├── .gitattributes ├── .gitignore ├── .travis.yml ├── Dockerfile ├── Dockerfile-debug ├── LICENSE ├── README.md ├── demo.py ├── pics └── sample_report.png ├── setup.py └── stagesep2 ├── __init__.py ├── analyser ├── __init__.py ├── base.py ├── match_template.py ├── ocr.py └── trend.py ├── config.py ├── executor.py ├── loader.py ├── logger.py ├── painter.py ├── reporter.py └── utils.py /.gitattributes: -------------------------------------------------------------------------------- 1 | # Auto detect text files and perform LF normalization 2 | * text=auto 3 | -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | # Byte-compiled / optimized / DLL files 2 | __pycache__/ 3 | *.py[cod] 4 | *$py.class 5 | 6 | # C extensions 7 | *.so 8 | 9 | # Distribution / packaging 10 | .Python 11 | build/ 12 | develop-eggs/ 13 | dist/ 14 | downloads/ 15 | eggs/ 16 | .eggs/ 17 | lib/ 18 | lib64/ 19 | parts/ 20 | sdist/ 21 | var/ 22 | wheels/ 23 | *.egg-info/ 24 | .installed.cfg 25 | *.egg 26 | MANIFEST 27 | 28 | # PyInstaller 29 | # Usually these files are written by a python script from a template 30 | # before PyInstaller builds the exe, so as to inject date/other infos into it. 31 | *.manifest 32 | *.spec 33 | 34 | # Installer logs 35 | pip-log.txt 36 | pip-delete-this-directory.txt 37 | 38 | # Unit test / coverage reports 39 | htmlcov/ 40 | .tox/ 41 | .nox/ 42 | .coverage 43 | .coverage.* 44 | .cache 45 | nosetests.xml 46 | coverage.xml 47 | *.cover 48 | .hypothesis/ 49 | .pytest_cache/ 50 | 51 | # Translations 52 | *.mo 53 | *.pot 54 | 55 | # Django stuff: 56 | *.log 57 | local_settings.py 58 | db.sqlite3 59 | 60 | # Flask stuff: 61 | instance/ 62 | .webassets-cache 63 | 64 | # Scrapy stuff: 65 | .scrapy 66 | 67 | # Sphinx documentation 68 | docs/_build/ 69 | 70 | # PyBuilder 71 | target/ 72 | 73 | # Jupyter Notebook 74 | .ipynb_checkpoints 75 | 76 | # IPython 77 | profile_default/ 78 | ipython_config.py 79 | 80 | # pyenv 81 | .python-version 82 | 83 | # celery beat schedule file 84 | celerybeat-schedule 85 | 86 | # SageMath parsed files 87 | *.sage.py 88 | 89 | # Environments 90 | .env 91 | .venv 92 | env/ 93 | venv/ 94 | ENV/ 95 | env.bak/ 96 | venv.bak/ 97 | 98 | # Spyder project settings 99 | .spyderproject 100 | .spyproject 101 | 102 | # Rope project settings 103 | .ropeproject 104 | 105 | # mkdocs documentation 106 | /site 107 | 108 | # mypy 109 | .mypy_cache/ 110 | .dmypy.json 111 | dmypy.json 112 | 113 | # Pyre type checker 114 | .pyre/ 115 | 116 | # custom 117 | .idea/ 118 | -------------------------------------------------------------------------------- /.travis.yml: -------------------------------------------------------------------------------- 1 | os: linux 2 | dist: xenial 3 | language: python 4 | sudo: true 5 | python: 6 | - "3.6" 7 | - "3.7" 8 | 9 | install: 10 | - sudo apt-get -qq update 11 | - sudo apt-get -y install tesseract-ocr tesseract-ocr-chi-sim libtesseract-dev libleptonica-dev pkg-config 12 | - sudo apt-get -y install libglib2.0 libsm6 libxrender1 libxext-dev 13 | - pip install Cython 14 | - pip install . 15 | 16 | script: 17 | - git clone https://github.com/williamfzc/stagesep2-sample.git 18 | - cd stagesep2-sample 19 | - python sample.py 20 | -------------------------------------------------------------------------------- /Dockerfile: -------------------------------------------------------------------------------- 1 | FROM python:3-slim 2 | 3 | USER root 4 | 5 | RUN apt-get update \ 6 | && apt-get -y install gcc build-essential \ 7 | && apt-get -y install tesseract-ocr tesseract-ocr-chi-sim libtesseract-dev libleptonica-dev pkg-config \ 8 | && apt-get -y install libglib2.0 libsm6 libxrender1 libxext-dev 9 | 10 | # install stagesep2 11 | WORKDIR /usr/src/app 12 | COPY . . 