15 |

16 |

17 |

setInput(e.target.value)}
24 |             onKeyDown={handleKeyDown}
25 |           />
26 |           <button
27 |             type="submit"
28 |             className="ml-2 px-2 py-1 rounded-lg bg-blue-500 text-white focus:outline-none hover:bg-blue-600 md:ml-4 md:px-4 md:py-2"
29 |           >
30 |             Send
31 |           </button>
32 |         </div>
33 |       </form>
34 |       <div className="pb-2 text-center text-xs text-gray-400 md:pb-4">
35 |         This is still a WIP, and answers may not be correct. Algorithm code last updated 2023-04-14.
36 |       </div>
37 |     </div>
38 |   )
39 | }
40 | 
41 | export default InputBar

--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # Chat With the Algorithm
 2 | 
 3 | Ask questions about the open-sourced [Twitter algorithm](https://github.com/twitter/the-algorithm).
 4 | 
 5 | The app is publicly hosted here: https://chat-twitter.vercel.app/. Instructions for hosting it yourself are below.
 6 | 
 7 | ## Basic architecture
 8 | 
 9 | The app is a NextJS/Tailwind CSS frontend with a FastAPI backend. The frontend is hosted on Vercel and the backend is hosted on a small node on [fly.io](https://fly.io/). The backend uses a Pinecone vector DB on the free tier. There is a Dockerfile provided.
10 | 
11 | Right now, I'm footing the OpenAI bill on the public instance. But I may require users to bring their own key in the future.
12 | 
13 | ## Running locally
14 | 
15 | 1. Set up environment variables
16 | 
17 | ```
18 | OPENAI_API_KEY=...
19 | OPENAI_ORG_ID=... # organization id, found in Manage account > settings
20 | PINECONE_API_KEY=...
21 | ```
22 | 
23 | 2. Clone the repo
24 | 
25 | ```
26 | git clone https://github.com/mtenenholtz/chat-twitter.git
27 | cd chat-twitter
28 | ```
29 | 
30 | 3. Install Node dependencies
31 | 
32 | ```
33 | npm i
34 | ```
35 | 
36 | 4. Run the Node server
37 | 
38 | ```
39 | npm run dev
40 | ```
41 | 
42 | 5. In another terminal, install the Python dependencies
43 | 
44 | ```
45 | # in the backend/ directory
46 | pip install -r requirements.txt
47 | ```
48 | 
49 | 6. Embed the Twitter codebase
50 | 
51 | ```
52 | # in the backend/ directory
53 | python create_vector_db.py
54 | ```
55 | 
56 | 7. Set up a Pinecone index. Give it a vector dimension of 1536 and name it `pinecone-index`. You can change this in `backend/main.py` if you want.
57 | 
58 | 8. Run the backend server
59 | 
60 | ```
61 | uvicorn main:app --reload
62 | ```
63 | 
64 | 9. The URL for the backend is currently hard coded to the live server URL. You will have to change this to localhost or your other server name.
65 | 
66 | ## Potential improvements
67 | 
68 | I will continue development on this project as demand exists. But, if I abandon it, here are some ideas:
69 | 
70 | - The dependency on Pinecone could be removed and replaced with a simple NumPy array. I just wanted to try Pinecone.
71 | - Replace the `chat_stream` endpoint with a websocket implementation.
72 | - Ask the model not to generatively reference its sources. Instead, simply copy the code snippet directly.
73 | - The splitter could be improved. Right now, it's a character splitter that favors newlines, but OpenAI has implemented a similar one that splits on tokens instead.
74 | - The embeddings and retrieval mechanisms could account for the hierarchy of the Algorithm's code structure, like Replit's Ghostwriter does.
75 | - The UI could be improved **a lot**. I suck at JS.

