he said, “and could severely hamper our ability to sense and to communicate from space.” \\nThe U.S. military “historically has not really had a good response to gray zone activities” and that should concern the Space Force, said John Klein, a senior fellow and strategist at Falcon Research, and adjunct professor at George Washington University’s Space Policy Institute. \\n“We kind of let it happen with no repercussions,” Klein said. “It doesn’t have to be a military response. But the U.S. has to make it known it’s unacceptable behavior, and that there will be consequences.” \\nIt is notable, though, that the Space Force is openly talking about deterring China and about the importance of resilience in U.S. space networks, Klein said. \\nHe observed that space resilience is not a new concept, as the Pentagon for decades has studied the issue of how to reduce the vulnerability of U.S. satellites. \\n“What is new is that the U.S. Space Force more specifically recognizes resilience as part of deterrence. It’s deterrence by denial of benefits,” Klein said. “It’s telling enemies that no matter what you do, it’s not going to matter. You’re not going to stop me.” \\nU.S. still vulnerable \\nA move to more diversified constellations is necessary but might still not be enough to deter China, warned Charles Galbreath, senior fellow for space power studies at the Air and Space Forces Association’s Mitchell Institute. \\n“While the proliferated LEO approach garners a great deal of attention, it is not the only method the Space Force can employ to increase the resiliency of its architecture,” Galbreath, former deputy chief technology and innovation officer of the U.S. Space Force, said June 26 at a Mitchell Institute event. \\nIn a white paper titled “Building U.S. Space Force Counter-Space Capabilities,” Galbreath suggested the Space Force consider the “enduring military practice of deception to confuse adversaries and complicate their ability to target U.S. satellites.” \\nFor example, the U.S. could build satellite payloads or components in ways that would camouflage their functions. \\nThe Space Force also should expand its use of protection measures such as nuclear hardening and anti-jam systems, Galbreath said. \\nSpeaking June 12 at the Mitchell Institute, Thompson, the vice chief of space operations, recognized that deterrence may work in theory but not in practice. And if diplomacy and deterrence fail, the military has to prepare for the worstcase scenario. \\n“This is where we’ve spent a lot of time working with the other services,” said Thompson, to make sure they understand their dependence on space assets and figure out how to ensure the Army, Navy and Air Force can continue operating if U.S. satellites were targeted. \\nOther discussions on this subject are taking place with the private sector, as the Space Force tries to figure out contracting options to secure access to commercial space services during conflicts. \\nUnder an initiative known as Commercial Augmentation Space Reserves, the Space Force is looking at establishing agreements with companies to ensure that services like satellite communication and remote sensing are prioritized for U.S. government use during national security emergencies. \\nThe role of the private space sector in national security is significant, said Thompson. “When you think about proliferation and diversity, it’s not just the number of satellites, it’s also allies and commercial partners.” \\nNeed to keep space usable \\nEven if the U.S. has superior space technology, a vibrant private space sector and more powerful weaponry, the reality is that rival nations have the means today to destroy satellites and create “devastating impacts on the environment that will be harmful to the use of space for decades and perhaps centuries to come,” Thompson said. \\n“As we look at the proliferation of space capabilities, it will be increasingly difficult to deny some level of use of space to an adversary,” he added. \\nWhile preparing for a protracted competition with rival powers, said Thompson, the Space Force has to make sure the U.S. government has accurate intelligence so diplomats and military leaders know what’s happening in the space domain. \\n“If we’re operationally or strategically surprised, shame on us.”\\n\\nThis article originally appeared in the July 2023 issue of SpaceNews magazine.\\n'}\n", 655 | "{'date': datetime.date(2023, 7, 11), 'title': 'Astranis to deliver GEO broadband satellite for the Philippines next year', 'author': 'Jason Rainbow', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/Astranis-Philppines-scaled.jpg?resize=800%2C600&ssl=1', 'text': 'TAMPA, Fla. — Astranis has sold a small broadband satellite launching to geostationary orbit next year to a telco in the Philippines looking for support from the country’s government, the Californian manufacturer announced July 11.\\nOrbits Corp, the satellite services arm of Philippine internet service provider HTechCorp, plans to sell at least some of the capacity to the government to help connect up to two million people across 5,000 remote and rural communities in the archipelago.\\nOnly 11,000 of the country’s 42,000 local communities are covered by fiber, and the government has identified many of those left unconnected as Geographically Isolated and Disadvantaged Areas (GIDA).\\nThe Philippines has made connecting GIDA communities a major priority, Astranis CEO John Gedmark said, and the companies are seeking ways to help the government bring internet to areas where people make less than $5,000 a year on average.\\n“Most of the internet penetration in the Philippines is confined to the metropolitan areas,” said Atty Augusto Baculio, a former legislator for the Philippines who now leads Orbits Corp.\\n“Outside of that, going to the inner villages, over mountains, across islands — that’s when you have intermittent access to connectivity, or none at all.”\\n\\n\\n\\nThe companies did not disclose government commitments or financial details about a satellite they say would be the first dedicated to providing internet services to the country.\\nLast year, the Philippines permitted SpaceX to provide services from its global Starlink broadband network in low Earth orbit to help bridge the country’s digital divide.\\nAstranis operates its satellites on behalf of customers, who lease the capacity over their eight-year lifetimes.\\xa0\\nAt around 400 kilograms, the company’s dishwasher-sized satellites are smaller than typical geostationary spacecraft that weigh thousands of kilograms and are scaled for smaller, regional coverage.\\nThe satellite for the Philippines is one of five slated to launch together in 2024 on a dedicated rocket Astranis has not disclosed. A pair of satellites for Mexican telco Apco Networks is joining this mission, which Astranis calls Block 3, and customers for the other two remain undisclosed.\\nLater this year, Astranis is slated to deploy four satellites on a dedicated SpaceX Falcon 9 mission as part of Block 2. It has only revealed customers for three of these: two satellites for\\xa0\\nU.S.-based mobile satellite connectivity specialist Anuvu and one for cellular backhaul provider Andesat of Peru.\\nFirst deal following inaugural launch\\nEight-year-old Astranis launched its debut satellite April 30 as a secondary payload to a SpaceX Falcon Heavy carrying the 6,400-kilogram ViaSat-3 broadband satellite to orbit.\\nCalled Arcturus, this inaugural satellite was sold to Alaska-based telco Pacific Dataport Inc., meaning Astranis has disclosed customers for seven of the 10 satellites it says are on order.\\nGedmark said May 24 that Arcturus was performing at speeds of around 9 gigabits per second (Gbps) in early tests, despite being specced for 7.5 Gbps. The company, which had expected at the time to have completed calibration and health checks by mid-June so the satellite could enter service, declined to give an update on its roll-out.\\n'}\n", 656 | "{'date': datetime.date(2023, 7, 11), 'title': 'Former NASA astronaut to advise Vast on commercial space station efforts', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/reisman.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — Vast Space has brought in a former NASA astronaut and SpaceX official to serve as an adviser for its plans to develop commercial space stations.\\nVast announced July 11 that it has appointed Garrett Reisman as a human spaceflight adviser. He will assist the company, which announced plans in May to develop a single-module station called Haven-1 as a precursor for future, larger space stations.\\nReisman joined NASA’s astronaut corps in 1998 and spent three months on the International Space Station in 2008 as part of the Expedition 16 and 17 crews, and another 12 days on the STS-132 shuttle mission to the ISS in 2010. He retired from the astronaut corps in 2011 and worked for SpaceX for several years in various capacities, including director of space operations. He is currently a professor of astronautical engineering at the University of Southern California.\\nIn an interview, Reisman said a former SpaceX colleague, now working at Vast, reached out to him with a technical question. That led to discussions with Vast executives about assisting the company. “I got excited by what I saw, which was a really interesting company doing some really cool work,” he said.\\nHe said he is particularly interested in the ability of Vast’s space stations to spin and create artificial gravity. “We have a lot of data at zero G and tons of data at one G about what it does to the human body. We have no idea what happens in between,” he said. Those stations, he said, could help answer questions such as how much partial gravity is needed to prevent some of the deleterious effects of microgravity on human physiology.\\nMax Haot, president of Vast, said the company was looking for someone intimately familiar with human spaceflight to assist its development of Haven-1. “The company had not yet gotten to the point where we had anyone with human spaceflight experience, who had gone to space,” he said. “I think it’s obvious to everyone that if you design a space station, you should have at least one, if not more, astronauts on board to help you with the process.”\\nReisman, he said, was an ideal fit because of both his NASA astronaut experience and work at SpaceX. Vast is working with SpaceX to launch Haven-1 and use Crew Dragon spacecraft to transport astronauts to the station.\\nReisman said he plans to focus on operations and safety. “I think those two things go hand-in-hand,” he said, to “make sure we hope for the best and plan for all possible outcomes.”\\nSince Vast announced Haven-1 in May, the company has signed an agreement with Impulse Space to use thrusters from that company on the spacecraft. It was also one of seven companies that won unfunded NASA Space Act Agreements to support work on commercial space capabilities.\\n“This is a great opportunity for us to build the relationship with NASA but also leverage a lot of their expertise,” Haot said of that agreement. It also includes “many technology milestones” to track its progress, he noted.\\nReisman, besides advising Vast, will continue as a senior adviser for SpaceX. He said that company had no issues with him also advising Vast since the two companies are working together.\\nHe also continues his teaching at USC, where he says he sees strong interest from students in careers in commercial human spaceflight. Some of his students have taken his class on human spaceflight out of a “vague interest” in the topic but have since gone on to pursue careers in the field.\\n“In a way I’m very jealous because it’s such a wonderful time to be working in this industry,” he said. “Coming as an engineer straight out of college there’s so many opportunities now that I didn’t have when I was finishing up my education.”\\n'}\n", 657 | "{'date': datetime.date(2023, 7, 11), 'title': 'Muon wins option to monitor ionosphere for Space Force', 'author': 'Debra Werner', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/rsz_1preview_009_1-1.png?resize=800%2C600&ssl=1', 'text': 'SAN FRANCISCO – Muon Space will deliver space weather data to the U.S. Space Force under a $400,000 contract option announced July 11.\\nUnder the original $2.8 million contract with Air Force Life Cycle Management Center Weather Systems Branch and the Defense Innovation Unit awarded in 2022, Muon will deliver terrestrial weather products to the Air Force\\xa0557th Weather Wing from a space-based prototype microwave sensor. The option directs Muon to also monitor the ionosphere for the U.S. Space Force through September 2024.\\n“Obviously solar activity influences the ionosphere,” Muon CEO Johnny Dyer told\\xa0SpaceNews.\\xa0“This is a really important measurement for users of RF communication systems.”\\nThe Department of Defense plans to evaluate data from Muon’s second satellite, MuSat-2, for operational weather forecasting, ionospheric modeling and climate change assessment. Data collected during the pilot program will be available for government-run Observing System Simulation Experiments or OSSEs. Government agencies conduct OSSEs to assess the impact of new datasets on forecasts and models. \\nFor example, the Air Force could simulate the impact of data gathered during the MuSat-2 pilot on numerical weather prediction. Pilot project data will be compatible with the USAF’s Weather Virtual Private Cloud. \\nDefense Department Customers\\nThe latest contract option expands Muon’s customer base within the Defense Department. It also demonstrates the versatility of Muon’s software-defined instrument, Dyer said.\\n“We’ll be delivering three different products using essentially the same instrument in orbit,” Dyer said. \\xa0\\nMuon will supply the Air Force with global soil moisture and sea surface wind measurements in addition to ionospheric data for the Space Force.\\nFirst Satellite Launch\\nMuon’s first satellite launched in June on a SpaceX Falcon 9 rideshare flight is “outperforming our expectations,” Dyer said. “It’s been a fantastic first platform for us to have in space, and it is going to be a valuable asset as we continue to upgrade the software onboard.”\\nMuSat-2, is scheduled to launch in February. While MuSat-2 shares the same spacecraft bus as MuSat-1, it will be equipped with Muon’s first software-defined sensor.\\n“Muon Space is honored by the Air Force, Space Force and DIU’s belief in our capabilities to bring new insights to DoD weather and ionospheric models with new data to include soil moisture, ocean winds and total electron content.\\xa0 We’re excited to showcase the operational relevance of this commercial space dataset to the DoD,” Dyer said in a statement.\\n'}\n", 658 | "{'date': datetime.date(2023, 7, 11), 'title': 'Thermal imagery sector heats up', 'author': 'Debra Werner', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/rsz_solar.jpg?resize=800%2C600&ssl=1', 'text': 'British investment firm Seraphim Space surveyed the Earth observation sector a few years ago, categorizing startups by sensor type.\\nAnalysts determined that firms focused on electro-optical and synthetic aperture radar were advancing rapidly. Thermal imagery startups were not.\\n“Much to our surprise, it was the only sensor area that didn’t have any companies that had really progressed,” said Seraphim Space CEO Mark Boggett. “None of the companies had raised $10 million, let alone $100 million or $500 million like some of the other sensor areas.”\\nAll that is changing. Startups focused on gathering thermal imagery via satellite are attracting investment, making acquisitions and winning contracts.\\n“It’s the next big thing in Earth observation,” said Anthony Baker, founder and CEO of Satellite Vu, a British Earth-observation startup. “No one has opened up the frontier on infrared.”\\nThe new businesses focused on thermal imagery vary widely. What the founders share is the conviction that startups can provide the type of data only expensive government satellites supplied in the past.\\nUrgent Climate Action\\nClimate change is propelling much of the work. Venture capital firms are flocking to startups promising to track emissions or mitigate the impact of droughts, floods and forest fires.\\n“There’s an urgency to do something,” said Max Gulde, CEO and co-founder of German startup constellr. “Having a thermal picture of our planet at an actionable resolution and frequency is something which is missing in our understanding of climate change. Suddenly, there’s a push for that.”\\nThermal imagery startups also are benefiting from recent declines in launch costs and technological advances.\\n“When I started founding the company, everyone was explaining to methat the sensor we wanted to build was impossible,” said Thomas Grübler, CEO and co-founder of Munich-based OroraTech. “Now, it’s possible to shrink these big complex systems to smallsat and cubesat scale.”\\nWhat’s more, founders have identified government and commercial customers willing to pay for thermal imagery. Farmers are buying data that help them irrigate crops without wasting water. And fire departments are eager for access to satellite images that reduce the need for dangerous aerial flights over wildfires.\\n“It’s getting cheaper to build a specialist constellation that can perform certain tasks extremely well,” Gulde said. “Suddenly, you’re passing a threshold where people are willing to pay.”\\nThe European Commission and the European Space Agency awarded contracts in June to constellr, OroraTech and Spanish startup Aistech Space.\\nThe companies will supply thermal data to complement observations collected through the European Union Copernicus Earth-observation program. \\nData Fusion\\nWhen Albedo, a Colorado startup, was founded in 2020, the business plan focused on collecting 10-centimeter-resolution optical imagery from telescopes in very low Earth orbit. The founders soon realized they could obtain longwave infrared imagery from the same satellites without much additional cost.\\n“With the optical resolution, we can go from counting cars to identifying cars,” said Albedo founder and CEO Topher Haddad. “With thermal, you can see where a car probably just pulled out from a parking spot.”\\nPlus, the combination of optical and thermal imagery helps observers distinguish hot tubs from pools or trampolines, and backyard dwelling units from sheds.\\nAlbedo has raised $58 million for refrigerator-size satellites scheduled to begin launching in 2025. \\nWork on Albedo’s infrared technology is being funded by the U.S. Air Force National Air and Space Intelligence Center under a $1.25 million contract. Under another $1.25 million contract, the Air Force is working with Albedo to look for ways to integrate Albedo imagery tasking with government systems.\\nConstellr’s celestr Land Surface Imaging product shows temperatures in Melbourne, Australia, an urban heat island. Credit: constellr Credit: constellr\\nEcosphere Health\\nConstellr is preparing to launch satellites in 2024 to gather thermal data for agricultural and environmental-monitoring applications.\\nSince the company was founded in 2020, constellr has raised about $14 million in venture capital and received an additional $14 million in grants. \\nEarlier this year, constellr acquired ScanWorld, a Belgian hyperspectral satellite imagery and analytics startup. With four shoebox-size satellites equipped with infrared sensors, constellr plans to gather daily imagery of agricultural fields around the world. By adding hyperspectral data, constellr can help farmers identify crop disease and manage fertilization schedules.\\nUnder the five-year, $5 million Copernicus contract announced in June, constellr will provide thermal imagery to thousands of European institutions.\\n“We’ve been working with the European Space Agency for quite some time,” Gulde said. “Government support has been exceptional.”\\nHydrosat, a startup building a constellation of thermal infrared imaging satellites, shows how its sensor highlights wildfire hot spots in contrast to visible imagery, which can be obscured by smoke. Credit: Hydrosat Credit: Hydrosat\\nWater Scarcity\\nWashington-based Hydrosat has raised $35.6 million to obtain thermal data from space. Supporting sustainable agriculture and helping customers reduce carbon emissions are the company’s primary goals.\\nThermal data can pick up initial signs of drought two to four weeks before optical imagery shows a change in the color of vegetation, said Pieter Fossel, Hydrosat CEO and co-founder.\\nIn June, Hydrosat acquired IrriWatch, a Netherlands company that delivers daily climate, crop, soil and irrigation updates to farmers.\\n“With IrriWatch, we can point to examples of how customers usingthis product, based on thermal satellite insights, are able to reduce water use,” Fossel said. “In a lot of parts of the world, reducing water use means less electricity for pumping that water out of the ground, less electricity for operating that mechanized center pivot irrigation system. And if the irrigation system is being run on diesel fuel, that’s a direct, measurable carbon reduction for the farm in addition to improvements in production, the boosting of yields and the reduction of water.”\\nHydrosat, founded in 2017, plans to launch its first two satellites next year to “deliver thermal infrared data as well as multispectral infrared data with higher resolution and greater revisit than what is available today,” Fossel said. “As a climate-oriented business, being able to have very tangible, measurable impacts is something that’s really important to us.”\\nOroraTech obtained thermal imagery of fires in Alberta, Canada, in June 2023. Credit: OroraTech Credit: OroraTech\\nFire Danger\\nAfter being told thermal imagery could not possibly be captured with cubesats, OroraTech’s founders raised $22.4 million to prove naysayers wrong. The German startup demonstrated its first uncooled thermalinfrared sensor on a Spire Global cubesat in 2022.\\n“People told me I wouldn’t see anything with this camera,” Grübler said. “We see fires at comparable quality to the Visible Infrared Imaging Sensor. We also see temperature quite well.”\\nOroraTech plans to collect global imagery every 30 minutes with a constellation of 96 satellites. The first eight satellites, built and operated by Spire Global under an agreement announced June 28, are slated to launch into a late-afternoon sun-synchronous orbit in 2024.\\n“We want to understand the trends in temperature, whether it’s for urban heat, industrial activity or forest fires,” Grübler said. “We want to be the first one to not only know where the forest fire is but the intensity of the forest fire and how it will behave.”\\nHigh-resolution image of the Satellite Vu Leeds test flight. Satellite Vu gathers data on heat usage, leakage and management from an airborne sensor to test its technology prior to launching thermal-imaging satellites. Credit: SatVu Credit: Satellite Vu\\n\\nData Sharing\\nSatVu was founded in 2016 to gather high-resolution thermal imagery with 160-kilogram satellites. After raising 12.7 British pounds ($16 million) for satellites designed and manufactured with Surrey Satellite Technologies Ltd., SatVu launched its first spacecraft in June.\\nA second satellite is slated to launch in 2024. SatVu’s eight-satellite constellation should be operating in a couple of years, Baker said.\\nTo date, 66 companies have committed 128 million pounds to SatVu’s Early Access Programme. SatVu customers who sign up for the Early Access Programme can task SatVu’s airborne sensor to collect thermal imagery and obtain discounts on future satellite-tasking orders.\\nDefense and intelligence agencies may be SatVu’s first customers “because they already know what the data looks like,” Baker said. “They just don’t have a commercial data source they can share with allies.”\\nAnother potential application is industrial activity monitoring.\\n“With infrared, you can see activity in a building,” Baker said. “If it’s dormant, it probably has no heat signature. If it’s an active factory, you can see which parts of the factory are operating.”\\nFrom left to right: SuperSharp chief science officer Ian Parry, chief technology officer George Hawker, and CEO Marco Gomez-Jenkins, with Satlantis CEO Juan Tomas Hernani and business development manager Ignacio Mares. Credit: SuperSharp Credit: Satlantis\\nFoldable Telescopes\\nEarlier this year, Spain’s Satlantis acquired a majority stake in SuperSharp, a British university spin-out developing unfolding space telescopes to obtain thermal infrared imagery.\\nUK government agencies have provided initial funding for SuperSharp. \\nThe next milestone is to get to a space-rated version of our telescope, which we plan to do by the end of this year,” said Marco Gomez-Jenkins, SuperSharp co-founder and CEO. “After that, we’re focusing on testing our telescope in space by early 2025.”\\nSatlantis, meanwhile, sells a variety of Earth observation payloads and manufactures satellites like Armenia’s Armsat-1, launched in 2022. \\nOnce SuperSharp’s infrared telescope is flight-proven, the company will begin selling foldable telescopes for microwave-size cubesats and developing a larger version of the telescope for higher-resolution imagery.\\nGuardian Satellites\\nOne of the first startups to launch a thermal infrared imaging satellite was Aistech. The Spanish startup’s first cubesat equipped with a multi-spectral telescope to collect visible, near-infrared and thermal infraredimagery reached orbit in 2022.\\n“Temperature is a unique data source that is key to understanding activity on the Earth’s surface and the changes that can occur,” Aistech founder Carles Franquesa told SpaceNews by email. “And for this, Aistech’s strategic plan involves deploying its space infrastructure to monitor these changes continuously and accurately.”\\nAistech is preparing to launch 20 Guardian thermal-imagery satellites into a constellation, scheduled to be completed in 2027, with applications including water management, forestry, environmental monitoring and maritime security. Two additional Guardians are set to launch in the second quarter of 2024. \\nIn June, Aistech was named a Copernicus Contribution Mission contract.\\n“Becoming part of the select group of companies that provide data generated by their own constellation of satellites through the Copernicus Contributing Missions programme marks an important milestone for the company, since a reference client such as ESA validates that both the technology developed and the data generated by Aistech satellites will provide a new vision of the changes that occur on the Earth’s surface and a value to society facing new challenges on the planet,” Franquesa said. “ESA, through this program, is carrying out important work in promoting and developing new applications based on geospatial intelligence, which will allow a new generation of companies to provide new\\xa0solutions to specific problems; and this entails an exponential growth in the need for new satellite\\xa0data, and therefore a great development of the space thermal imaging market in the coming years.”\\nThis article originally appeared in the July 2023 issue of SpaceNews magazine.\\n'}\n", 659 | "{'date': datetime.date(2023, 7, 10), 'title': 'NASA awards “crossover” spacesuit task orders to Axiom and Collins', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/03/axiom-suit.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — NASA awarded task orders to two companies already working on spacesuits for the International Space Station and Artemis missions to develop alternative versions of their suits.\\nNASA announced July 10 that it issued task orders valued at $5 million each to Axiom Space and Collins Aerospace to begin design work on alternative versions of their suits already in development. Axiom’s task order begins work on a version of its suit for the ISS while Collins will begin design of a suit intended for moonwalks.\\nNASA awarded contracts to the two companies in June 2022 through its Exploration Extravehicular Activity Services program to support development of new Artemis and ISS spacesuits. NASA would then acquire spacesuit services rather than the suits themselves, effectively renting them versus owning them.\\nThe contracts, though, required the companies to compete for specific task orders for spacesuit development. NASA awarded one task order to Axiom Space in September 2022 to develop an Artemis spacesuit, valued at $228.5 million. It awarded another to Collins Aerospace in December 2022 for an ISS spacesuit, valued at $97.2 million.\\nThe companies will use the new “crossover” task orders to adapt the suits they are developing for one application to the other. NASA said in a statement that, after completing initial design work, the agency will then consider exercising options for further suit development.\\nDoing so, the agency said, provides redundancy by having a backup suit for both the ISS and Artemis missions. “Using this competitive approach we will enhance redundancy, expand future capabilities, and further invest in the space economy,” said Lara Kearney, manager of the Extravehicular Activity and Human Surface Mobility Program at the Johnson Space Center.\\nThe awards may also more closely align with the companies’ plans. Axiom, which won the original task order for lunar spacesuits, is also developing a commercial space station that may require spacewalks either for maintenance or to serve customer requirements. Collins, which won the ISS suit task order last December, had earlier emphasized its work on lunar spacesuit designs.\\n“We are excited to add our orbital spacesuits as an option for NASA,” Mark Greeley, EVA program manager at Axiom Space, said in a company statement. The company said work on a low Earth orbit version of its spacesuit is already underway.\\n“Our next-generation spacesuit design is nearly 90% compatible with a lunar mission,” said Dave Romero, director of EVA and human space mobility systems at Collins, in a company statement. “This formal contract award will support continued efforts to modify our next-generation spacesuit, making it suitable to tasks on the moon.”\\nAxiom said that the full value of this new task order, if all options are exercised, is $142 million over four years. Collins did not disclose the full value of its task order.\\n'}\n", 660 | "{'date': datetime.date(2023, 7, 10), 'title': 'Astra to raise up to $65 million in stock sale', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/05/astra-rocket4.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — Faced with dwindling cash and a stock delisting, Astra Space announced plans July 10 to perform a reverse split of its stock and sell up to $65 million of it.\\nIn a filing with the U.S. Securities and Exchange Commission published after the markets closed, Astra said it had signed a sales agreement with Roth Capital Partners under which it will sell up to $65 million of its stock in an “at-the-market” offering, where shares are sold at the going market rate.\\nNet proceeds from the stock sale, the company said, would go towards working capital and general corporate purposes. That includes development of its next-generation launch vehicle, Rocket 4, as well as continued production of its Astra Spacecraft Engine electric thrusters.\\nThe stock sale comes as the company was running low on cash. Astra reported having $62.7 million in cash as of the end of the first quarter, with a net loss of $44.9 million. The company reported no revenue in the first quarter.\\nAstra executives said on an earnings call May 15 they projected ending the second quarter with $30 million to $33 million of cash remaining. The company was considering at the time both debt and equity options to raise additional cash to keep the company operating as it continues work on Rocket 4, which the company does not anticipate entering commercial service until some time in 2024.\\nAstra also faced a potential delisting of its stock from the Nasdaq because its price had fallen below $1 per share. Astra was originally given 180 days by Nasdaq, starting in October 2022, to get its share price above $1 for at least 10 consecutive business days. The company won a 180-day extension on April 10.\\nAstra said July 10 that its board had approved a plan for a 1-for-15 reverse stock split that will occur no later than Oct. 2. Under that plan, 15 shares of existing Astra stock will be exchanged for one new share of Astra stock. That will have the effect of boosting the share price but not the company’s overall market cap. Shareholders had approved a proposal at the company’s annual shareholder meeting June 8 to allow the board to enact a reverse split of between 1-for-5 and 1-for-15.\\nShares in Astra closed July 10 at 40.2 cents per share, which would be $6.03 after the reverse split. Those shares have fallen more than 97% in value from their peak in early July 2021, shortly after Astra completed its merger with Holicity, a special-purpose acquisition company, and started trading publicly on Nasdaq.\\n'}\n", 661 | "{'date': datetime.date(2023, 7, 10), 'title': 'Voyager Space deepens India ties for commercial space station plans', 'author': 'Jason Rainbow', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/03/starlab-voyagerairbus.jpg?resize=800%2C600&ssl=1', 'text': 'TAMPA, Fla. — Voyager Space is considering using India’s proposed Gaganyaan crewed spacecraft to serve the commercial space station it aims to be operating by the end of the decade.\\nThe Denver-based space technology provider announced a Memorandum of Understanding (MoU) July 10 with India to explore using Gaganyaan, among other potential collaborations to deepen ties with the country’s space industry.\\xa0\\nThe MoU paves the way to other partnerships across exploration, research, and commercial activities, Voyager chief revenue officer Clay Mowry said.\\nIndia expects to perform Gaganyaan’s first crewed flight no earlier than 2025 following delays that have pushed out its schedule by at least three years.\\nThe MoU with India is Voyager’s first with a crewed spacecraft provider outside the United States, Mowry told SpaceNews.\\nHe said Voyager is working with multiple undisclosed providers to supply crew and cargo services for Starlab, which would use a standard docking system aiming to be compatible with various spacecraft.\\n“We are targeting our single-launch configuration to be operational in 2028,” he added.\\nIndia partnerships\\nGaganyaan would launch to low Earth orbit on a version of India’s heavy-lift Geosynchronous Satellite Launch Vehicle Mark 3.\\nVoyager announced a separate MoU July 7 to explore launch and deployment opportunities for small satellites orbited by two smaller Indian rockets: The Small Satellite Launch Vehicle and Polar Satellite Launch Vehicle, or SSLV and PLSV.\\nVoyager’s customers have previously flown payloads on two PLSV missions, according to the company, which said the deal further expands their access to space.\\xa0\\nThe agreement also enables Voyager to study using space-qualified components from the recently created commercial arm of India’s space agency, covering spacecraft manufacturing, deployment, operations, and other areas of interest.\\nVoyager offers a broad range of space technologies following a series of acquisitions since being founded four years ago, ranging from laser and radio frequency communications systems to mission-data transmitters and cameras.\\nThe company announced its latest acquisition March 13 in a deal for engineering company ZIN Technologies, known for microgravity research equipment that Voyager said would support plans for its Starlab space station.\\nUnder development in partnership with Lockheed Martin, Starlab is one of three commercial concepts in the running to help NASA transition from the aging International Space Station.\\nIn January, Voyager said Airbus is also providing technical design support and expertise for the project, potentially making it easier for European governments to use Starlab.\\nVoyager’s partnerships in India come as the country relaxes regulatory rules over its commercial space sector and the involvement of foreign businesses.\\nDuring Indian Prime Minister Narendra Modi’s recent visit to the United States, the two countries also announced plans to create a strategic framework for human spaceflight cooperation by the end of the year. The plans include a joint effort to the International Space Station in 2024 that has yet to be detailed.\\nEfforts to galvanize India’s space industry also include plans by the Indian Space Research Organization to auction off SSLV to the private sector, reported the country’s Economic Times July 9, citing an unnamed senior official at the space agency.\\nCapable of carrying up to 500 kilograms to mid-inclination low Earth orbits, SSLV is designed to be cheaper and more flexible than its two larger Indian cousins for deploying small payloads.\\nAfter failing to reach orbit in its August debut, SSLV’s second flight successfully placed three satellites into low Earth orbit in February.\\n'}\n", 662 | "{'date': datetime.date(2023, 7, 10), 'title': 'Benchmark raises $33 million in Series B round', 'author': 'Debra Werner', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2022/11/rsz_big_propulsion_demand_is_driving_thruster_production_at_benchmark.jpg?resize=800%2C600&ssl=1', 'text': 'SAN FRANCISCO – Benchmark Space Systems raised $33 million in a Series B funding round.\\nWith the money raised, Benchmark plans to shift its focus from propulsion system research and development to manufacturing and testing. \\xa0\\xa0\\nNews of the investment round came on the heels of layoffs by the Burlington, Vermont-based startup.\\n“The fundraising process began months ago and is not directly related to our team realignment, which recently resulted in about a 15 percent reduction of our team of 118 people,” Benchmark CEO Ryan McDevitt told SpaceNews by email. “We’ve been strategizing about how to best scale to successfully execute on our next phase of development in tandem with the fundraising process, and believe we are very well positioned to meet the current and future needs and demands of commercial and government.”