--------------------------------------------------------------------------------
/backend/create_vector_db.py:
--------------------------------------------------------------------------------
 1 | from langchain.text_splitter import RecursiveCharacterTextSplitter
 2 | from langchain.embeddings import OpenAIEmbeddings
 3 | from langchain.vectorstores import Pinecone
 4 | import pinecone
 5 | 
 6 | import pandas as pd
 7 | import tiktoken
 8 | 
 9 | from tqdm import tqdm
10 | 
11 | import os
12 | import zipfile
13 | from urllib.request import urlopen
14 | from io import BytesIO
15 | 
16 | def embed_document(vector_db, splitter, document_id, document):
17 |     metadata = [{'document_id': document_id}]
18 |     split_documents = splitter.create_documents([str(document)], metadatas=metadata)
19 | 
20 |     texts = [d.page_content for d in split_documents]
21 |     metadatas = [d.metadata for d in split_documents]
22 | 
23 |     docsearch = vector_db.add_texts(texts, metadatas=metadatas)
24 | 
25 | def zipfile_from_github():
26 |     http_response = urlopen('https://github.com/twitter/the-algorithm/archive/refs/heads/main.zip')
27 |     zf = BytesIO(http_response.read())
28 |     return zipfile.ZipFile(zf, 'r')
29 | 
30 | embeddings = OpenAIEmbeddings(
31 |     openai_api_key=os.environ['OPENAI_API_KEY'],
32 |     openai_organization=os.environ['OPENAI_ORG_ID'],
33 | )
34 | encoder = tiktoken.get_encoding('cl100k_base')
35 | 
36 | pinecone.init(
37 |     api_key=os.environ['PINECONE_API_KEY'],
38 |     environment='us-east1-gcp'
39 | )
40 | vector_store = Pinecone(
41 |     index=pinecone.Index('pinecone-index'),
42 |     embedding_function=embeddings.embed_query,
43 |     text_key='text',
44 |     namespace='twitter-algorithm'
45 | )
46 | 
47 | splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
48 | 
49 | total_tokens, corpus_summary = 0, []
50 | file_texts, metadatas = [], []
51 | with zipfile_from_github() as zip_ref:
52 |     zip_file_list = zip_ref.namelist()
53 |     
54 |     pbar = tqdm(zip_file_list, desc=f'Total tokens: 0')
55 |     for file_name in pbar:
56 |         if (file_name.endswith('/') or 
57 |             any(f in file_name for f in ['.DS_Store', '.gitignore']) or 
58 |             any(file_name.endswith(ext) for ext in ['.png', '.jpg', '.jpeg'])
59 |         ):
60 |             continue
61 |         else:
62 |             with zip_ref.open(file_name, 'r') as file:
63 |                 file_contents = str(file.read())
64 |                 file_name_trunc = str(file_name).replace('the-algorithm-main/', '')
65 |                 
66 |                 n_tokens = len(encoder.encode(file_contents))
67 |                 total_tokens += n_tokens
68 |                 corpus_summary.append({'file_name': file_name_trunc, 'n_tokens': n_tokens})
69 | 
70 |                 file_texts.append(file_contents)
71 |                 metadatas.append({'document_id': file_name_trunc})
72 |                 pbar.set_description(f'Total tokens: {total_tokens}')
73 | 
74 | split_documents = splitter.create_documents(file_texts, metadatas=metadatas)
75 | vector_store.from_documents(
76 |     documents=split_documents, 
77 |     embedding=embeddings,
78 |     index_name='pinecone-index',
79 |     namespace='twitter-algorithm'
80 | )
81 | 
82 | pd.DataFrame.from_records(corpus_summary).to_csv('data/corpus_summary.csv', index=False)

--------------------------------------------------------------------------------
/components/ChatMessages.js:
--------------------------------------------------------------------------------
  1 | import React, { useState, useEffect } from 'react';
  2 | import { Prism as SyntaxHighlighter } from 'react-syntax-highlighter';
  3 | import { oneDark } from 'react-syntax-highlighter/dist/esm/styles/prism';
  4 | 
  5 | const ChatMessages = ({ messages }) => {
  6 |   const [fileData, setFileData] = useState([]);
  7 | 
  8 |   const filteredMessages = messages.filter(
  9 |     (message) =>
 10 |       !(message.sender === 'systemMessage' | (message.sender !== 'user' && message.text.length === 0)),
 11 |   );
 12 | 
 13 |   const formatMessage = (text) => {
 14 |     const codeBlockRegex = /(```[\w]*[\s\S]+?```)/g;
 15 |     const parts = text.split(codeBlockRegex);
 16 |   
 17 |     return parts.map((part, index) => {
 18 |       if (codeBlockRegex.test(part)) {
 19 |         const languageRegex = /```(\w*)\n/;