\\nGrowing Backlog\\nBenchmark has grown rapidly in the past couple of years due in part to its acquisition of Alameda Applied Science Corp.’s electric propulsion technologies. After acquiring the electric propulsion technology, Benchmark expanded its product line to include chemical, electric and hybrid propulsion systems.\\n“The funding provides the capital we need to deliver on a growing number of government and commercial contracts coming in for a wide range of non-toxic chemical, electric and hybrid propulsion systems,” McDevitt said. “Benchmark is quickly transitioning from research and development to a strong focus on production.”\\nAs Benchmark customers expand satellite production, Benchmark “is scaling up our assembly and testing capabilities,” McDevitt said. “Our refined strategy and capital raise will enable Benchmark to better address and meet the evolving demands and innovations required for successful commercial and government missions set to launch over the near term 12 to 24 months and longer-term projects throughout the next three to five years.”\\n'}\n", 663 | "{'date': datetime.date(2023, 7, 10), 'title': 'AI, quantum and nuclear technologies are key to Lockheed Martin’s vision for Space 2050', 'author': 'Miriam Klaczynska', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/rsz_2screenshot_2023-07-10_at_93725_am.png?resize=800%2C549&ssl=1', 'text': 'LOS ANGELES – Artificial intelligence, quantum computing and nuclear power are among the key technologies Lockheed Martin sees as important for future space missions.\\nThrough a project called Destination: Space 2050, Lockheed Martin executives are exploring, for example, how AI could assist scientific exploration of locations where communications with remote sensors would be disrupted by high latency.\\nIn that type of environment, “you really can’t interact with the robotic sensors,” David Lackner, Lockheed Martin senior manager strategy and business development, said during a June 28 webinar. “You have to have something that is super autonomous that can deal with unknown unknowns. We’ve got some really interesting causal autonomy tools that … allow the AI to be super smart about running into something that it hasn’t encountered before.”\\nAI also has important applications for remote-sensing data, said Aura Roy, Lockheed Martin deputy program manager for Multi-slit Solar Explorer mission, known as MUSE.\\nData gathered by hundreds or thousands of satellites traveling in different orbits could provide a “vast amount of information which would be beyond the ability of any number of human operators to parse through,” Roy said. “The goal is to use AI to determine truly optimal and trusted decisions from that raw data,” which may not be intuituve.\\nIn the future, “we will need to rely on a AI to augment human decision makers at all levels of command with advanced AI data processing and course-of-action generation that will support all types of operations,” Roy said.\\nQuantum Computing\\nIn addition, Lockheed Martin’s Space 2050 report focuses on quantum computing, quantum communications and quantum remote sensing, technologies.\\n“That computing infrastructure utilizing quantum will be there for us in the 2050 timeframe,” Lackner said. \\nAs a result, Lockheed Martin is developing “quantum algorithms to make use of quantum computers, quantum remote sensing and quantum communications,” Lackner said. “Specific space applications of quantum are going to be super enabling for what we want to do.”\\nNuclear Power, Crew Habitats and Mobility\\nAdvancements in power and propulsion, including nuclear technologies, “are going to be absolutely critical in terms of dramatically improving the types of missions that you can do and the types of science that you can collect,” said Kate Watts, Lockheed Martin’s vice president of Mission Strategy and Advanced Capabilities for Human and Scientific Exploration. “Think high-power generation, dramatically improved ISP related to propulsion, so you have more maneuverability” or the ability to reach distant locations more quickly.\\nBy 2050 more people will be traveling to low-Earth orbit as well as to the moon or Mars, which will create demand for habitats and vehicles to move around the lunar surface.\\n“When you’re on other worlds, you need to be able to move around, modify things, change the surface as needed,” Watts said. “You need to give the crew autonomy to manage science with machines that have capability like they do here on Earth.”\\nLunar Commercial Opportunities \\nBy 2050, “I think we’ll see regular commercial delivery services from the Earth to the moon and back for both cargo or people,” said Crescent Space CEO Joe Landon. Lockheed Martin established Crescent Space earlier this year to provide lunar communications and navigation services.\\nOn the lunar surface, crews will “be able to find, extract and process valuable resources to create fuel and to sustain life and support human operations,” Landon said.\\n“Every space mission today has to be completely self sufficient, has to bring everything you need with you and account for every contingency,” Landon said. “So, even the smallest missions or capabilities end up being very complex and costly.”\\nTo achieve Lockheed Martin’s Space 2050 vision, new infrastructure including computational capacity in cislunar space will be needed.\\nProviding access to lunar cloud computing would enable both governments and private actors to play a larger economic role in space. Instead of purchasing large assets and then having to maintain them, buyers would instead be able to buy capabilities as needed, Lackner said.\\nLockheed Martin selected 2050 as a goal for its technology report because that timeframe is long enough away “that we can truly develop disruptive technologies and disruptive capabilities,” said Nelson Pedreiro, vice president at Lockheed Martin’s Space Advanced Technology Center. \\xa0\\n'}\n", 664 | "{'date': datetime.date(2023, 7, 10), 'title': 'Can space governance keep up with space sustainability?', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/2021-Russian-ASAT-COMPSOC-volumetric.jpeg?resize=800%2C562&ssl=1', 'text': 'It is not uncommon to be faced with a problem but struggle to find a solution. In the case of space sustainability, as low Earth orbit fills with both active satellites and debris, the challenge is as much coming up with a solution to deal with that congestion as it is determining who should do it. \\nThe rapid growth in the number of space objects, caused by the rise of satellite constellations as well as debris-generating events like anti-satellite tests and collisions, is testing the international governance model for space activities developed in the early years of the Space Age that many in the industry believe can no longer keep up. \\n“More objects have been launched in the last 10 years than in the previous 50 combined,” — Guy Ryder, undersecretary-general for policy at the United Nations, speaking June 13, 2023 at the Secure World Foundation’s Summit for Space Sustainability. Credit: Marcel Crozet/ILO\\n“More objects have been launched in the last 10 years than in the previous 50 combined,” noted Guy Ryder, undersecretary-general for policy at the United Nations, during a speech at the Summit for Space Sustainability by the Secure World Foundation (SWF) in New York June 13. That creates, he said, “boundless development opportunities and governance needs.” \\nSpace governance has been handled at the international level by the U.N.’s Committee on the Peaceful Uses of Outer Space (COPUOS), which now has 102 nations as members. In the 1960s, COPUOS helped guide the development of the Outer Space Treaty, the foundation of international space law, followed by several related agreements. \\nHowever, the large size of COPUOS and its use of a consensus-based model — all nations must agree — can make progress slow or nonexistent. Attendees of the most recent COPUOS meeting in Vienna, which concluded in early June, noted that one agenda item known as “dark and quiet skies” to study the effects of satellite constellations on astronomy had widespread support but was dropped because of objections from one nation, Iran, which is opposed to constellations in general on sovereignty grounds. \\n“It’s slow, it’s frustrating,” said Valda Vikmanis Keller, director of the U.S. State Department’s Office of Space Affairs, of COPUOS, during one panel at the Summit for Sustainability. But, she said, the discussions there were essential. “It’s the only way forward.” \\nHowever, some are looking for alternative mechanisms to address those growing space sustainability concerns. That includes efforts both within the United Nations itself and among other governments and organizations, seeking binding agreements or simply widely adopted norms and guidelines as the population of satellites and debris continues to grow. \\nWindow of opportunity \\nOne such effort is at the United Nations. It is beginning preparations for its Summit of the Future, a two-day meeting in New York in September 2024 where member nations will discuss key global issues. The U.N. calls it a “once-in-a-generation opportunity to enhance cooperation on critical challenges and address gaps in global governance.” \\nThe agenda of that meeting will include space. “We have a window of opportunity over the next 15 months,” Ryder said, “where we can accelerate space diplomacy and advance the governance issue.” \\nThe U.N. started planning for the Summit of the Future with the release of a policy paper in May on outer space governance. It highlighted as key topics space traffic coordination as well as space resource utilization and concerns about conflict in outer space. \\nThe report doesn’t offer specific solutions, but Ryder said the ultimate goal is to develop “one single united regime to facilitate data sharing, cooperation and continuity” in space traffic management and related issues. Those issues could be addressed individually, though, “if that path looks likelier to achieve results.” \\nCOPUOS will play a role in developing concepts to be considered at the Summit of the Future through its meetings next year, but it will not be the only mechanism. Portugal will host a conference in the spring of 2024 to develop proposals for consideration at the fall summit. \\nHugo André Costa, a member of the executive board of the Portuguese Space Agency, said at the SWF summit that the spring meeting will be preceded by two virtual workshops, one this October on technology issues and a second in March 2024 on policy issues. Those meetings will be open to representatives from industry, academia and governments. “This is the only way that we can prepare for the future,” he argued. \\nDeveloping a global space government framework in just 15 months is an ambitious task. If it is successful, though, Ryder suggested that the framework that emerges might look different from existing approaches. One model, he said, might be existing treaties like the U.N. Convention on the Law of the Sea. “All of this provides us with the confidence that the kinds of agreements concluded in the past are possible in the future, even in today’s admittedly challenging geopolitical climate.” \\n“The notion is that we live in a rapidly changing world,” he said, looking for new approaches while maintaining existing mechanisms. “Let’s be honest: what worked yesterday will not necessarily work tomorrow. That is not a recipe or a reason to jettison everything that has worked up until this point.” \\n“But,” he added, “it can be a very strong reason to adapt, modify, improve what we have.” \\nPhone a friend \\nThe international community has demonstrated it can move quickly on space sustainability issues. After Russia conducted an ASAT demonstration in November 2021, destroying a defunct Russian satellite and creating thousands of pieces of debris, the United States and others started pushing for mechanisms to halt any future such tests. \\nU.S. officials said they didn’t expect Russia to perform a destructive direct-ascent (DA) ASAT test. “We were shocked. I’ll be honest with you, I was shocked,” said Audrey Schaffer, director of space policy at the National Security Council, during a talk at the Summit for Space Sustainability. “How could they do something so brazen, so reckless, and so clearly contrary to the safety, sustainability and security of an environment that so many of us depend upon?” \\nAt left, Lt. Gen. John Shaw, deputy commander of U.S. Space Command. Credit: U.S. Space Command\\nLater at the summit, Lt. Gen. John Shaw, deputy commander of U.S. Space Command, said that the U.S. military suspected Russia was planning some sort of test. “They consider themselves the senior spacefaring nation,” he said, and thus expected Russia would do an “offset test” that fired a missile to deliberately miss the target. “They want to continue that tradition of being that senior, responsible spacefaring nation. I was wrong.” \\nThe Defense Department, he said, became “one of the earliest and biggest proponents” of what Vice President Kamala Harris announced the following April, that the U.S. would refrain from conducting similar ASAT tests and request other countries to do the same. \\nThat effort resulted in a U.N. General Assembly vote last December, just over a year after the Russian ASAT test, on a resolution regarding such a test ban. A total of 155 nations voted in favor of the resolution, while nine, including China and Russia, voted against it. Nine other nations, including India, abstained. \\n“That kind of vote count indicates a very strong base of support,” Schaffer said but argued it was not enough. The resolution, she noted, was non-binding, simply encouraging countries to make such commitments. As of June 2023, 13 countries have done so. \\n“To truly establish an internationally recognized norm banning destructive DAASAT missile testing, we need a critical mass of nations to actually make the commitment,” she said. \\nThat has been a slow process. Only three countries — Austria, Italy and the Netherlands — have formally committed not to conduct destructive ASAT tests since the U.N. vote. That slow progress, some suggested, may be due to domestic politics. \\nA photo exhibition called drawing attention to the fragility of Earth’s near-space environment was on display during last month’s meeting in Vienna of the U.N.’s Committee on the Peaceful Uses of Outer Space. Credit: UNIS Vienna via Flickr\\n“We understand that other states that voted for the resolution but have not yet joined the commitment need some time to thoroughly review the domestic effects,” said Hyerin Kim, second secretary in the disarmament and non-proliferation division of South Korea’s Ministry of Foreign Affairs. South Korea announced its commitment to refrain from such tests last October. “Korea is also making efforts to raise awareness of the danger posed by ASAT testing.” \\n“We had a great start out of the gate, but there’s more road ahead of us,” Schaffer said, urging conference attendees representing nations that had not made a commitment to consider doing so. And for those who have, “phone a friend and ask them to make the commitment too.” \\nAlternative approaches \\nOther efforts to address space sustainability are proliferating. At the Summit for Space Sustainability, the World Economic Forum (WEF) announced a new set of guidelines for mitigating the growth of orbital debris. The document is the latest foray into space sustainability for the WEF, which previously led the development of a Space Sustainability Rating. \\nAmong the recommendations published by the WEF is to establish a success rate for “post-mission disposal,” or removal of satellites from orbit after the end of their missions, of 95% to 99%. That disposal, it added, should be done no later than five years after the end of a satellite’s life, versus earlier guidelines of 25 years. \\n“We wanted to push the envelope a little bit on some of these concrete, specific targets,” said Nikolai Khlystov, lead for the WEF’s Future of Space initiative, at the conference. The document also calls for satellites to be maneuverable, preferably through onboard propulsion, when operating at altitudes above 375 kilometers, and for operators to agree to share orbital data. \\nThe WEF convinced 27 companies to endorse the guidelines, including several major satellite operators. Notably absent, though, were some planning or deploying large constellations, such as Amazon and SpaceX. \\nKhlystov said the WEF worked with more than the 27 companies who signed the new guidelines. “If some actors didn’t sign on, I don’t think it’s a sign that they are against these standards,” he said. (An Amazon spokesperson later said that while the company helped craft those guidelines, it was not ready yet to endorse them as it assesses various other proposals.) \\nESA Director General Joseph Aschbacher, third from left, is shown discussing the Zero Debris charter during last month’s Paris Air Show with executives from Airbus Defence and Space, OHB SE and Thales Alenia Space. Credit: ESA\\nJust over a week later, the European Space Agency and three European satellite manufacturers — Airbus Defence and Space, OHB and Thales Alenia Space — announced their intent to develop a “Zero Debris Charter” to mitigate the growth of orbital debris. The charter, announced during the Paris Air Show, doesn’t exist yet beyond the most general outlines. \\n“The principle is a very simple one,” ESA Director General Josef Aschbacher said at the announcement. “The Zero Debris Charter is a principle where we would like to ensure that there is zero debris left behind in space.” He compared it a national park, where visitors are expected not to leave behind any garbage. \\nThe goal is that, under the charter, satellite operators would be expected by 2030 to either deorbit their satellites on their own at the end of their lives or hire a company that offers active debris removal services to do so. ESA and the companies plan to refine the details and complete the text of the charter by the end of the year. \\nWhat benefit the charter will have was not exactly clear at the announcement. All three companies participating have already endorsed the WEF’s guidelines, with some executives at the event praising that document for its specific requirements. \\nAschbacher suggested the charter could be incorporated into regulations so that governments agree to work only with those companies that agree to follow it. “We need to achieve a status where we demand that only data or information is bought from those satellite providers who are adhering to certain standards,” Aschbacher said. “The charter may be one vehicle for doing so.” \\nExecutives, though, were cautious about a solution to a worldwide problem that applied only to European companies. “It’s important that those regulations are worldwide,” said Hervé Derrey, chief executive of Thales Alenia Space. “If this is not applied to the rest of the world it will have no effect at the end. It will collectively fail. And, on top of that, European industry will not be on a level playing field with its competitors. That would be the worst situation.” \\n“Only with a real international regulation will we get it under control,” said Lutz Bertling, an OHB board member. \\nThat, then, returns to the United Nations and its often slow efforts. Ryder, at the Summit for Space Sustainability, was careful to praise COPUOS for its work even while suggesting changes are in order. “We have a very solid record of achievement,” he said. “It’s a very solid platform from which to start.” \\n“It’s very difficult to have 102 different ways of seeing the same thing,” Portugal’s Costa said, describing the long discussions at COPUOS meetings over things like wording of a single sentence in a statement all will accept. “You just need to have nerves of steel to wait until the very last moment when all the agreements are reached.” \\nAs low Earth orbit fills with satellites and debris, that very last moment for space sustainability may be rapidly approaching.\\n\\nThis article originally appeared in the July 2023 issue of SpaceNews magazine.\\n'}\n", 665 | "{'date': datetime.date(2023, 7, 7), 'title': 'Plasmos pivots from rocket engines to VC', 'author': 'Debra Werner', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/e734dad0-1b72-47eb-bfe2-4177ccf63e24.jpeg?resize=800%2C528&ssl=1', 'text': 'SAN FRANCISCO – Plasmos, the Los Angeles-based startup developing rocket engines, is pivoting to become an artificial intelligence-driven venture capital firm.\\nPlasmos officially changed its business model in late May after struggling to attract investors and strategic partners due in part to the background of Plasmos CEO Ali Baghchehsara. Born in Iran, Baghchehsara moved to Germany as a teenager to earn a master’s degree in aeronautical engineering. There, he worked for the German Aerospace Center DLR and Airbus, before moving to the United States in 2021 to pursue his dream of developing hybrid electric-chemical\\xa0rocket\\xa0engine.\\nWhile Plasmos could address the business challenges startups commonly face, the political challenges seemed insurmountable, Baghchehsara said.\\n“I saw where Plasmos will end,” Baghchehsara told SpaceNews. “If we are successful, I’ll have to deal with the U.S. government. And if I sell to the U.S. government it may become impossible for me to visit my family in Iran. I had to choose a path: company success or family.”\\nAI for VC\\nPlasmos has raised about $275,000 in cash and $275,000 in in-kind contributions including 3D printing services. With remaining funds, Baghchehsara is establishing PlasmOS, an AI-driven venture capital firm.\\nUnlike venture capital partners who base investment decisions on their knowledge, experience or instinct, PlasmOS will use AI to select promising startups.\\n“We will use data to make almost instantaneous decisions, a quick yes or no on funding,” Baghchehsara said.\\nAmbitious Plan\\nAs a space startup, Plasmos was developing a Space Truck, powered by the company’s dual-mode propulsion system. The Space Truck was designed to transport payloads in Earth orbit and to support in-space manufacturing, last-mile delivery, point-to-point transportation, on-orbit servicing and active debris removal.\\n'}\n", 666 | "{'date': datetime.date(2023, 7, 7), 'title': 'Merger rumors swirl around Dish and EchoStar', 'author': 'Jason Rainbow', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/03/rsz_jupiter_3_-_launch_configuration.jpg?resize=800%2C540&ssl=1', 'text': 'TAMPA, Fla. — Satellite TV broadcaster Dish Network is rumored to be considering recombining with internet-focused sister company EchoStar to strengthen its financial resources.\\nThe companies have engaged advisers to flesh out a potential deal, reported news publication Semafor July 6, citing people familiar with the matter.\\nDish and satellite fleet operator EchoStar — both controlled by billionaire Charlie Ergen — declined to comment on the speculation.\\nA deal would need to navigate starkly contrasting financial standings following their split into separate companies and stocks back in 2008.\\nDish has been investing heavily to meet regulatory deployment deadlines for expanding a terrestrial 5G network across the United States, putting its balance sheet under strain as its core satellite TV business bleeds subscribers.\\nMeanwhile, EchoStar is sitting on $1.7 billion in cash and is poised for subscriber and revenue growth from Jupiter 3, its long-awaited broadband satellite that Maxar recently delivered for a Falcon Heavy launch in the coming months.\\nEchoStar sold underperforming broadcast satellite services assets to Dish in 2019, seemingly doubling down on the strategy to keep their businesses apart.\\nStill, rumors about a potential recombination have periodically resurfaced in the market over the years, according to Raymond James analyst Ric Prentiss.\\nNotably, EchoStar CEO Hamid Akhavan’s Feb. 17, 2022, employment offer letter includes a clause covering the possibility of Dish owning more than half of EchoStar’s voting stock.\\nPrentiss said he was unsurprised to see the speculation resurface again because of how challenging it is to borrow money to finance Dish’s wireless plans in the current economic climate.\\nA combination could also put perennial rumors to rest about another attempt to merge Dish with its satellite broadcast rival DirecTV, majority owned by U.S. telecoms giant AT&T.\\nEarlier this year, Dish was also said to be one of a handful of U.S. companies in talks about selling wireless services through Amazon, according to reports including Bloomberg and the Wall Street Journal.\\nCiting people familiar with the situation, Bloomberg reported June 2 that Amazon was looking into offering a low-cost or free nationwide mobile service to Amazon Prime subscribers.\\nAmazon, which is working toward providing initial broadband services next year from its proposed Project Kuiper satellite constellation, said in response that it was not currently planning to add wireless service to its Prime offering.\\nIts package of Amazon Prime services includes online TV streaming, which has contributed to the decline of the satellite broadcast market.\\n'}\n", 667 | "{'date': datetime.date(2023, 7, 7), 'title': 'Interest grows for human spaceflight in Europe', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/marcuswandt.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — As the European Space Agency continues to develop proposals for human space exploration efforts, more European countries are showing an interest in launching astronauts.\\nAt a June 29 briefing after a meeting of the ESA Council, ESA Director General Josef Aschbacher announced that Poland was subscribing an additional 295 million euros ($325 million) for agency programs, a figure that includes the cost of flying a Polish astronaut to the International Space Station on a commercial mission.\\nThat agreement, he said, was similar to one announced in April that involved ESA, the Swedish National Space Agency and Axiom Space that will fund the flight of a Swedish astronaut on an Axiom commercial flight to the ISS. ESA announced in June that Marcus Wandt, a Swedish Air Force pilot selected by ESA as a reserve astronaut in November 2022, would go on that mission.\\nESA has not announced when Wandt will fly. Anna Rathsman, director general of the Swedish National Space Agency and outgoing chair of the ESA Council, said at the briefing that the flight had to be coordinated among the various organizations involved, including NASA.\\n“Sweden has been able to make very quickly decisions that allowed this flight to happen,” Aschbacher said. “In less than two months we have gone, I would say, from zero to signature of the agreements.”\\nWandt, he said, is already in Houston undergoing training for the flight. The earliest opportunity would be the Ax-3 mission, slated to fly to the ISS in late 2023 or early 2024. Axiom Space has not announced the crew for that mission.\\nOther European countries have expressed an interest in flying short-duration missions to the ISS. Walter Villadei, an Italian Air Force pilot who flew on Virgin Galactic’s first commercial suborbital flight June 29, also has trained with Axiom Space and was a backup on the Ax-2 mission to the station in May. Hungary’s foreign minister said at the ESA ministerial meeting in November 2022 that the country was spending $100 million on its own private astronaut to fly with Axiom.\\nThat activity comes as ESA is working on proposals for human spaceflight programs of its own. A report in March by an independent High-Level Advisory Group recommended ESA pursue the ability to launch its own astronauts, developing crewed spacecraft and potentially commercial space stations and lunar missions. In response, Aschbacher said in April that the agency was working on proposals to implement those recommendations.\\nIn a July 1 interview after the launch of ESA’s Euclid space telescope on a Falcon 9 from Cape Canaveral, Florida, Aschbacher said the agency is continuing work on those proposals through the summer in preparation for a Space Summit of European nations in Seville, Spain, in November.\\nThose efforts included a discussion during a side meeting of last week’s ESA Council meeting. “The discussion was very constructive, very positive, and there’s a lot of work that is being done right now by our teams, with industry and with the member states to establish scenarios and options of what Europe could be doing and should be doing, and what the price tag is attached to it,” he said.\\nBy September or October, he expected to have “more clarity” on the various options, which he said currently are too premature to discuss publicly. He did note that one requirement is that the proposals not take funding from other ESA programs. “That’s why I’m calling for the Space Summit because this engages the top leaders in Europe to really discuss what Europe should be doing, because these decisions need to be made there.”\\nThe interest by several ESA member states in private astronaut missions is a positive sign for the agency’s plans for human space exploration, he argued. “Sweden, Poland and many others are now inspired by this ambition to go to space and have an astronaut flying into space,” he said. “This momentum starts developing, and I can only say it’s nice to see.”\\nAschbacher added that he didn’t think that such short-term missions, funded in Sweden’s case as part of a public-private partnership, would detract from efforts to craft a long-term human spaceflight program at the agency. “They have certainly raised the appetite and the attention, but you cannot compare a 10-day mission with a six-month mission on the space station,” he said. “But it’s a new element that we can add to our portfolio.”\\n'}\n", 668 | "{'date': datetime.date(2023, 7, 7), 'title': 'Chinese launch firm secures fresh funding for reusable rocket', 'author': 'Andrew Jones', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/Tianlong2-launch-2april2023-Space-pioneer-2.jpg?resize=800%2C600&ssl=1', 'text': 'HELSINKI — Chinese rocket firm Space Pioneer has secured C-round funding for its Tianlong-3 medium-lift reusable launch vehicle.\\xa0\\nSpace Pioneer, full name Beijing Tianbing Technology Co., Ltd, announced the funding round July 5. The company’s first launch attempt in April saw the Tianlong-2 rocket make Space Pioneer the first Chinese commercial outfit to reach orbit with a liquid propellant launcher.\\nThe company has now raised a total of three billion Chinese yuan ($414 million) across 11 rounds since its establishment in 2018. In February the firm announced it had secured “B+” and “Pre-C” strategic funding rounds.\\nA number of the leading investors are linked to the Chinese state. These include China International Capital Corporation (CICC), a partially state-owned investment vehicle, CCB International, belonging to the China Construction Bank Corporation, and CITIC Construction, the engineering and construction arm of Chinese state-owned CITIC Group, and Zhejiang University Lianchuang. Venture capital also plays a part in the funding.\\nThe new funds will be used for the development of the Tianlong-3 kerosene-liquid oxygen launch vehicle and its engine and the construction of a dedicated launch complex at the Jiuquan national spaceport. The funds will also go towards developing mass production capabilities and recruitment.\\xa0\\nThe first launch of the Tianlong-3 rocket is currently scheduled for May 2024. It is intended to be somewhat comparable to the Falcon 9. If successful, Space Pioneer claims it will be able to launch 30 times per year from 2025.\\nTianlong-3 (“Sky Dragon-3”) is a two-stage kerosene-liquid oxygen rocket with a reusable first stage. Space Pioneer’s webpages state that the rocket will be capable of lifting 17 tons of payload to low Earth orbit, or 14 tons to 500-kilometer sun-synchronous orbit.\\xa0\\n\\xa0It will have a takeoff mass of 590 tons and produce 770 tons of thrust. The 71-meter-long Tianlong-3 will have a diameter of 3.8 meters. Previously, many commercial Chinese rockets used 2.25 or 3.35-meter-diameter stages, the dimensions as those of most Long March rockets.\\xa0\\nThe Tianlong-3 is, according to Space Pioneer, tailor-made for launching satellites for China’s national communications megaconstellation. Tianlong-3 will provide “important strategic support for the country’s new satellite Internet infrastructure,” according to the July 5 statement. The planned 13,000-strong Guowang constellation is seen as China’s answer to Starlink.\\nMore recently established Chinese commercial launch companies including Space Pioneer and OrienSpace are opting to develop much larger launch vehicles than earlier movers, which have committed to launchers with payload capacities on the order of a few thousand kilograms. This shift has been initiated by the emergence of the possibility of contracts for Guowang and commercial cargo missions.\\nSpace Pioneer says it has already signed 10 orders for launches on the Tianlong-2.\\nThe company also has plans for a Tianlong-3H, a triple-core version in the same style as the SpaceX Falcon Heavy, and the Tianlong-3M, a single core rocket tipped with a reusable spaceplane.\\nSpace Pioneer initially started out developing engines burning green monopropellant before changing direction. The firm also apparently scrapped development of the Tianlong-1 rocket.\\nThe firm has received sponsorship and investment from Zhangjiagang, a city in Jiangsu province on the Yangtze river. Space Pioneer is building rocket manufacturing facilities in the city and the Tianlong-2 launch also bore the name “Zhangjiagang.”\\nMeanwhile, fellow Chinese commercial firm Landspace is gearing up for the second launch of its Zhuque-2 methane-liquid oxygen rocket from Jiuquan. Airspace closure notices indicate a launch window of 1:53 a.m. to 4:14 a.m. Eastern July 12.\\n\\nMore: pic.twitter.com/AG2OeQSFGN— Cosmic Penguin (@Cosmic_Penguin) July 7, 2023\\n\\nBeijing-based iSpace also recently tested engines for its Hyperbola-2 methalox launcher.\\n'}\n", 669 | "{'date': datetime.date(2023, 7, 6), 'title': 'Europe leans on SpaceX to bridge launcher gap', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/f9-euclid-prelaunch1.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — Europe, temporarily lacking its own access to space, plans to rely more on SpaceX to launch key science and navigation spacecraft while working to restore its launch capabilities.\\nThe successful final Ariane 5 launch July 5 means that Europe temporarily has no ability to launch payloads into orbit. The Ariane 5’s successor, Ariane 6, is still in development and appears increasingly unlikely to be ready for its inaugural launch before 2024. The Soyuz is no longer available in Europe after Russia’s invasion of Ukraine in February 2022. The Vega C remains grounded after a December 2022 launch failure, and its return to flight, previously planned for late this year, is facing delays after an anomaly during a static-fire test June 28 of that rocket’s Zefiro 40 motor.\\nThe original version of the Vega, which does not use the Zefiro 40, is scheduled to resume launches in September, but most launches on the Vega manifest are of the more powerful Vega C. European startups such as Isar Aerospace and Rocket Factory Augsburg are working on small launch vehicles whose first flights could take place before the end of the year. But for larger payloads, there are few near-term options for European organizations.\\nThat situation led the European Space Agency to announce in October 2022 it was moving two missions to SpaceX’s Falcon 9. The first of those, the Euclid astronomy mission, was set to launch on Soyuz but instead lifted off on a Falcon 9 July 1 from Cape Canaveral, Florida. The Hera asteroid mission will launch on a Falcon 9 in October 2024 after being originally manifested on Ariane 6.\\nMore European missions are likely to fly on Falcon 9. At a June 29 briefing after a meeting of the ESA Council, the agency announced that the Earth Clouds, Aerosols and Radiation Explorer, or EarthCARE, mission that has been moved from Soyuz to Vega C last October would instead likely fly on Falcon 9 in the second quarter of 2024.\\nWhile the announcement coincided with the news of the Zefiro 40 test anomaly, ESA Director General Josef Aschbacher said the change was not directly linked to that but instead to the failed launch last December as well as changes in the size of EarthCARE that would have required modifications to the Vega C payload fairing to accommodate it.\\n“With these two elements, I asked my inspector general to reassess the situation,” he said at the briefing. “The conclusion of this assessment by the inspector general was recommending to me that we should not fly with Vega C.” That was based on the earlier Vega C failure review, which recommended no changes to the configuration, like the payload fairing, on the initial series of Vega C launches.\\nThe Falcon 9 payload fairing with the logos of ESA and the Euclid spacecraft. Credit: SpaceNews/Jeff Foust\\nAt the same briefing, ESA officials said they were also in discussions with SpaceX for the launch of up to four Galileo satellites on Falcon 9 vehicles. “We are moving ahead with negotiations to conclude hopefully soon with SpaceX,” said Javier Benedicto, ESA’s director of navigation. That is contingent on concluding the negotiations with SpaceX as well as securing approvals from the European Union and its security agreement with SpaceX.\\nIn a July 1 interview after the Euclid launch, Aschbacher said it will be up to the European Commission to decide when and how to launch those Galileo satellites. “We have provided all the technical information with regards to launcher compatibility, which the Commission has,” he said. “Now it’s up to them to make a decision.”\\nHe noted that Euclid was not the first time that ESA had used the Falcon 9. Sentinel-6A, also known as Sentinel-6 Michael Freilich, was a joint U.S.-Europe Earth science mission launched on a Falcon 9 in November 2020. ESA astronauts have also flown on Crew Dragon missions. However, in those cases the launches were procured and overseen by NASA, not ESA.\\nAschbacher praised SpaceX for its role in launching Euclid. “SpaceX has been very proactive, very quick, very professional in providing this launch service. And I’m very happy now that this has been conducted successfully.”