 20 |         const languageMatch = part.match(languageRegex);
 21 |         const language = languageMatch && languageMatch[1] ? languageMatch[1] : '';
 22 |   
 23 |         return (
 24 |           <SyntaxHighlighter
 25 |             key={index}
 26 |             language={language}
 27 |             style={oneDark}
 28 |             customStyle={{ backgroundColor: '#2d2d2d', borderRadius: '0.375rem', padding: '1rem' }}
 29 |           >
 30 |             {part.replace(languageRegex, '').replace(/```$/, '')}
 31 |           </SyntaxHighlighter>
 32 |         );
 33 |       } else {
 34 |         return (
 35 |           <span
 36 |             key={index}
 37 |             dangerouslySetInnerHTML={{
 38 |               __html: replaceFileNamesWithLinks(part, fileData),
 39 |             }}
 40 |           />
 41 |         );
 42 |       }
 43 |     });
 44 |   };
 45 | 
 46 |   useEffect(() => {
 47 |     async function fetchData() {
 48 |       const response = await fetch('api/fetchCsvData', {
 49 |           headers: {
 50 |             'Content-Type': 'application/json',
 51 |           },
 52 |       })
 53 |       .catch((error) => {
 54 |         console.log(error)
 55 |       });
 56 |       const data = await response.json();
 57 |       setFileData(data);
 58 |     };
 59 |     fetchData();
 60 |   }, []);
 61 | 
 62 |   const replaceFileNamesWithLinks = (text, data) => {
 63 |     const validExtensions = ['.md', '.py', '.scala', '.java', '.cpp', '.rs'];
 64 |     const filenameRegex = /`?((?:[\w/_-]+\/README)|[\w-_]+(?:\.(?:md|py|scala|java|cpp|rs)))`?/g;
 65 |     return text.replace(filenameRegex, (match, filename) => {
 66 |       // Check if the filename is "README" with a full file path
 67 |       const isReadmeWithPath = filename.endsWith('/README');
 68 |       const fileExtension = filename.split('.').pop();
 69 |       const isValidExtension = validExtensions.includes(`.${fileExtension}`);
 70 |   
 71 |       if (isReadmeWithPath) {
 72 |         // Use the full file path directly to create the URL for "README"
 73 |         const url = `https://github.com/twitter/the-algorithm/blob/main/${filename}`;
 74 |         return `<a class="text-blue-500 underline cursor-pointer" href="${url}" target="_blank" rel="noopener noreferrer">${filename}</a>`;
 75 |       } else if (isValidExtension) {
 76 |         // Find the record in the data list that matches the basename and has a valid extension
 77 |         const fileRecord = data.find((record) => {
 78 |           const recordBasename = record.file_name.split('/').pop();
 79 |           return recordBasename === filename;
 80 |         });
 81 |         if (fileRecord) {
 82 |           const url = `https://github.com/twitter/the-algorithm/blob/main/${fileRecord.file_name}`;
 83 |           return `<a class="text-blue-500 underline cursor-pointer" href="${url}" target="_blank" rel="noopener noreferrer">${filename}</a>`;
 84 |         }
 85 |       }
 86 |       return filename;
 87 |     });
 88 |   };
 89 | 
 90 |   return (
 91 |     <div className="w-full md:w-1/2 md:max-w-xl">
 92 |       {filteredMessages.map((message, index) => (
 93 |         <div
 94 |           key={index}
 95 |           className={`mb-4 p-3 text-lg rounded-lg shadow-md whitespace-pre-wrap ${
 96 |             message.sender === 'user' ? 'bg-gray-100 text-gray-800' : 'bg-gray-600'
 97 |           }`}
 98 |         >
 99 |           {formatMessage(message.text)}
100 |         </div>
101 |       ))}
102 |     </div>
103 |   );
104 | };
105 | 
106 | export default ChatMessages;

--------------------------------------------------------------------------------
/app/page.js:
--------------------------------------------------------------------------------
  1 | 'use client';
  2 | 
  3 | import { useState, useEffect, useRef } from 'react'
  4 | 
  5 | import Head from 'next/head'
  6 | import Header from '../components/Header'
  7 | import ChatMessages from '../components/ChatMessages'
  8 | import InputBar from '../components/InputBar'
  9 | 
 10 | export default function Home() {
 11 |     const [messages, setMessages] = useState([])
 12 |     const [input, setInput] = useState('')
 13 |     const inputRef = useRef(null)
 14 | 
 15 |     const handleKeyDown = (e) => {
 16 |         if (e.key === 'Enter' && !e.shiftKey) {
 17 |             handleSubmit(e)
 18 |         }