\\nHe reiterated past comments that the problems with Ariane 6 and Vega C and the loss of Soyuz had created a “launcher crisis” in Europe. “We are in a crisis and we should use the opportunity to convert this crisis into actions and changes that need to be adopted in order, in the future, to develop a robust launcher system for Europe,” he said, looking beyond Ariane 6 and Vega C.\\nHe also took the long view, expecting that this temporary gap in European launch capabilities will be forgotten in the long term, once Ariane 6 and Vega C are in regular operations. “Then this few months will be a blip,” he said. “Of course, now we are right in the midst of it and it’s difficult for everyone to accept the current situation. But you have to see this in the longer term.”\\n'}\n", 670 | "{'date': datetime.date(2023, 7, 6), 'title': 'Viasat signs deal to commercialize European airspace tracking service', 'author': 'Jason Rainbow', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/ESSP.jpg?resize=800%2C593&ssl=1', 'text': 'TAMPA, Fla. — A group founded by European air traffic controllers has signed a deal to bring improved airspace-tracking capabilities from Viasat’s constellation to market next year.\\nThe European Satellite Services Provider group, or ESSP, said July 6 it will be responsible for leading the commercialization of Iris, an air traffic modernization program the European Space Agency developed with Viasat’s recently acquired satellite operator Inmarsat.\\nIris — not to be confused with Europe’s proposed IRIS² connectivity constellation — promises to help aircraft fly more efficient routes by using Inmarsat’s L-band satellites to complement congested VHF data links.\\nAccording to ESA, fitting aircraft with higher bandwidth Iris technology would give air traffic controllers more data to schedule landings in advance, minimize fuel consumption, and maximize airspace and airport capacity.\\nCommunications between pilots and controllers could also move from voice to text messages under the upgraded air traffic management (ATM) system, improving operational safety and efficiency.\\nIris forms a part of the Single European Sky’s Air Traffic Management Research master plan proposed in 2020 to create more environmentally sustainable and efficient flight paths.\\nFrance-based ESSP was initially set up in 2001 so their air navigation shareholders could participate in the European Geostationary Navigation Overlay Service program, which Europe uses to augment and improve GPS services in the region.\\nESSP currently comprises the air navigation service providers of France, Portugal, Italy, Switzerland, Germany, Spain, and the United Kingdom.\\nU.K.-based easyJet and ITA Airways of Italy are among the first airlines intending to operate early Iris services next year, according to ESSP.\\n'}\n", 671 | "{'date': datetime.date(2023, 7, 6), 'title': 'China’s Landspace set for second methalox rocket launch', 'author': 'Andrew Jones', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2022/12/zhuque-2-Landspace-jiuquan-2022-2.jpg?resize=800%2C600&ssl=1', 'text': 'HELSINKI — Chinese commercial launch firm Landspace is set for a second attempt to reach orbit with its Zhuque-2 rocket July 12.\\nThe second Zhuque-2 methane-liquid oxygen rocket was rolled out to the pad at Jiuquan Satellite Launch Center in the Gobi Desert July 6, according to Chinese social media posts.\\xa0\\nSatellite imagery later confirmed the development. Earlier Chinese media reports indicate the launch is scheduled for July 12 Beijing time.\\nIf successful, Zhuque-2 (“Vermillion Bird-2”) will become the world’s first methalox launch vehicle to achieve orbit. A range of methalox rockets, including SpaceX’s Starship, the ULA Vulcan, Blue Origin’s New Glenn, Rocket Lab’s Neutron and Terran R from Relativity Space, are at various stages of development and testing.\\n\\nSatellite imagery confirms that the second Zhuque-2 launch vehicle rolled out to Launch Complex 96 earlier today ahead of its upcoming launch attempt currently scheduled for July 12. pic.twitter.com/iIATKtbN1J— Harry Stranger (@Harry__Stranger) July 6, 2023\\n\\nA successful Zhuque-2 flight would also make Landspace the second private Chinese rocket company to perform a successful launch with a liquid propellant rocket. Space Pioneer, full name Beijing Tianbing Technology Co., Ltd, became the first such company in April with the launch of its Tianlong-2 rocket.\\nThe upcoming launch comes almost seven months after the debut flight of the Zhuque-2, also from Jiuquan. That December 2022 launch ended in failure due to an issue with a liquid oxygen inlet pipe feeding four vernier thrusters on the rocket’s second stage.\\nZhuque-2 is powered by gas generator engines producing 268 tons of thrust. It is capable of delivering a 6,000-kilogram payload capacity to a 200-kilometer low Earth orbit (LEO), or 4,000 kilograms to 500-kilometer sun-synchronous orbit (SSO), according to Landspace.\\nThe rocket has a diameter of 3.35 meters—the same as a number of national Long March rockets—a total length of 49.5 meters and a take-off mass of 219 tons.\\xa0\\nThe rocket is currently expendable but Landspace is working on a s tested a restartable version of the 80-ton-thrust TQ-12 engine which powers the Zhuque-2 first stage.\\xa0\\nThe firm is also working on an improved second-stage engine which will not require vernier engines. This upgraded engine will not be present on the second Zhuque-2.\\nLandspace is one of the earliest and best-funded of China’s emerging commercial launch firms. The company’s first launch took place four years ago with the much smaller and simpler solid-propellant Zhuque-1 and ended in failure.\\xa0\\nThe Chinese government opened up parts of the space sector to private capital in late 2014, seen to be a reaction to developments in the U.S.\\xa0\\nThis policy shift and subsequent policy support and guidance has been the catalyst for the emergence of hundreds of space-related companies engaged in a range of activities, including launch, satellite operation and manufacture, ground stations, downstream applications, and more.\\nA number of commercial launch companies are now looking to China’s “Guowang” national satellite internet project as a potential source of contracts and revenue.\\nChina recently opened a call for space station commercial cargo proposals, further indicating that commercial firms will have a growing role to play in the country’s space sector.\\nLandspace and other early movers in China’s commercial launch sector such as iSpace and Galactic Energy committed to plans to develop light-lift solid and liquid propellant rockets.\\xa0\\nNewer entrants such as Space Pioneer and Orienspace are moving directly towards medium-lift and heavier classes of launchers, likely informed by the emergence of the possible satellite internet and programs.\\nA range of remote sensing constellations are also being designed and planned by a range of state and commercial actors, providing further opportunities for launch contracts.The first launch of China’s communications megaconstellation satellites is expected in the second half of the year, using a Long March rocket developed by China’s state-owned main space contractor, CASC.\\n'}\n", 672 | "{'date': datetime.date(2023, 7, 6), 'title': 'Space Command argues for shift from static to dynamic satellite operations', 'author': 'Sandra Erwin', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/Screenshot-2023-07-06-at-4.15.21-PM.png?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — In order to better keep tabs on adversaries, the U.S. military needs satellites that can actively maneuver in orbit, the deputy commander of U.S. Space Command said July 6.\\n“The way we’ve been doing space operations since the dawn of the space age, we’ve been doing it wrong,” Lt. Gen. John Shaw said at a Mitchell Institute for Aerospace Studies event.\\xa0\\nShaw specifically alluded to the military’s “neighborhood watch” satellites — known as GSSAP (Geosynchronous Space Situational Awareness Program) — used to monitor the geostationary belt, where the Pentagon deploys its most valuable space assets.\\xa0\\nGSSAPs were built to last for decades but have a limited fuel supply, and maneuvers must be carefully planned to minimize fuel consumption.\\xa0\\nShaw has raised the issue at several public appearances in recent months. At the Mitchell forum, he once again insisted that the current approach of deploying satellites in designated orbits and minimizing maneuvers does not work in the age of great power competition.\\nHe said Space Command is advocating for a new approach he described as “dynamic space operations,” in contrast to the current practice of minimizing maneuvers for fear of depleting a satellite’s fuel supply.\\xa0\\n“This could be the most fundamental doctrinal shift that we’re probably going to see in the next four to five years,” he said.\\xa0\\nNeed maneuverable platforms\\nPerhaps not all satellites have to be highly maneuverable, Shaw said. But surveillance and reconnaissance platforms need to be more mobile in order to better monitor adversaries’ activities.\\xa0\\n“We launch a platform into orbit, and we tend to leave it right in that orbit,” Shaw said. “But we’re coming to the realization that that is not going to be sufficient anymore.”\\nShaw said he recently spent time with GSSAP operators at Schriever Space Force Base, Colorado. Their feedback reinforced the need for change in how these assets are developed and deployed, he added.\\n“GSSAP’s sole purpose is to move around the geosynchronous belt and to look at other platforms … And if there’s something that’s behaving suspiciously, we’ll look at that,” Shaw said.\\xa0\\nU.S. military satellites are operated to ensure they stay in service for many years or decades. “And so we are actually largely constrained by what we can do with those platforms,” he said. “I had a chance to sit down with operators a few weeks ago and spend a good amount of time understanding how they approach it.”\\nThey have to operate with significant constraints, said Shaw. “We can’t have those constraints in the future. And so we’re trying to articulate a requirement to the Space Force that we need to be able to have sustained space maneuver for those platforms.”\\nPotential solutions\\nIt will be up to the Space Force to come up with solutions to this problem, Shaw said. Among the options being considered is to build satellites with refueling ports so they can take advantage of commercial refueling services.\\xa0\\n“If I could refuel GSSAP satellites once a month, we would be operating them completely differently than we do now. They’d be operating at maximum trust levels and delta V levels that are unlike anything we’re doing right now,” Shaw said.\\nAnother possible solution is to rely on low-cost small satellites that would be replaced more frequently. “I’ll get another one in two weeks because I’m going to fly the heck out of it. And it’s going to empty that gas tank in a hurry.”\\nThe goal is to “achieve surprise and initiative against an adversary in ways that I can’t today,” Shaw added.\\xa0\\nThe Space Force is working to address these needs, he said.\\xa0\\n“We’ve asked for a demo by 2026 on how we would do sustained space maneuvers for a given platform.”\\n'}\n", 673 | "{'date': datetime.date(2023, 7, 6), 'title': 'HawkEye 360 satellites to monitor illegal fishing in Pacific Islands', 'author': 'Sandra Erwin', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/Cluster7_beautyShot-scaled.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — \\xa0HawkEye 360, a commercial operator of remote-sensing satellites, announced July 6 it was selected by the government of Australia to help detect illegal fishing activity using radio-frequency sensors.\\nThe company received a contract of undisclosed value from Australia’s Department of Foreign Affairs and Trade for a pilot program in support of the Pacific Islands Forum Fisheries Agency.\\nThe agency, based in Honiara, Solomon Islands, was established in 1979\\xa0 to help its 17 member countries manage their tuna fisheries and track illicit activities.\\nHawkEye 360, headquartered in Herndon, Virginia, uses radio-frequency data analytics to geolocate electronic emissions and draw insights.\\xa0\\nRadio-frequency identification\\nThe company operates a constellation of 21 satellites that detect, characterize and geolocate radio frequency signals from emitters used for communication, navigation and security.\\xa0\\nHawkEye 360’s satellites fly in triangular formations in low Earth orbit. As a cluster passes over an area, each satellite observes signal waveforms and downlinks the data to a cloud system on the ground where it’s analyzed.\\xa0\\nUnder the contract with Australia, HawkEye 360 will provide satellite RF maritime analytics and training through 2023, said Alex Fox, chief growth officer of HawkEye 360.\\n“The Pacific Islands encompass a vast and highly trafficked region with rich fisheries resources that present complex challenges for maritime domain awareness,” he said.\\xa0\\nThe Pacific Islands Forum Fisheries Agency and its members will get data, analytics services and training support to identify illicit maritime activity within their waters, Fox said. By analyzing RF data, he added, HawkEye 360 can track maritime activity that is not detectable by Automatic Identification Systems (AIS).\\n'}\n", 674 | "{'date': datetime.date(2023, 7, 5), 'title': 'Ariane 5 launches for the final time', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/ariane5-lastlaunch2.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — One chapter in European access to space came to a close July 5 with the final launch of the Ariane 5, but the beginning of the next chapter faces additional delays.\\nAn Ariane 5 lifted off from the European spaceport at Kourou, French Guiana, at 6 p.m. Eastern. The launch had been scheduled for June 16 but was postponed a day in advance after Arianespace concluded that three pyrotechnical transmission lines used for the separation of the rocket’s solid rocket boosters needed to be replaced. The company rescheduled the launch for July 4, only to delay it an additional day because of strong upper-level winds.\\nAs with so many Ariane 5 missions, this launch, designated VA261, carried two communications satellites destined for geostationary transfer orbit. Nearly 30 minutes after liftoff, the rocket deployed Heinrich-Hertz-Satellit, a spacecraft built by OHB for the German Space Agency on behalf of other German government agencies. The 3,400-kilogram satellite will test advanced communications technologies.\\nAbout three and a half minutes later, the rocket deployed the other payload, the Syracuse 4B satellite for the French military. The 3,570-kilogram satellite was developed by a consortium of Airbus Defence and Space and Thales Alenia Space, using an Airbus Eurostar 3000 bus.\\n“It is a success for ‘Team Europe’ tonight with this last and final Ariane 5,” Stéphane Israël, chief executive of Arianespace, said on the company’s webcast of the launch after confirmation of successful payload deployment.\\n“Thanks for ArianeGroup, Arianespace and CNES. It was a wonderful launch, even if it is the last one,” said Gen. Michel Sayegh, director of space programs for the French armaments agency DGA during the launch webcast.\\nThe launch was the 117th and final flight of the Ariane 5 over 27 years. The vehicle made its first, unsuccessful launch in June 1996, and suffered a partial failure on its second launch in October 1997 before an unqualified success on its third launch in October 1998. The rocket’s ability to carry two large geostationary communications satellites at once made it a key vehicle for many years in the commercial space industry during an era when geostationary communications satellites dominated the market.\\nThe European Space Agency also regularly used the rocket for several science missions as well as the launch of five Automated Transfer Vehicle cargo spacecraft to the International Space Station between 2008 and 2014. In perhaps the rocket’s highest-profile launch, it successfully launched the James Webb Space Telescope for NASA on Christmas Day 2021, delivering it on a trajectory so accurate it had the effect of significantly increasing the spacecraft’s lifetime by reducing the amount of propellant needed for trajectory correction maneuvers.\\n“Ariane 5 is now over, and Ariane 5 has perfectly finished its work and really is now a legendary launcher,” Israël said. “But Ariane 6 is coming.”\\nWaiting on Ariane 6\\nIn the lead up to the final Ariane 5 launch, Arianespace has advertised a “spaceflight continuum,” of past and future rockets, but that continuum is not necessarily continuous. The Ariane 5 overlapped with the end of the Ariane 4 rocket, which made its last launch in 2003. ESA had originally planned for a similar overlap between the end of the Ariane 5 and the introduction of its successor, the Ariane 6.\\nHowever, the development of Ariane 6 has been plagued by delays that have pushed out its first launch, once planned for 2020, by several years. In October 2022, ESA said it projected the first launch to take place in the fourth quarter of 2023, but it is increasingly likely the launch will slip into 2024.\\nExecutives with OHB, a supplier on the Ariane 6 program, said in an earnings call in May that they expected the first Ariane 6 launch to take place in early 2024, and no later than May 2024. “I am getting more and more confident we will see the first launch of Ariane 6 early next year,” Marco Fuchs, chief executive of OHB, said during the call.\\nESA and Arianespace have declined to provide an updated launch date for that inaugural mission. “Today it would be speculative to mention a launch date,” ESA Director General Josef Aschbacher said during a press briefing June 29 after an ESA Council meeting in Stockholm. “We have to go through a number of technical milestones over the summer period but I promise, after the summer in September, we will indicate a period which is the target period for the Ariane 6.”\\nThose milestones include a hot-fire test of the Ariane 6 upper stage scheduled for July at a test facility in Lampoldshausen, Germany, which will be followed by a second test in the fall to test its performance in what ESA calls “degraded cases.” Assembly of the first flight model of the Ariane 6 is planned to begin in November in French Guiana, according to an ESA update published June 8.\\n'}\n", 675 | "{'date': datetime.date(2023, 7, 5), 'title': 'Rivada gets more breathing room to deploy constellation', 'author': 'Jason Rainbow', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2022/04/global_network_connection_covering_earth_with_lines_innovative_perception_copy_2_34-1536x952-1.jpeg?resize=800%2C600&ssl=1', 'text': 'TAMPA, Fla. — International regulators have waived a requirement for Rivada Space Networks to launch 10% of its proposed 576 satellites by September, boosting plans to fund the multibillion-dollar connectivity constellation.\\nRivada expects to start deploying commercial satellites in 2025 under contracts with manufacturer Terran Orbital and launcher SpaceX, easily missing the first deployment deadline under International Telecommunication Union (ITU) rules.\\nHowever, Rivada announced July 5 that the ITU granted a waiver for this milestone after reviewing evidence of the funding, manufacturing, and launch contracts in place for its non-geostationary orbit network (NGSO).\\nAmid a flood of applications for new NGSO constellations, the ITU adopted deployment milestone rules in late 2019 to help separate those truly planning to deploy satellites from others more interested in hoarding spectrum.\\nRivada’s waiver is a positive sign for others hoping for leniency as pandemic-related supply chain issues drag on the space industry, such as Canada’s Telesat, which continues to seek funds for an NGSO constellation running six years behind schedule.\\xa0\\n“A shortage of launch capacity and delays in technology development have been significant challenges to overcome in order to deploy these constellations in the stipulated timeframe,” said Rainer Schnepfleitner, director of Liechtenstein’s telecoms regulator, which licensed Rivada’s constellation.\\nTwo major deployment hurdles remain\\xa0\\nRivada has two spectrum filings lodged with the ITU, an affiliate of the United Nations, each covering 288 satellites.\\nThe venture must deploy 50% of its total constellation, or 288 satellites, by mid-2026 under the second ITU deployment milestone that remains in place, or risk losing priority Ka-band spectrum. All 576 satellites must be deployed by mid-2028.\\nRivada head of corporate communications Brian Carney said the venture is also on track to launch four precursor satellites next year, which would test user terminals and networking protocols and not form part of the operational constellation.\\nHe said negotiations with debt providers are ongoing to fully fund the constellation, which includes a $2.4 billion contract announced in February with Florida-headquartered Terran Orbital to build the 500-kilogram satellites.\\nDeclan Ganley, CEO of U.S. wireless technology company Rivada Networks, the parent of Germany-based Rivada Space Networks, recently said the venture hopes to secure support from U.S. Ex-Im Bank, the country’s export credit agency.\\nWhile the ITU required Rivada to show it has sufficient funding commitments for the constellation, it did not require the plans to be fully funded with money in the bank.\\n'}\n", 676 | "{'date': datetime.date(2023, 7, 5), 'title': 'Radio noise from satellite constellations could interfere with astronomers', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/rfi-radioastronomy.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — Large satellite constellations can unintentionally generate electromagnetic noise, creating an additional source of interference for radio astronomers.\\nAstronomers announced July 5 that they detected radio emissions at relatively low frequencies from dozens of SpaceX Starlink satellites as they passed over a Dutch radio observatory. The emissions, at frequencies between 110 and 188 megahertz, are different from the deliberate transmissions from the satellites used to provide broadband internet access between 10.7 and 12.7 gigahertz.\\nUsing the Low Frequency Array (LOFAR) radio telescope, astronomers detected emissions from 47 of 68 Starlink satellites they monitored. They detected several narrowband emissions at specific frequencies in that range, primarily from Starlink satellites in their operational orbits rather than those still in the process of raising their orbits after launch. They also detected broadband emissions over the entire range as well as one signal at 143.05 megahertz that was likely reflections from the French GRAVES space surveillance radar.\\nThe emissions, researchers said, likely come from electromagnetic interference from subsystems within the spacecraft. There are no international regulations regarding such emissions from spacecraft, in contrast to rules for terrestrial equipment.\\n“This study represents the latest effort to better understand satellite constellations’ impact on radio astronomy,” said Federico Di Vruno, co-director of the International Astronomical Union’s Center for the Protection of the Dark and Quiet Sky from Satellite Constellation Interference and lead author of the study being published in the journal Astronomy & Astrophysics. Astronomers had theorized that such emissions could be detected, he noted in a statement. “Our observations confirm it is measurable.”\\nIt’s unclear what effect the electromagnetic interference emissions from Starlink satellites would have on radio astronomy, but astronomers noted that it spans one frequency range between 150.05 and 153 megahertz that is protected for radio astronomy by the International Telecommunication Union.\\nThe emissions do not violate any regulations, and astronomers noted that SpaceX has been willing to talk about ways to mitigate any interference. That includes design changes already made to its next generation of Starlink satellites that can reduce those emissions.\\nHowever, astronomers said that similar emissions are likely from other satellite constellations, creating additional interference. “Our simulations show that the larger the constellation, the more important this effect becomes as the radiation from all satellites adds up,” said Benjamin Winkel of the Max-Planck-Institute for Radio Astronomy (MPIfR), a co-author on the study, in a statement.\\n“This makes us not only worried about the existing constellations but even more about the planned ones, and also about the absence of clear regulation that protects the radio astronomy bands from unintended radiation,” he said.\\nAstronomers have raised concerns for several years about the potential deleterious impact of satellite constellations on astronomy. That has included both the deliberate radio transmissions from such satellites as well as reflected sunlight that can interfere with optical observations.\\nIn January, the U.S. National Science Foundation (NSF), which funds several optical and radio astronomy observatories, announced a coordination agreement with SpaceX to mitigate the impacts of Starlink satellites on astronomy. That features efforts to reduce the brightness of Starlink satellites to least magnitude 7 as well as agreeing not to transmit while passing over major radio observatories.\\nDuring a town hall session at a meeting of the American Astronomical Society June 5, NSF officials said a similar coordination agreement had been reached with OneWeb, with more details to be released in the near future.\\nAstronomers involved in this study said they hope the detection of the electromagnetic interference emissions from the Starlink satellites will encourage other operators to take steps to mitigate similar emissions from their satellites.\\n“We believe that the early recognition of this situation gives astronomy and large constellation operators an opportunity to work together on technical mitigations pro-actively, in parallel to the necessary discussions to develop suitable regulations,” said Gyula Józsa of MPIfR and Rhodes University in South Africa.\\n'}\n", 677 | "{'date': datetime.date(2023, 7, 5), 'title': 'Regulatory uncertainty as commercial human spaceflight takes off', 'author': 'Jeff Foust', 'thumbnail': 'https://i0.wp.com/spacenews.com/wp-content/uploads/2023/07/galactic01-takeoff2.jpg?resize=800%2C600&ssl=1', 'text': 'WASHINGTON — As two companies prepare to begin or resume commercial suborbital human spaceflights, they are facing uncertainty about how the safety of the people on those flights will be regulated.\\nVirgin Galactic conducted its first commercial flight of its VSS Unity SpaceShipTwo vehicle June 29, flying three Italian payload specialists on a research mission designated Galactic 01. The company plans to begin monthly flights of private astronauts on that vehicle as soon as early August.\\nIt joins Blue Origin, which started flying paying customers on its New Shepard suborbital vehicle in 2021. New Shepard has been grounded after an engine problem on a September 2022 payload-only flight, although Bob Smith, chief executive of Blue Origin, said at a June 6 conference that the company would be ready to resume flights “within the next few weeks.” The company has not provided any further updates on flight plans.\\nBoth companies operate under a regulatory regime by the Federal Aviation Administration that focuses on the safety of the uninvolved public. Federal law sharply restricts the FAA’s ability to enact regulations for the safety of spaceflight participants on commercial vehicles. This restriction, called a “moratorium” by some in the field and a “learning period” by others, limits the FAA’s ability to enact such regulations to accidents that caused death or injury, or had the potential to cause death or injury.\\nThat restriction was included in the Commercial Space Launch Amendments Act of 2004 and, at the time, was to last eight years, giving industry time to build up flight experience upon which regulations could be based. It has been extended several times and is now currently set to expire at the end of September.\\nMany in industry have been lobbying for another extension, arguing that companies are only now beginning to fly vehicles on a routine basis, allowing them to develop that experience the restriction was designed to foster.\\n“The issue of the learning period is, should the government be limited to only regulate if there is evidence requiring regulation or should they be allowed to regulate prospectively without data, without any specific reason to regulate?” said Jim Muncy of PoliSpace during a panel discussion on the topic by the Beyond Earth Institute in May.\\nA report by the RAND Corporation in April, though, recommended that the current restrictions be allowed to expire this year. It concluded that doing so would allow the FAA and industry to move forward on developing regulations in gradual, coordinated process.\\n“It doesn’t mean we’re recommending a large stockpile of regulations immediately. In fact, it’s just the opposite,” said Bruce McClintock, senior policy researcher at RAND, during the Beyond Earth webinar. “We would broadly call this expanding the learning period to include more tools and resources.”\\nIt’s unclear if the current restriction can be extended. The House and Senate have been working on their versions of reauthorization legislation for the FAA, but neither currently includes any language about the learning period. The House Science Committee is working on a separate commercial space bill that could address the issue. However, any bill faces long odds of passage before Oct. 1.\\n“We have a divided Congress, so the ability to move an extension through may be a bit challenging this year,” said Caryn Schenewerk, president of CS Consulting who previously worked on regulatory issues for Relativity Space and SpaceX, during the webinar. Failure to achieve an extension, she said, may be less of an active decision by Congress than a side effect of broader debates between the Republican-led House and Democratic-led Senate.\\nShould the restriction expire on Oct. 1, FAA officials have said they do not have a set of regulations ready to be enacted, but are looking at ways of cooperating with industry to shorten the development time for spaceflight participant safety rules.\\nMike Moses, president of spaceline missions and safety at Virgin Galactic, said in an interview after the Galactic 01 flight he has been encouraged about discussions his company, along with Blue Origin and SpaceX, have had with the FAA on how regulations could be developed.\\n“The idea would be to take the data we’ve learned and use that to craft very specific, focused development areas,” he said, emphasizing the need for performance-based regulations given the different technical approaches those three companies are taking for human spaceflight. “One set of regulations just won’t apply to all.”\\nHe also played down any impact on the industry, or the debate on regulations, from the June 18 accident of a commercial deep-sea submersible, Titan, built and operated by OceanGate. That accident, which killed the five people on board, raised scrutiny about the level of regulation of that industry and parallels with commercial spaceflight, given both are forms of adventure tourism with similar clientele; one of the people killed on Titan, Hamish Harding, flew on Blue Origin’s New Shepard in 2022.\\nMoses said any comparison of that accident with commercial spaceflight is an “apples-to-oranges” one given that there is existing regulation of commercial spacecraft to protect the uninvolved public. “It certainly drives accountability. You’re not totally unsupervised,” he said. “It’s a very different thing in comparison to OceanGate.”\\n'}\n" 678 | ] 679 | } 680 | ], 681 | "source": [ 682 | "for d in data:\n", 683 | " # if d['date'] != crawl_to:\n", 684 | " # break\n", 685 | " print(d)\n", 686 | " with open(f\"{d['title']}.txt\", \"w\") as file:\n", 687 | " file.write(d['title'] + \"\\n\")\n", 688 | " file.write(d['text'])\n" 689 | ] 690 | } 691 | ], 692 | "metadata": { 693 | "colab": { 694 | "provenance": [] 695 | }, 696 | "kernelspec": { 697 | "display_name": "Python 3", 698 | "name": "python3" 699 | }, 700 | "language_info": { 701 | "codemirror_mode": { 702 | "name": "ipython", 703 | "version": 3 704 | }, 705 | "file_extension": ".py", 706 | "mimetype": "text/x-python", 707 | "name": "python", 708 | "nbconvert_exporter": "python", 709 | "pygments_lexer": "ipython3", 710 | "version": "3.8.10" 711 | } 712 | }, 713 | "nbformat": 4, 714 | "nbformat_minor": 0 715 | } 716 | -------------------------------------------------------------------------------- /custom_module/crawl_papers_with_code.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "metadata": { 7 | "id": "sTeE737yB2ij" 8 | }, 9 | "outputs": [], 10 | "source": [ 11 | "from bs4 import BeautifulSoup\n", 12 | "from urllib.request import Request, urlopen\n", 13 | "from datetime import datetime\n", 14 | "import pandas as pd\n", 15 | "import requests\n", 16 | "import re\n", 17 | "import os\n", 18 | "\n", 19 | "keyword = 'satellite'\n", 20 | "crawl_base_url = f\"https://paperswithcode.com/search?q_meta=&q_type=&q={keyword}\"\n", 21 | "\n", 22 | "crawl_from = 'July 5, 2023'\n", 23 | "crawl_to = 'July 12, 2023'" 24 | ] 25 | }, 26 | { 27 | "cell_type": "code", 28 | "execution_count": 2, 29 | "metadata": { 30 | "id": "7kDa_tXqOXsE" 31 | }, 32 | "outputs": [], 33 | "source": [ 34 | "os.mkdir('pdf')" 35 | ] 36 | }, 37 | { 38 | "cell_type": "code", 39 | "execution_count": 3, 40 | "metadata": { 41 | "colab": { 42 | "base_uri": "https://localhost:8080/" 43 | }, 44 | "id": "H4v3ZL98CvMu", 45 | "outputId": "c1d67312-8703-4fa1-b30d-96da2cfee7c2" 46 | }, 47 | "outputs": [ 48 | { 49 | "name": "stdout", 50 | "output_type": "stream", 51 | "text": [ 52 | "Crawling https://paperswithcode.com/paper/deepmask-an-algorithm-for-cloud-and-cloud\n", 53 | "Crawling https://paperswithcode.com/paper/self-attention-for-raw-optical-satellite-time\n", 54 | "Crawling https://paperswithcode.com/paper/you-only-look-twice-rapid-multi-scale-object\n", 55 | "Crawling https://paperswithcode.com/paper/efficient-statistical-classification-of\n", 56 | "Crawling https://paperswithcode.com/paper/deepglobe-2018-a-challenge-to-parse-the-earth\n", 57 | "Crawling https://paperswithcode.com/paper/highres-net-recursive-fusion-for-multi-frame\n", 58 | "Crawling https://paperswithcode.com/paper/past-ai-physical-layer-authentication-of\n", 59 | "Crawling https://paperswithcode.com/paper/semantic-stereo-for-incidental-satellite\n", 60 | "Crawling https://paperswithcode.com/paper/sen12ms-a-curated-dataset-of-georeferenced\n", 61 | "Crawling https://paperswithcode.com/paper/xbd-a-dataset-for-assessing-building-damage\n", 62 | "Crawling https://paperswithcode.com/paper/identifying-centromeric-satellites-with-dna\n", 63 | "Crawling https://paperswithcode.com/paper/temporal-convolutional-neural-network-for-the\n" 64 | ] 65 | }, 66 | { 67 | "ename": "KeyboardInterrupt", 68 | "evalue": "", 69 | "output_type": "error", 70 | "traceback": [ 71 | "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m", 72 | "\u001b[1;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)", 73 | "\u001b[1;32mc:\\Users\\phkha\\OneDrive\\Máy tính\\crawl_papers_with_code.ipynb Cell 3\u001b[0m in \u001b[0;36m\u001b[1;34m()\u001b[0m\n\u001b[0;32m 31\u001b[0m \u001b[39mcontinue\u001b[39;00m\n\u001b[0;32m 32\u001b[0m info[\u001b[39m'\u001b[39m\u001b[39mpdf_url\u001b[39m\u001b[39m'\u001b[39m] \u001b[39m=\u001b[39m soup\u001b[39m.\u001b[39mfind(\u001b[39m'\u001b[39m\u001b[39ma\u001b[39m\u001b[39m'\u001b[39m, {\u001b[39m'\u001b[39m\u001b[39mclass\u001b[39m\u001b[39m'\u001b[39m: \u001b[39m'\u001b[39m\u001b[39mbadge badge-light\u001b[39m\u001b[39m'\u001b[39m})[\u001b[39m'\u001b[39m\u001b[39mhref\u001b[39m\u001b[39m'\u001b[39m]\n\u001b[1;32m---> 33\u001b[0m response \u001b[39m=\u001b[39m requests\u001b[39m.\u001b[39;49mget(info[\u001b[39m'\u001b[39;49m\u001b[39mpdf_url\u001b[39;49m\u001b[39m'\u001b[39;49m])\n\u001b[0;32m 34\u001b[0m \u001b[39mwith\u001b[39;00m \u001b[39mopen\u001b[39m(\u001b[39mf\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m./pdf/\u001b[39m\u001b[39m{\u001b[39;00minfo[\u001b[39m'\u001b[39m\u001b[39mpdf_url\u001b[39m\u001b[39m'\u001b[39m]\u001b[39m.\u001b[39msplit(\u001b[39m'\u001b[39m\u001b[39m/\u001b[39m\u001b[39m'\u001b[39m)[\u001b[39m-\u001b[39m\u001b[39m1\u001b[39m]\u001b[39m}\u001b[39;00m\u001b[39m\"\u001b[39m, \u001b[39m'\u001b[39m\u001b[39mwb\u001b[39m\u001b[39m'\u001b[39m) \u001b[39mas\u001b[39;00m f:\n\u001b[0;32m 35\u001b[0m f\u001b[39m.\u001b[39mwrite(response\u001b[39m.\u001b[39mcontent)\n", 74 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\api.py:73\u001b[0m, in \u001b[0;36mget\u001b[1;34m(url, params, **kwargs)\u001b[0m\n\u001b[0;32m 62\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39mget\u001b[39m(url, params\u001b[39m=\u001b[39m\u001b[39mNone\u001b[39;00m, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwargs):\n\u001b[0;32m 63\u001b[0m \u001b[39mr\u001b[39m\u001b[39m\"\"\"Sends a GET request.\u001b[39;00m\n\u001b[0;32m 64\u001b[0m \n\u001b[0;32m 65\u001b[0m \u001b[39m :param url: URL for the new :class:`Request` object.\u001b[39;00m\n\u001b[1;32m (...)\u001b[0m\n\u001b[0;32m 70\u001b[0m \u001b[39m :rtype: requests.Response\u001b[39;00m\n\u001b[0;32m 71\u001b[0m \u001b[39m \"\"\"\u001b[39;00m\n\u001b[1;32m---> 73\u001b[0m \u001b[39mreturn\u001b[39;00m request(\u001b[39m\"\u001b[39m\u001b[39mget\u001b[39m\u001b[39m\"\u001b[39m, url, params\u001b[39m=\u001b[39mparams, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwargs)\n", 75 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\api.py:59\u001b[0m, in \u001b[0;36mrequest\u001b[1;34m(method, url, **kwargs)\u001b[0m\n\u001b[0;32m 55\u001b[0m \u001b[39m# By using the 'with' statement we are sure the session is closed, thus we\u001b[39;00m\n\u001b[0;32m 56\u001b[0m \u001b[39m# avoid leaving sockets open which can trigger a ResourceWarning in some\u001b[39;00m\n\u001b[0;32m 57\u001b[0m \u001b[39m# cases, and look like a memory leak in others.