 19 |     }
 20 | 
 21 |     const getSystemMessage = async (userInputMessage) => {
 22 |         const response = await fetch('https://chat-twitter-backend.fly.dev/system_message', {
 23 |             method: 'POST',
 24 |             headers: {
 25 |                 'Content-Type': 'application/json',
 26 |             },
 27 |             body: JSON.stringify(userInputMessage),
 28 |         })
 29 |         const systemMessage = await response.json();
 30 |         return { text: systemMessage.system_message, sender: 'systemMessage' }
 31 |     }
 32 | 
 33 |     const handleSubmit = async (e) => {
 34 |         e.preventDefault()
 35 | 
 36 |         let updatedMessages = []
 37 |         if (input.trim()) {
 38 |             const userInputMessage = { text: input, sender: 'user' }
 39 |             if (messages.length === 0) {
 40 |                 const systemMessage = await getSystemMessage(userInputMessage);
 41 |                 updatedMessages = [systemMessage, userInputMessage];
 42 |             } else {
 43 |                 updatedMessages = [...messages, userInputMessage]
 44 |             }
 45 |             setMessages(updatedMessages)
 46 | 
 47 |             await handleChat(updatedMessages)
 48 | 
 49 |             setInput('')
 50 |         }
 51 |     }
 52 | 
 53 |     useEffect(() => {
 54 |         if (inputRef.current) {
 55 |             inputRef.current.style.height = 'auto'
 56 |             inputRef.current.style.height = inputRef.current.scrollHeight + 'px'
 57 |         }
 58 |     }, [input])
 59 | 
 60 |     const handleChat = async (updatedMessages) => {
 61 |         let accumulatedText = "";
 62 |         fetch('https://chat-twitter-backend.fly.dev/chat_stream', {
 63 |             method: 'POST',
 64 |             headers: {
 65 |                 'Content-Type': 'application/json',
 66 |             },
 67 |             body: JSON.stringify(updatedMessages),
 68 |         })
 69 |         .then(response => {
 70 |             const reader = response.body.getReader();
 71 |             return new ReadableStream({
 72 |                 async start(controller) {
 73 |                     while (true) {
 74 |                         const { done, value } = await reader.read();
 75 |                         if (done) {
 76 |                             break;
 77 |                         }
 78 |                         let newToken = new TextDecoder().decode(value);
 79 |                         accumulatedText += newToken;
 80 |                         controller.enqueue(newToken);
 81 |                     }
 82 |                     controller.close();
 83 |                     reader.releaseLock();
 84 |                 }
 85 |             });
 86 |         })
 87 |         .then(stream => {
 88 |             updatedMessages = [...updatedMessages, { text: '', sender: 'llm' }];
 89 |             setMessages(updatedMessages);
 90 |             const reader = stream.getReader();
 91 |             reader.read().then(function processText({ done, value }) {
 92 |                 if (done) {
 93 |                     return;
 94 |                 }
 95 |                 setMessages((prevMessages) => {
 96 |                     let outputMessage = prevMessages[prevMessages.length - 1];
 97 |                     outputMessage.text = accumulatedText;
 98 |                     return [...prevMessages.slice(0, -1), outputMessage];
 99 |                 });
100 |                 return reader.read().then(processText);
101 |             });
102 |         });
103 |     };
104 | 
105 |   
106 | return (
107 |     <>
108 |         <Head>
109 |             <title>Chat With the Twitter Algorithm</title>
110 |             <meta
111 |                 name="description"
112 |                 content="Chat with the Twitter algorithm."
113 |             />
114 |             <link rel="icon" href="/favicon.ico" />
115 |             <link href="https://fonts.googleapis.com/css2?family=Roboto:wght@400;700&display=swap" rel="stylesheet" />
116 |         </Head>
117 |   
118 |         <div className="h-screen flex flex-col bg-gray-800 text-gray-100 font-sans font-roboto">
119 |             <Header />
120 |             <div className="flex-1 overflow-auto p-4 flex justify-center">
121 |                 <ChatMessages messages={messages} />
122 |             </div>
123 |   