\u001b[39;00m\n\u001b[0;32m 58\u001b[0m \u001b[39mwith\u001b[39;00m sessions\u001b[39m.\u001b[39mSession() \u001b[39mas\u001b[39;00m session:\n\u001b[1;32m---> 59\u001b[0m \u001b[39mreturn\u001b[39;00m session\u001b[39m.\u001b[39mrequest(method\u001b[39m=\u001b[39mmethod, url\u001b[39m=\u001b[39murl, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39mkwargs)\n", 76 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\sessions.py:587\u001b[0m, in \u001b[0;36mSession.request\u001b[1;34m(self, method, url, params, data, headers, cookies, files, auth, timeout, allow_redirects, proxies, hooks, stream, verify, cert, json)\u001b[0m\n\u001b[0;32m 582\u001b[0m send_kwargs \u001b[39m=\u001b[39m {\n\u001b[0;32m 583\u001b[0m \u001b[39m\"\u001b[39m\u001b[39mtimeout\u001b[39m\u001b[39m\"\u001b[39m: timeout,\n\u001b[0;32m 584\u001b[0m \u001b[39m\"\u001b[39m\u001b[39mallow_redirects\u001b[39m\u001b[39m\"\u001b[39m: allow_redirects,\n\u001b[0;32m 585\u001b[0m }\n\u001b[0;32m 586\u001b[0m send_kwargs\u001b[39m.\u001b[39mupdate(settings)\n\u001b[1;32m--> 587\u001b[0m resp \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39msend(prep, \u001b[39m*\u001b[39m\u001b[39m*\u001b[39msend_kwargs)\n\u001b[0;32m 589\u001b[0m \u001b[39mreturn\u001b[39;00m resp\n", 77 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\sessions.py:745\u001b[0m, in \u001b[0;36mSession.send\u001b[1;34m(self, request, **kwargs)\u001b[0m\n\u001b[0;32m 742\u001b[0m \u001b[39mpass\u001b[39;00m\n\u001b[0;32m 744\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mnot\u001b[39;00m stream:\n\u001b[1;32m--> 745\u001b[0m r\u001b[39m.\u001b[39;49mcontent\n\u001b[0;32m 747\u001b[0m \u001b[39mreturn\u001b[39;00m r\n", 78 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\models.py:899\u001b[0m, in \u001b[0;36mResponse.content\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m 897\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_content \u001b[39m=\u001b[39m \u001b[39mNone\u001b[39;00m\n\u001b[0;32m 898\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[1;32m--> 899\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_content \u001b[39m=\u001b[39m \u001b[39mb\u001b[39;49m\u001b[39m\"\u001b[39;49m\u001b[39m\"\u001b[39;49m\u001b[39m.\u001b[39;49mjoin(\u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49miter_content(CONTENT_CHUNK_SIZE)) \u001b[39mor\u001b[39;00m \u001b[39mb\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[0;32m 901\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_content_consumed \u001b[39m=\u001b[39m \u001b[39mTrue\u001b[39;00m\n\u001b[0;32m 902\u001b[0m \u001b[39m# don't need to release the connection; that's been handled by urllib3\u001b[39;00m\n\u001b[0;32m 903\u001b[0m \u001b[39m# since we exhausted the data.\u001b[39;00m\n", 79 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\requests\\models.py:816\u001b[0m, in \u001b[0;36mResponse.iter_content..generate\u001b[1;34m()\u001b[0m\n\u001b[0;32m 814\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mhasattr\u001b[39m(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mraw, \u001b[39m\"\u001b[39m\u001b[39mstream\u001b[39m\u001b[39m\"\u001b[39m):\n\u001b[0;32m 815\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[1;32m--> 816\u001b[0m \u001b[39myield from\u001b[39;00m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mraw\u001b[39m.\u001b[39mstream(chunk_size, decode_content\u001b[39m=\u001b[39m\u001b[39mTrue\u001b[39;00m)\n\u001b[0;32m 817\u001b[0m \u001b[39mexcept\u001b[39;00m ProtocolError \u001b[39mas\u001b[39;00m e:\n\u001b[0;32m 818\u001b[0m \u001b[39mraise\u001b[39;00m ChunkedEncodingError(e)\n", 80 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\urllib3\\response.py:579\u001b[0m, in \u001b[0;36mHTTPResponse.stream\u001b[1;34m(self, amt, decode_content)\u001b[0m\n\u001b[0;32m 577\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m 578\u001b[0m \u001b[39mwhile\u001b[39;00m \u001b[39mnot\u001b[39;00m is_fp_closed(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_fp):\n\u001b[1;32m--> 579\u001b[0m data \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49mread(amt\u001b[39m=\u001b[39;49mamt, decode_content\u001b[39m=\u001b[39;49mdecode_content)\n\u001b[0;32m 581\u001b[0m \u001b[39mif\u001b[39;00m data:\n\u001b[0;32m 582\u001b[0m \u001b[39myield\u001b[39;00m data\n", 81 | "File \u001b[1;32md:\\miniconda3\\lib\\site-packages\\urllib3\\response.py:522\u001b[0m, in \u001b[0;36mHTTPResponse.read\u001b[1;34m(self, amt, decode_content, cache_content)\u001b[0m\n\u001b[0;32m 520\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m 521\u001b[0m cache_content \u001b[39m=\u001b[39m \u001b[39mFalse\u001b[39;00m\n\u001b[1;32m--> 522\u001b[0m data \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_fp\u001b[39m.\u001b[39;49mread(amt) \u001b[39mif\u001b[39;00m \u001b[39mnot\u001b[39;00m fp_closed \u001b[39melse\u001b[39;00m \u001b[39mb\u001b[39m\u001b[39m\"\u001b[39m\u001b[39m\"\u001b[39m\n\u001b[0;32m 523\u001b[0m \u001b[39mif\u001b[39;00m (\n\u001b[0;32m 524\u001b[0m amt \u001b[39m!=\u001b[39m \u001b[39m0\u001b[39m \u001b[39mand\u001b[39;00m \u001b[39mnot\u001b[39;00m data\n\u001b[0;32m 525\u001b[0m ): \u001b[39m# Platform-specific: Buggy versions of Python.\u001b[39;00m\n\u001b[1;32m (...)\u001b[0m\n\u001b[0;32m 531\u001b[0m \u001b[39m# not properly close the connection in all cases. There is\u001b[39;00m\n\u001b[0;32m 532\u001b[0m \u001b[39m# no harm in redundantly calling close.\u001b[39;00m\n\u001b[0;32m 533\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_fp\u001b[39m.\u001b[39mclose()\n", 82 | "File \u001b[1;32md:\\miniconda3\\lib\\http\\client.py:455\u001b[0m, in \u001b[0;36mHTTPResponse.read\u001b[1;34m(self, amt)\u001b[0m\n\u001b[0;32m 452\u001b[0m \u001b[39mif\u001b[39;00m amt \u001b[39mis\u001b[39;00m \u001b[39mnot\u001b[39;00m \u001b[39mNone\u001b[39;00m:\n\u001b[0;32m 453\u001b[0m \u001b[39m# Amount is given, implement using readinto\u001b[39;00m\n\u001b[0;32m 454\u001b[0m b \u001b[39m=\u001b[39m \u001b[39mbytearray\u001b[39m(amt)\n\u001b[1;32m--> 455\u001b[0m n \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49mreadinto(b)\n\u001b[0;32m 456\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39mmemoryview\u001b[39m(b)[:n]\u001b[39m.\u001b[39mtobytes()\n\u001b[0;32m 457\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m 458\u001b[0m \u001b[39m# Amount is not given (unbounded read) so we must check self.length\u001b[39;00m\n\u001b[0;32m 459\u001b[0m \u001b[39m# and self.chunked\u001b[39;00m\n", 83 | "File \u001b[1;32md:\\miniconda3\\lib\\http\\client.py:499\u001b[0m, in \u001b[0;36mHTTPResponse.readinto\u001b[1;34m(self, b)\u001b[0m\n\u001b[0;32m 494\u001b[0m b \u001b[39m=\u001b[39m \u001b[39mmemoryview\u001b[39m(b)[\u001b[39m0\u001b[39m:\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mlength]\n\u001b[0;32m 496\u001b[0m \u001b[39m# we do not use _safe_read() here because this may be a .will_close\u001b[39;00m\n\u001b[0;32m 497\u001b[0m \u001b[39m# connection, and the user is reading more bytes than will be provided\u001b[39;00m\n\u001b[0;32m 498\u001b[0m \u001b[39m# (for example, reading in 1k chunks)\u001b[39;00m\n\u001b[1;32m--> 499\u001b[0m n \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49mfp\u001b[39m.\u001b[39;49mreadinto(b)\n\u001b[0;32m 500\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mnot\u001b[39;00m n \u001b[39mand\u001b[39;00m b:\n\u001b[0;32m 501\u001b[0m \u001b[39m# Ideally, we would raise IncompleteRead if the content-length\u001b[39;00m\n\u001b[0;32m 502\u001b[0m \u001b[39m# wasn't satisfied, but it might break compatibility.\u001b[39;00m\n\u001b[0;32m 503\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_close_conn()\n", 84 | "File \u001b[1;32md:\\miniconda3\\lib\\socket.py:704\u001b[0m, in \u001b[0;36mSocketIO.readinto\u001b[1;34m(self, b)\u001b[0m\n\u001b[0;32m 702\u001b[0m \u001b[39mwhile\u001b[39;00m \u001b[39mTrue\u001b[39;00m:\n\u001b[0;32m 703\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[1;32m--> 704\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_sock\u001b[39m.\u001b[39;49mrecv_into(b)\n\u001b[0;32m 705\u001b[0m \u001b[39mexcept\u001b[39;00m timeout:\n\u001b[0;32m 706\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_timeout_occurred \u001b[39m=\u001b[39m \u001b[39mTrue\u001b[39;00m\n", 85 | "File \u001b[1;32md:\\miniconda3\\lib\\ssl.py:1241\u001b[0m, in \u001b[0;36mSSLSocket.recv_into\u001b[1;34m(self, buffer, nbytes, flags)\u001b[0m\n\u001b[0;32m 1237\u001b[0m \u001b[39mif\u001b[39;00m flags \u001b[39m!=\u001b[39m \u001b[39m0\u001b[39m:\n\u001b[0;32m 1238\u001b[0m \u001b[39mraise\u001b[39;00m \u001b[39mValueError\u001b[39;00m(\n\u001b[0;32m 1239\u001b[0m \u001b[39m\"\u001b[39m\u001b[39mnon-zero flags not allowed in calls to recv_into() on \u001b[39m\u001b[39m%s\u001b[39;00m\u001b[39m\"\u001b[39m \u001b[39m%\u001b[39m\n\u001b[0;32m 1240\u001b[0m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m\u001b[39m__class__\u001b[39m)\n\u001b[1;32m-> 1241\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49mread(nbytes, buffer)\n\u001b[0;32m 1242\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m 1243\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39msuper\u001b[39m()\u001b[39m.\u001b[39mrecv_into(buffer, nbytes, flags)\n", 86 | "File \u001b[1;32md:\\miniconda3\\lib\\ssl.py:1099\u001b[0m, in \u001b[0;36mSSLSocket.read\u001b[1;34m(self, len, buffer)\u001b[0m\n\u001b[0;32m 1097\u001b[0m \u001b[39mtry\u001b[39;00m:\n\u001b[0;32m 1098\u001b[0m \u001b[39mif\u001b[39;00m buffer \u001b[39mis\u001b[39;00m \u001b[39mnot\u001b[39;00m \u001b[39mNone\u001b[39;00m:\n\u001b[1;32m-> 1099\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_sslobj\u001b[39m.\u001b[39;49mread(\u001b[39mlen\u001b[39;49m, buffer)\n\u001b[0;32m 1100\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m 1101\u001b[0m \u001b[39mreturn\u001b[39;00m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_sslobj\u001b[39m.\u001b[39mread(\u001b[39mlen\u001b[39m)\n", 87 | "\u001b[1;31mKeyboardInterrupt\u001b[0m: " 88 | ] 89 | } 90 | ], 91 | "source": [ 92 | "metadata = []\n", 93 | "# Crawl 3 trang dau\n", 94 | "for page in range(1,3):\n", 95 | " req = Request(\n", 96 | " url=f\"{crawl_base_url}&page={page}\",\n", 97 | " # url=f\"{crawl_base_url}&page=1\",\n", 98 | " headers={'User-Agent': 'Mozilla/5.0'}\n", 99 | " )\n", 100 | " html = urlopen(req).read().decode('utf-8')\n", 101 | " soup = BeautifulSoup(html, features=\"lxml\")\n", 102 | " articles = soup.find_all('div', {'class': 'row infinite-item item paper-card'})\n", 103 | " for article in articles:\n", 104 | " info = {}\n", 105 | " info[\"title\"] = article.find('h1').get_text()\n", 106 | " info[\"stars\"] = article.find('div', {'class': 'entity-stars'}).get_text()\n", 107 | " detail = article.find('a', {'class': 'badge badge-light'})['href']\n", 108 | " pdf = \"https://paperswithcode.com\" + detail\n", 109 | " print(\"Crawling\", pdf)\n", 110 | " req = Request(\n", 111 | " url=pdf,\n", 112 | " headers={'User-Agent': 'Mozilla/5.0'}\n", 113 | " )\n", 114 | " html = urlopen(req).read().decode('utf-8')\n", 115 | " soup = BeautifulSoup(html, features=\"lxml\")\n", 116 | " authors = soup.find_all('span', {'class': 'author-span'})\n", 117 | " info['date'] = authors[0].get_text()\n", 118 | " info['authors'] = ''\n", 119 | " for a in authors[1:]:\n", 120 | " info['authors'] += a.get_text()\n", 121 | " if not soup.find('a', {'class': 'badge badge-light'}):\n", 122 | " continue\n", 123 | " info['pdf_url'] = soup.find('a', {'class': 'badge badge-light'})['href']\n", 124 | " response = requests.get(info['pdf_url'])\n", 125 | " with open(f\"./pdf/{info['pdf_url'].split('/')[-1]}\", 'wb') as f:\n", 126 | " f.write(response.content)\n", 127 | " metadata.append(info)" 128 | ] 129 | }, 130 | { 131 | "cell_type": "code", 132 | "execution_count": 4, 133 | "metadata": { 134 | "colab": { 135 | "base_uri": "https://localhost:8080/", 136 | "height": 1000 137 | }, 138 | "id": "1lW7Ln2ELTGs", 139 | "outputId": "fd81d882-10d3-48e4-a595-e5bade2653e4" 140 | }, 141 | "outputs": [ 142 | { 143 | "data": { 144 | "text/html": [ 145 | "
\n", 146 | "\n", 159 | "\n", 160 | " \n", 161 | " \n", 162 | " \n", 163 | " \n", 164 | " \n", 165 | " \n", 166 | " \n", 167 | " \n", 168 | " \n", 169 | " \n", 170 | " \n", 171 | " \n", 172 | " \n", 173 | " \n", 174 | " \n", 175 | " \n", 176 | " \n", 177 | " \n", 178 | " \n", 179 | " \n", 180 | " \n", 181 | " \n", 182 | " \n", 183 | " \n", 184 | " \n", 185 | " \n", 186 | " \n", 187 | " \n", 188 | " \n", 189 | " \n", 190 | " \n", 191 | " \n", 192 | " \n", 193 | " \n", 194 | " \n", 195 | " \n", 196 | " \n", 197 | " \n", 198 | " \n", 199 | " \n", 200 | " \n", 201 | " \n", 202 | " \n", 203 | " \n", 204 | " \n", 205 | " \n", 206 | " \n", 207 | " \n", 208 | " \n", 209 | " \n", 210 | " \n", 211 | " \n", 212 | " \n", 213 | " \n", 214 | " \n", 215 | " \n", 216 | " \n", 217 | " \n", 218 | " \n", 219 | " \n", 220 | " \n", 221 | " \n", 222 | " \n", 223 | " \n", 224 | " \n", 225 | " \n", 226 | " \n", 227 | " \n", 228 | " \n", 229 | " \n", 230 | " \n", 231 | " \n", 232 | " \n", 233 | " \n", 234 | " \n", 235 | " \n", 236 | " \n", 237 | " \n", 238 | " \n", 239 | " \n", 240 | " \n", 241 | " \n", 242 | " \n", 243 | " \n", 244 | " \n", 245 | " \n", 246 | " \n", 247 | " \n", 248 | " \n", 249 | " \n", 250 | " \n", 251 | " \n", 252 | " \n", 253 | " \n", 254 | " \n", 255 | " \n", 256 | " \n", 257 | " \n", 258 | " \n", 259 | " \n", 260 | "
titlestarsdateauthorspdf_url
0DeepMask: an algorithm for cloud and cloud sha...\\n 3,114\\n9 Nov 2019\\nKe Xu\\nKaiyu Guan\\nJian Peng\\nYunan Luo\\nSib...https://arxiv.org/pdf/1911.03607v1.pdf
1Self-attention for raw optical Satellite Time ...\\n 3,686\\n23 Oct 2019\\nMarc Rußwurm\\nMarco Körnerhttps://arxiv.org/pdf/1910.10536v3.pdf
2You Only Look Twice: Rapid Multi-Scale Object ...\\n 635\\n24 May 2018\\nAdam Van Ettenhttps://arxiv.org/pdf/1805.09512v1.pdf
3Efficient statistical classification of satell...\\n 490\\n10 Feb 2012\\nPeter Millshttps://arxiv.org/pdf/1202.2194v4.pdf
4DeepGlobe 2018: A Challenge to Parse the Earth...\\n 312\\n17 May 2018\\nIlke Demir\\nKrzysztof Koperski\\nDavid Linden...https://arxiv.org/pdf/1805.06561v1.pdf
5HighRes-net: Recursive Fusion for Multi-Frame ...\\n 257\\n15 Feb 2020\\nMichel Deudon\\nAlfredo Kalaitzis\\nIsrael Goy...https://arxiv.org/pdf/2002.06460v1.pdf
6PAST-AI: Physical-layer Authentication of Sate...\\n 291\\n12 Oct 2020\\nGabriele Oligeri\\nSimone Raponi\\nSavio Scian...https://arxiv.org/pdf/2010.05470v1.pdf
7Semantic Stereo for Incidental Satellite Images\\n 124\\n21 Nov 2018\\nMarc Bosch\\nKevin Foster\\nGordon Christie\\nS...https://arxiv.org/pdf/1811.08739v1.pdf
8SEN12MS -- A Curated Dataset of Georeferenced ...\\n 3,193\\n18 Jun 2019\\nMichael Schmitt\\nLloyd Haydn Hughes\\nChunpin...https://arxiv.org/pdf/1906.07789v1.pdf
9xBD: A Dataset for Assessing Building Damage f...\\n 165\\n21 Nov 2019\\nRitwik Gupta\\nRichard Hosfelt\\nSandra Sajeev...https://arxiv.org/pdf/1911.09296v1.pdf
10Identifying centromeric satellites with dna-brnn\\n 66\\n22 Jan 2019\\nHeng Lihttps://arxiv.org/pdf/1901.07327v1.pdf
\n", 261 | "
" 262 | ], 263 | "text/plain": [ 264 | " title stars \\\n", 265 | "0 DeepMask: an algorithm for cloud and cloud sha... \\n 3,114\\n \n", 266 | "1 Self-attention for raw optical Satellite Time ... \\n 3,686\\n \n", 267 | "2 You Only Look Twice: Rapid Multi-Scale Object ... \\n 635\\n \n", 268 | "3 Efficient statistical classification of satell... \\n 490\\n \n", 269 | "4 DeepGlobe 2018: A Challenge to Parse the Earth... \\n 312\\n \n", 270 | "5 HighRes-net: Recursive Fusion for Multi-Frame ... \\n 257\\n \n", 271 | "6 PAST-AI: Physical-layer Authentication of Sate... \\n 291\\n \n", 272 | "7 Semantic Stereo for Incidental Satellite Images \\n 124\\n \n", 273 | "8 SEN12MS -- A Curated Dataset of Georeferenced ... \\n 3,193\\n \n", 274 | "9 xBD: A Dataset for Assessing Building Damage f... \\n 165\\n \n", 275 | "10 Identifying centromeric satellites with dna-brnn \\n 66\\n \n", 276 | "\n", 277 | " date authors \\\n", 278 | "0 9 Nov 2019 \\nKe Xu\\nKaiyu Guan\\nJian Peng\\nYunan Luo\\nSib... \n", 279 | "1 23 Oct 2019 \\nMarc Rußwurm\\nMarco Körner \n", 280 | "2 24 May 2018 \\nAdam Van Etten \n", 281 | "3 10 Feb 2012 \\nPeter Mills \n", 282 | "4 17 May 2018 \\nIlke Demir\\nKrzysztof Koperski\\nDavid Linden... \n", 283 | "5 15 Feb 2020 \\nMichel Deudon\\nAlfredo Kalaitzis\\nIsrael Goy... \n", 284 | "6 12 Oct 2020 \\nGabriele Oligeri\\nSimone Raponi\\nSavio Scian... \n", 285 | "7 21 Nov 2018 \\nMarc Bosch\\nKevin Foster\\nGordon Christie\\nS... \n", 286 | "8 18 Jun 2019 \\nMichael Schmitt\\nLloyd Haydn Hughes\\nChunpin... \n", 287 | "9 21 Nov 2019 \\nRitwik Gupta\\nRichard Hosfelt\\nSandra Sajeev... \n", 288 | "10 22 Jan 2019 \\nHeng Li \n", 289 | "\n", 290 | " pdf_url \n", 291 | "0 https://arxiv.org/pdf/1911.03607v1.pdf \n", 292 | "1 https://arxiv.org/pdf/1910.10536v3.pdf \n", 293 | "2 https://arxiv.org/pdf/1805.09512v1.pdf \n", 294 | "3 https://arxiv.org/pdf/1202.2194v4.pdf \n", 295 | "4 https://arxiv.org/pdf/1805.06561v1.pdf \n", 296 | "5 https://arxiv.org/pdf/2002.06460v1.pdf \n", 297 | "6 https://arxiv.org/pdf/2010.05470v1.pdf \n", 298 | "7 https://arxiv.org/pdf/1811.08739v1.pdf \n", 299 | "8 https://arxiv.org/pdf/1906.07789v1.pdf \n", 300 | "9 https://arxiv.org/pdf/1911.09296v1.pdf \n", 301 | "10 https://arxiv.org/pdf/1901.07327v1.pdf " 302 | ] 303 | }, 304 | "execution_count": 4, 305 | "metadata": {}, 306 | "output_type": "execute_result" 307 | } 308 | ], 309 | "source": [ 310 | "df = pd.DataFrame(metadata)\n", 311 | "df" 312 | ] 313 | } 314 | ], 315 | "metadata": { 316 | "colab": { 317 | "provenance": [] 318 | }, 319 | "kernelspec": { 320 | "display_name": "Python 3", 321 | "name": "python3" 322 | }, 323 | "language_info": { 324 | "codemirror_mode": { 325 | "name": "ipython", 326 | "version": 3 327 | }, 328 | "file_extension": ".py", 329 | "mimetype": "text/x-python", 330 | "name": "python", 331 | "nbconvert_exporter": "python", 332 | "pygments_lexer": "ipython3", 333 | "version": "3.9.5" 334 | } 335 | }, 336 | "nbformat": 4, 337 | "nbformat_minor": 0 338 | } 339 | -------------------------------------------------------------------------------- /custom_module/pdf/1910.10536v3.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/phkhanhtrinh23/text_summarization/4efd25dfedd13ecb6b88dd3b1a1668a0fe7d96e2/custom_module/pdf/1910.10536v3.pdf -------------------------------------------------------------------------------- /custom_module/pdf/1911.03607v1.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/phkhanhtrinh23/text_summarization/4efd25dfedd13ecb6b88dd3b1a1668a0fe7d96e2/custom_module/pdf/1911.03607v1.pdf -------------------------------------------------------------------------------- /custom_module/requirements.txt: -------------------------------------------------------------------------------- 1 | nltk 2 | tesseract 3 | pytessract 4 | pdf2image 5 | torch 6 | torchvision 7 | transformers -------------------------------------------------------------------------------- /custom_module/templates/base.html: -------------------------------------------------------------------------------- 1 | 2 | 3 | 4 | 5 | 19 | 20 | 21 | 24 |
25 |
26 |