124 |             <div className="border-t border-gray-700">
125 |                 <InputBar
126 |                     input={input}
127 |                     setInput={setInput}
128 |                     handleKeyDown={handleKeyDown}
129 |                     handleSubmit={handleSubmit}
130 |                 />
131 |             </div>
132 |       </div>
133 |     </>
134 |   )
135 | }

--------------------------------------------------------------------------------
/backend/main.py:
--------------------------------------------------------------------------------
  1 | import os
  2 | import threading
  3 | import queue
  4 | 
  5 | from typing import List, Optional, Dict
  6 | from fastapi import FastAPI, HTTPException
  7 | from fastapi.middleware.cors import CORSMiddleware
  8 | from fastapi.responses import StreamingResponse
  9 | from pydantic import BaseModel
 10 | 
 11 | from langchain.chat_models import ChatOpenAI
 12 | from langchain.schema import (
 13 |     AIMessage,
 14 |     HumanMessage,
 15 |     SystemMessage
 16 | )
 17 | from langchain.callbacks.base import CallbackManager
 18 | from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
 19 | from langchain.embeddings import OpenAIEmbeddings
 20 | from langchain.vectorstores import Pinecone
 21 | 
 22 | import pinecone
 23 | import openai
 24 | import tiktoken
 25 | import os
 26 | 
 27 | app = FastAPI()
 28 | 
 29 | origins = [
 30 |     "http://localhost",
 31 |     "http://localhost:8080",
 32 |     "http://localhost:3000",
 33 |     "https://chat-twitter.fly.dev",
 34 |     "https://chat-twitter.fly.dev:8000"
 35 |     "https://chat-twitter.fly.dev:8080"
 36 | ]
 37 | 
 38 | app.add_middleware(
 39 |     CORSMiddleware,
 40 |     allow_origins=['*'],
 41 |     allow_credentials=True,
 42 |     allow_methods=["*"],
 43 |     allow_headers=["*"],
 44 | )
 45 | 
 46 | pinecone.init(
 47 |     api_key=os.environ['PINECONE_API_KEY'],
 48 |     environment='us-east1-gcp'
 49 | )
 50 | 
 51 | class Message(BaseModel):
 52 |     text: str
 53 |     sender: str
 54 | 
 55 | class ContextSystemMessage(BaseModel):
 56 |     system_message: str
 57 |     
 58 | class Chat(BaseModel):
 59 |     messages: list[dict]
 60 | 
 61 | class ThreadedGenerator:
 62 |     def __init__(self):
 63 |         self.queue = queue.Queue()
 64 | 
 65 |     def __iter__(self):
 66 |         return self
 67 | 
 68 |     def __next__(self):
 69 |         item = self.queue.get()
 70 |         if item is StopIteration: raise item
 71 |         return item
 72 | 
 73 |     def send(self, data):
 74 |         self.queue.put(data)
 75 | 
 76 |     def close(self):
 77 |         self.queue.put(StopIteration)
 78 | 
 79 | class ChainStreamHandler(StreamingStdOutCallbackHandler):
 80 |     def __init__(self, gen):
 81 |         super().__init__()
 82 |         self.gen = gen
 83 | 
 84 |     def on_llm_new_token(self, token: str, **kwargs):
 85 |         self.gen.send(token)
 86 | 
 87 | def format_context(docs):
 88 |     context = '\n\n'.join([f'From file {d.metadata["document_id"]}:\n' + str(d.page_content) for d in docs])
 89 |     return context
 90 | 
 91 | def format_query(query, context):
 92 |     return f"""Relevant context: {context}
 93 |     
 94 |     {query}"""
 95 | 
 96 | def embedding_search(query, k):
 97 |     embeddings = OpenAIEmbeddings(
 98 |         openai_api_key=os.environ['OPENAI_API_KEY'],
 99 |         openai_organization=os.environ['OPENAI_ORG_ID'],
100 |     )
101 |     db = Pinecone(
102 |         index=pinecone.Index('pinecone-index'),
103 |         embedding_function=embeddings.embed_query,
104 |         text_key='text',
105 |         namespace='twitter-algorithm'
106 |     )
107 | 
108 |     return db.similarity_search(query, k=k)
109 | 
110 | @app.get("/heath")
111 | def health():
112 |     return "OK"
113 | 
114 | @app.post("/system_message", response_model=ContextSystemMessage)
115 | def system_message(query: Message):
116 |     docs = embedding_search(query.text, k=10)
117 |     context = format_context(docs)
118 | 
119 |     prompt = """Given the following context and code, answer the following question. Do not use outside context, and do not assume the user can see the provided context. Try to be as detailed as possible and reference the components that you are looking at. Keep in mind that these are only code snippets, and more snippets may be added during the conversation.
120 |     Do not generate code, only reference the exact code snippets that you have been provided with. If you are going to write code, make sure to specify the language of the code. For example, if you were writing Python, you would write the following:
121 | 
122 |     ```python
123 |     <python code goes here>
124 |     ```
125 |     
126 |     Now, here is the relevant context: 
127 | 
128 |     Context: {context}
129 |     """
130 | 
131 |     return {'system_message': prompt.format(context=context)}
132 | 
133 | @app.post("/chat_stream")
134 | async def chat_stream(chat: List[Message]):
135 |     model_name = 'gpt-3.5-turbo'
136 |     encoding_name = 'cl100k_base'
137 | 
138 |     def llm_thread(g, prompt):
139 |         try:
140 |             llm = ChatOpenAI(
141 |                 model_name=model_name,
142 |                 verbose=True,
143 |                 streaming=True,
144 |                 callback_manager=CallbackManager([ChainStreamHandler(g)]),
145 |                 temperature=0.7,
146 |                 openai_api_key=os.environ['OPENAI_API_KEY'],
147 |                 openai_organization=os.environ['OPENAI_ORG_ID']
148 |             )
149 | 
150 |             encoding = tiktoken.get_encoding(encoding_name)
151 | 
152 |             # the system message gets 10 new docs. Only include 2 more for new queries
153 |             if len(chat) > 2:
154 |                 system_message, latest_query = [chat[0].text, chat[-1].text]
155 |                 keep_messages = [system_message, latest_query]
156 |                 new_messages = []
157 | 
158 |                 token_count = sum([len(encoding.encode(m)) for m in keep_messages])
159 |                 # fit in as many of the previous human messages as possible
160 |                 for message in chat[1:-1:2]:
161 |                     token_count += len(encoding.encode(message.text))
162 | 
163 |                     if token_count > 750:
164 |                         break
165 | 
166 |                     new_messages.append(message.text)
167 |                     
168 |                 query_messages = [system_message] + new_messages + [latest_query]
169 |                 query_text = '\n'.join(query_messages)
170 | 
171 |                 # add some more context
172 |                 docs = embedding_search(query_text, k=2)
173 |                 context = format_context(docs)
174 |                 formatted_query = format_query(latest_query, context)
175 |             else:
176 |                 formatted_query = chat[-1].text
177 | 
178 |             # always include the system message and the latest query in the prompt
179 |             system_message = SystemMessage(content=chat[0].text)
180 |             latest_query = HumanMessage(content=formatted_query)
181 |             messages = [latest_query]
182 | 
183 |             # for all the rest of the messages, iterate over them in reverse and fit as many in as possible
184 |             token_limit = 4000
185 |             num_tokens = len(encoding.encode(chat[0].text)) + len(encoding.encode(formatted_query))
186 |             for message in reversed(chat[1:-1]):
187 |                 # count the number of new tokens
188 |                 num_tokens += 4
189 |                 num_tokens += len(encoding.encode(message.text))
190 | 
191 |                 if num_tokens > token_limit:
192 |                     # if we're over the token limit, stick with what we've got
193 |                     break
194 |                 else:
195 |                     # otherwise, add the new message in after the system prompt, but before the rest of the messages we've added
196 |                     new_message = HumanMessage(content=message.text) if message.sender == 'user' else AIMessage(content=message.text)
197 |                     messages = [new_message] + messages
198 | 
199 |             # add the system message to the beginning of the prompt
200 |             messages = [system_message] + messages
201 | 
202 |             llm(messages)
203 | 
204 |         finally:
205 |             g.close()
206 | 
207 |     def chat_fn(prompt):
208 |         g = ThreadedGenerator()
209 |         threading.Thread(target=llm_thread, args=(g, prompt)).start()
210 |         return g
211 | 
212 |     return StreamingResponse(chat_fn(chat), media_type='text/event-stream')
213 |

--------------------------------------------------------------------------------
/LICENSE:
--------------------------------------------------------------------------------
  1 |                                  Apache License
  2 |                            Version 2.0, January 2004
  3 |                         http://www.apache.org/licenses/
  4 | 
  5 |    TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
  6 | 
  7 |    1. Definitions.
  8 | 
  9 |       "License" shall mean the terms and conditions for use, reproduction,
 10 |       and distribution as defined by Sections 1 through 9 of this document.
 11 | 
 12 |       "Licensor" shall mean the copyright owner or entity authorized by
 13 |       the copyright owner that is granting the License.
 14 | 
 15 |       "Legal Entity" shall mean the union of the acting entity and all
 16 |       other entities that control, are controlled by, or are under common
 17 |       control with that entity. For the purposes of this definition,
 18 |       "control" means (i) the power, direct or indirect, to cause the
 19 |       direction or management of such entity, whether by contract or
 20 |       otherwise, or (ii) ownership of fifty percent (50%) or more of the
 21 |       outstanding shares, or (iii) beneficial ownership of such entity.
 22 | 
 23 |       "You" (or "Your") shall mean an individual or Legal Entity
 24 |       exercising permissions granted by this License.
 25 | 
 26 |       "Source" form shall mean the preferred form for making modifications,
 27 |       including but not limited to software source code, documentation
 28 |       source, and configuration files.