{% block heading %} {% endblock %}

27 | {% block context %} {% endblock %} 28 |
29 | 30 | -------------------------------------------------------------------------------- /custom_module/templates/create.html: -------------------------------------------------------------------------------- 1 | 22 | 23 | {% extends 'base.html' %} 24 | 25 | {% block context %} 26 |
27 |
28 | 29 |

30 | 33 | 38 |

39 | 40 |

41 | 46 |

47 | 48 |
49 |
50 | {% endblock %} -------------------------------------------------------------------------------- /custom_module/templates/index.html: -------------------------------------------------------------------------------- 1 | 22 | {% extends 'base.html' %} 23 | 24 | {% block heading %} Results {% endblock %} 25 | 26 | {% block context %} 27 | {% for message in messages %} 28 |
29 |

URL for website: {{ message['text_web'] }}

30 |
31 |

URL for downloading multiple pdf file: {{ message['text_pdf'] }}

32 |
33 |

Answer from a website:

34 |
35 | {% for values in message['answer_web'] %} 36 |

{{ values[0] }}

{{ values[1] }}

37 | {% endfor %} 38 |
39 |

Answer from multiple pdf files:

40 |
41 | {% for values in message['answer_pdf'] %} 42 |

{{ values[0] }}

{{ values[1] }}

43 | {% endfor %} 44 |
45 | {% endfor %} 46 |
47 | 48 |
49 | {% endblock %} -------------------------------------------------------------------------------- /custom_module/utils.py: -------------------------------------------------------------------------------- 1 | from pdf2image import convert_from_path 2 | import pytesseract 3 | from PIL import Image 4 | import os 5 | from bs4 import BeautifulSoup 6 | from urllib.request import Request, urlopen 7 | import requests 8 | import os 9 | from datetime import datetime 10 | 11 | 12 | def crawl_web(crawl_base_url, crawl_from='July 5, 2023', crawl_to='July 12, 2023'): 13 | date_format = '%B %d, %Y' 14 | crawl_from = datetime.strptime(crawl_from, date_format).date() 15 | crawl_to = date_object = datetime.strptime(crawl_to, date_format).date() 16 | 17 | metadata = [] 18 | stop = False 19 | for page in range(1, 3): 20 | if stop == True: 21 | break 22 | 23 | req = Request( 24 | url=f"{crawl_base_url}/page/{page}/", 25 | headers={'User-Agent': 'Mozilla/5.0'} 26 | ) 27 | html = urlopen(req).read().decode('utf-8') 28 | soup = BeautifulSoup(html, features='lxml') 29 | articles = soup.find('main', {'id': 'main'}).find_all('article') 30 | 31 | for i, article in enumerate(articles): 32 | info = {} 33 | date_string = article.find('time', {'class': 'entry-date published'}).get_text() 34 | date_object = datetime.strptime(date_string, date_format).date() 35 | 36 | if date_object > crawl_to: 37 | continue 38 | 39 | if date_object <= crawl_to and date_object >= crawl_from: 40 | info['date'] = date_object 41 | info['title'] = article.find('h2').get_text() 42 | info['author'] = article.find('span', {'class': 'author vcard'}).get_text() 43 | info['thumbnail'] = article.find('img')['src'] 44 | url = article.find('a')['href'] 45 | html = urlopen(url).read().decode('utf-8') 46 | soup = BeautifulSoup(html, features='lxml') 47 | content = soup.find('div', {'class': 'entry-content'}) 48 | info['text'] = '' 49 | 50 | for el in content.find_all(recursive=False): 51 | if 'Related' not in el.get_text(): 52 | info['text'] += el.get_text() + '\n' 53 | 54 | print("Crawled web: ", info) 55 | 56 | metadata.append(info) 57 | 58 | if date_object < crawl_from: 59 | stop = True 60 | break 61 | 62 | if i == 1: # Just 2 articles for demo 63 | break 64 | 65 | return metadata 66 | 67 | 68 | def crawl_pdf(crawl_base_url): 69 | metadata = [] 70 | if os.path.exists('pdf/') is False: 71 | os.mkdir('pdf/') 72 | for page in range(1, 2): # Crawl 1 web pages (not pdf pages) which includes many pdf files 73 | req = Request( 74 | url=f"{crawl_base_url}&page={page}", 75 | headers={'User-Agent': 'Mozilla/5.0'} 76 | ) 77 | html = urlopen(req).read().decode('utf-8') 78 | soup = BeautifulSoup(html, features="lxml") 79 | articles = soup.find_all('div', {'class': 'row infinite-item item paper-card'}) 80 | 81 | for i, article in enumerate(articles): 82 | info = {} 83 | info["title"] = article.find('h1').get_text() 84 | info["stars"] = article.find('div', {'class': 'entity-stars'}).get_text() 85 | detail = article.find('a', {'class': 'badge badge-light'})['href'] 86 | pdf = "https://paperswithcode.com" + detail 87 | print("Crawling pdf:", pdf) 88 | 89 | req = Request( 90 | url=pdf, 91 | headers={'User-Agent': 'Mozilla/5.0'} 92 | ) 93 | html = urlopen(req).read().decode('utf-8') 94 | soup = BeautifulSoup(html, features="lxml") 95 | authors = soup.find_all('span', {'class': 'author-span'}) 96 | info['date'] = authors[0].get_text() 97 | info['authors'] = '' 98 | 99 | for a in authors[1:]: 100 | info['authors'] += a.get_text() 101 | 102 | if not soup.find('a', {'class': 'badge badge-light'}): 103 | continue 104 | 105 | info['pdf_url'] = soup.find('a', {'class': 'badge badge-light'})['href'] 106 | response = requests.get(info['pdf_url']) 107 | 108 | with open(f"./pdf/{info['pdf_url'].split('/')[-1]}", 'wb') as f: 109 | f.write(response.content) 110 | metadata.append(info) 111 | 112 | if i == 1: # Just 2 pdfs for demo 113 | break 114 | 115 | return metadata 116 | 117 | 118 | def extract_OCR(file): 119 | pages = convert_from_path(file, 500) 120 | 121 | image_counter = 1 122 | 123 | for page in pages: 124 | filename = "page_" + str(image_counter) + ".jpg" 125 | page.save(filename, "JPEG") 126 | image_counter = image_counter + 1 127 | 128 | limit = image_counter-1 129 | text = "" 130 | 131 | for i in range(1, limit + 1): 132 | filename = "page_" + str(i) + ".jpg" 133 | page = str(((pytesseract.image_to_string(Image.open(filename))))) 134 | page = page.replace("-\n", "") 135 | text += page 136 | os.remove(filename) 137 | 138 | return text 139 | --------------------------------------------------------------------------------