 29 | 
 30 |       "Object" form shall mean any form resulting from mechanical
 31 |       transformation or translation of a Source form, including but
 32 |       not limited to compiled object code, generated documentation,
 33 |       and conversions to other media types.
 34 | 
 35 |       "Work" shall mean the work of authorship, whether in Source or
 36 |       Object form, made available under the License, as indicated by a
 37 |       copyright notice that is included in or attached to the work
 38 |       (an example is provided in the Appendix below).
 39 | 
 40 |       "Derivative Works" shall mean any work, whether in Source or Object
 41 |       form, that is based on (or derived from) the Work and for which the
 42 |       editorial revisions, annotations, elaborations, or other modifications
 43 |       represent, as a whole, an original work of authorship. For the purposes
 44 |       of this License, Derivative Works shall not include works that remain
 45 |       separable from, or merely link (or bind by name) to the interfaces of,
 46 |       the Work and Derivative Works thereof.
 47 | 
 48 |       "Contribution" shall mean any work of authorship, including
 49 |       the original version of the Work and any modifications or additions
 50 |       to that Work or Derivative Works thereof, that is intentionally
 51 |       submitted to Licensor for inclusion in the Work by the copyright owner
 52 |       or by an individual or Legal Entity authorized to submit on behalf of
 53 |       the copyright owner. For the purposes of this definition, "submitted"
 54 |       means any form of electronic, verbal, or written communication sent
 55 |       to the Licensor or its representatives, including but not limited to
 56 |       communication on electronic mailing lists, source code control systems,
 57 |       and issue tracking systems that are managed by, or on behalf of, the
 58 |       Licensor for the purpose of discussing and improving the Work, but
 59 |       excluding communication that is conspicuously marked or otherwise
 60 |       designated in writing by the copyright owner as "Not a Contribution."
 61 | 
 62 |       "Contributor" shall mean Licensor and any individual or Legal Entity
 63 |       on behalf of whom a Contribution has been received by Licensor and
 64 |       subsequently incorporated within the Work.
 65 | 
 66 |    2. Grant of Copyright License. Subject to the terms and conditions of
 67 |       this License, each Contributor hereby grants to You a perpetual,
 68 |       worldwide, non-exclusive, no-charge, royalty-free, irrevocable
 69 |       copyright license to reproduce, prepare Derivative Works of,
 70 |       publicly display, publicly perform, sublicense, and distribute the
 71 |       Work and such Derivative Works in Source or Object form.
 72 | 
 73 |    3. Grant of Patent License. Subject to the terms and conditions of
 74 |       this License, each Contributor hereby grants to You a perpetual,
 75 |       worldwide, non-exclusive, no-charge, royalty-free, irrevocable
 76 |       (except as stated in this section) patent license to make, have made,
 77 |       use, offer to sell, sell, import, and otherwise transfer the Work,
 78 |       where such license applies only to those patent claims licensable
 79 |       by such Contributor that are necessarily infringed by their
 80 |       Contribution(s) alone or by combination of their Contribution(s)
 81 |       with the Work to which such Contribution(s) was submitted. If You
 82 |       institute patent litigation against any entity (including a
 83 |       cross-claim or counterclaim in a lawsuit) alleging that the Work
 84 |       or a Contribution incorporated within the Work constitutes direct
 85 |       or contributory patent infringement, then any patent licenses
 86 |       granted to You under this License for that Work shall terminate
 87 |       as of the date such litigation is filed.
 88 | 
 89 |    4. Redistribution. You may reproduce and distribute copies of the
 90 |       Work or Derivative Works thereof in any medium, with or without
 91 |       modifications, and in Source or Object form, provided that You
 92 |       meet the following conditions:
 93 | 
 94 |       (a) You must give any other recipients of the Work or
 95 |           Derivative Works a copy of this License; and
 96 | 
 97 |       (b) You must cause any modified files to carry prominent notices
 98 |           stating that You changed the files; and
 99 | 
100 |       (c) You must retain, in the Source form of any Derivative Works
101 |           that You distribute, all copyright, patent, trademark, and
102 |           attribution notices from the Source form of the Work,
103 |           excluding those notices that do not pertain to any part of
104 |           the Derivative Works; and
105 | 
106 |       (d) If the Work includes a "NOTICE" text file as part of its
107 |           distribution, then any Derivative Works that You distribute must
108 |           include a readable copy of the attribution notices contained
109 |           within such NOTICE file, excluding those notices that do not
110 |           pertain to any part of the Derivative Works, in at least one
111 |           of the following places: within a NOTICE text file distributed
112 |           as part of the Derivative Works; within the Source form or
113 |           documentation, if provided along with the Derivative Works; or,
114 |           within a display generated by the Derivative Works, if and
115 |           wherever such third-party notices normally appear. The contents
116 |           of the NOTICE file are for informational purposes only and
117 |           do not modify the License. You may add Your own attribution
118 |           notices within Derivative Works that You distribute, alongside
119 |           or as an addendum to the NOTICE text from the Work, provided
120 |           that such additional attribution notices cannot be construed
121 |           as modifying the License.
122 | 
123 |       You may add Your own copyright statement to Your modifications and
124 |       may provide additional or different license terms and conditions
125 |       for use, reproduction, or distribution of Your modifications, or
126 |       for any such Derivative Works as a whole, provided Your use,
127 |       reproduction, and distribution of the Work otherwise complies with
128 |       the conditions stated in this License.
129 | 
130 |    5. Submission of Contributions. Unless You explicitly state otherwise,
131 |       any Contribution intentionally submitted for inclusion in the Work
132 |       by You to the Licensor shall be under the terms and conditions of
133 |       this License, without any additional terms or conditions.
134 |       Notwithstanding the above, nothing herein shall supersede or modify
135 |       the terms of any separate license agreement you may have executed
136 |       with Licensor regarding such Contributions.
137 | 
138 |    6. Trademarks. This License does not grant permission to use the trade
139 |       names, trademarks, service marks, or product names of the Licensor,
140 |       except as required for reasonable and customary use in describing the
141 |       origin of the Work and reproducing the content of the NOTICE file.
142 | 
143 |    7. Disclaimer of Warranty. Unless required by applicable law or
144 |       agreed to in writing, Licensor provides the Work (and each
145 |       Contributor provides its Contributions) on an "AS IS" BASIS,
146 |       WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
147 |       implied, including, without limitation, any warranties or conditions
148 |       of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
149 |       PARTICULAR PURPOSE. You are solely responsible for determining the
150 |       appropriateness of using or redistributing the Work and assume any
151 |       risks associated with Your exercise of permissions under this License.
152 | 
153 |    8. Limitation of Liability. In no event and under no legal theory,
154 |       whether in tort (including negligence), contract, or otherwise,
155 |       unless required by applicable law (such as deliberate and grossly
156 |       negligent acts) or agreed to in writing, shall any Contributor be
157 |       liable to You for damages, including any direct, indirect, special,
158 |       incidental, or consequential damages of any character arising as a
159 |       result of this License or out of the use or inability to use the
160 |       Work (including but not limited to damages for loss of goodwill,
161 |       work stoppage, computer failure or malfunction, or any and all
162 |       other commercial damages or losses), even if such Contributor
163 |       has been advised of the possibility of such damages.
164 | 
165 |    9. Accepting Warranty or Additional Liability. While redistributing
166 |       the Work or Derivative Works thereof, You may choose to offer,
167 |       and charge a fee for, acceptance of support, warranty, indemnity,
168 |       or other liability obligations and/or rights consistent with this
169 |       License. However, in accepting such obligations, You may act only
170 |       on Your own behalf and on Your sole responsibility, not on behalf
171 |       of any other Contributor, and only if You agree to indemnify,
172 |       defend, and hold each Contributor harmless for any liability
173 |       incurred by, or claims asserted against, such Contributor by reason
174 |       of your accepting any such warranty or additional liability.
175 | 
176 |    END OF TERMS AND CONDITIONS
177 | 
178 |    APPENDIX: How to apply the Apache License to your work.
179 | 
180 |       To apply the Apache License to your work, attach the following
181 |       boilerplate notice, with the fields enclosed by brackets "[]"
182 |       replaced with your own identifying information. (Don't include
183 |       the brackets!)  The text should be enclosed in the appropriate
184 |       comment syntax for the file format. We also recommend that a
185 |       file or class name and description of purpose be included on the
186 |       same "printed page" as the copyright notice for easier
187 |       identification within third-party archives.
188 | 
189 |    Copyright [yyyy] [name of copyright owner]
190 | 
191 |    Licensed under the Apache License, Version 2.0 (the "License");
192 |    you may not use this file except in compliance with the License.
193 |    You may obtain a copy of the License at
194 | 
195 |        http://www.apache.org/licenses/LICENSE-2.0
196 | 
197 |    Unless required by applicable law or agreed to in writing, software
198 |    distributed under the License is distributed on an "AS IS" BASIS,
199 |    WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
200 |    See the License for the specific language governing permissions and
201 |    limitations under the License.
202 |

--------------------------------------------------------------------------------