13 | RUN pip install . \ 14 | && apt-get purge -y --auto-remove gcc build-essential \ 15 | && apt-get clean \ 16 | && rm -rf /var/lib/apt/lists/* 17 | 18 | WORKDIR /root/stagesep2 19 | CMD ["bash"] 20 | -------------------------------------------------------------------------------- /Dockerfile-debug: -------------------------------------------------------------------------------- 1 | FROM python:3-slim 2 | 3 | USER root 4 | 5 | WORKDIR /usr/src/app 6 | COPY . . 7 | 8 | RUN apt-get update \ 9 | && apt-get -y install gcc build-essential tesseract-ocr tesseract-ocr-chi-sim libtesseract-dev libleptonica-dev pkg-config \ 10 | && apt-get -y install libglib2.0 libsm6 libxrender1 libxext-dev \ 11 | && pip install . \ 12 | && apt-get purge -y --auto-remove gcc build-essential \ 13 | && apt-get clean \ 14 | && rm -rf /var/lib/apt/lists/* 15 | 16 | WORKDIR /root/stagesep2 17 | CMD ["bash"] 18 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2018 William Feng 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | > 该项目处于低活跃状态，建议使用 https://github.com/williamfzc/stagesepx 2 | 3 | # stagesep2 4 | 5 | ![PyPI - Python Version](https://img.shields.io/pypi/pyversions/stagesep2.svg?style=flat-square) 6 | [![PyPI version](https://badge.fury.io/py/stagesep2.svg)](https://badge.fury.io/py/stagesep2) 7 | [![Build Status](https://travis-ci.org/williamfzc/stagesep2.svg?branch=master)](https://travis-ci.org/williamfzc/stagesep2) 8 | [![Maintainability](https://api.codeclimate.com/v1/badges/ad4c729fcf98d98497b8/maintainability)](https://codeclimate.com/github/williamfzc/stagesep2/maintainability) 9 | 10 | --- 11 | 12 | Analyse, and convert video into useful data. 13 | 14 | # 原理 15 | 16 | stagesep2 的两个核心功能： 17 | 18 | - ocr（tesseract） 19 | - 图像识别（opencv） 20 | 21 | 视频是由一系列连续的图片（帧）组成的，通过逐帧分析，我们能够从视频中提取出流程相关信息。理论上，每一帧都存在能够区分它与其他帧的标识，可能是文字或图像。 22 | 23 | 例如，我们需要测试 从桌面启动chrome，打开amazon 的速度： 24 | 25 | - 在操作前，我们在主页。主页上会有特定的文字与图像（例如chrome icon） 26 | - 在操作时，页面会有特定的变化（例如chrome icon变暗，或出现点击位置反馈） 27 | - 在操作后（chrome启动后），页面发生切换，页面上的文字与图像都会发生改变（例如amazon logo出现） 28 | 29 | 那么，我们可以通过两个方面来进行检测。首先是通过标志性图片： 30 | 31 | ![](pics/sample_report.png) 32 | 33 | 可以看到，在 0.2s 时chrome图标被点击，在 0.96s 时amazon图标正式出现。再看看文字的： 34 | 35 | ```json 36 | { 37 | "result_id": "c2e5116c-462b-11e9-9ed2-005056c00008", 38 | "video_name": "./videos/demo1.mp4", 39 | "frame_id": 24, 40 | "current_time": 0.96, 41 | "ocr": ["PO", "/", "1dX00d", "¥", ":", "00Xv00Yv00", "—", "ESRRGRERSize0.0215", ":", "27dSlsB", "(", "ayhttps", "/", "/", "www", ".", "amazon", ".", "comQO", "°", "oa", "\"", "四", "Ney", ",", "DepartmentsListsDealsVideoMusic", "©", "DelivertoChinaWeshipinternationallyWe", "'", "reshowingyouitemsthatshiptoChina", ".", "Toseeiemsthatshiptoadifferentcountrychangeyourgetiveryaddres5AdditionallanguagesettingsareavailableSomecontentmaybeauto", "-", "translatedCHANGETHEADDRESSrs", "并", "~", "Shopwith100%", "ConfidenceonAmazonLJ", "—", ")", "SigninforthebestexperienceCoO00", "@"], 42 | "match_template": { 43 | "amazon": { 44 | "min": -0.4684264361858368, 45 | "max": 0.6224471926689148 46 | }, 47 | "chrome": { 48 | "min": -0.4022962152957916, 49 | "max": 0.7294253706932068 50 | }, 51 | "chrome_clicked": { 52 | "min": -0.6132965087890625, 53 | "max": 0.7038567066192627 54 | } 55 | }, 56 | "trend": { 57 | "previous": 0.8785946933890821, 58 | "first": 0.8719320065296263, 59 | "last": 0.5842399940047383 60 | } 61 | } 62 | ``` 63 | 64 | 节选 0.96s 的检测结果，可以发现在该时刻amazon网页上的相关字样开始出现，与图片检测的结果是一致的。 65 | 66 | 通过对这些阶段进行分析，得到每个阶段及帧对应的时间戳，我们就能够准确地知晓视频每个阶段发生的事情。你的视频FPS越高，数据会越精确。 67 | 68 | # 目的 69 | 70 | > 值得注意的是，在性能测试范畴内，fps的稳定性是非常重要的。而目前软件录制手段实际上并不稳定，容易造成结果不准确的问题。如果用于生产环境，建议使用硬件方式（例如外置摄像头）。 71 | 72 | 全平台的 性能测试/有效性验证 方案 73 | 74 | - 渲染内容是否符合期望 75 | - 渲染性能 76 | 77 | ## 为什么介入图像识别 78 | 79 | ### UI 80 | 81 | 现阶段的UI测试大多属于纯代码层面的行为，而对于控件是否真的渲染成为我们希望的样子我们并不知晓。 82 | 83 | ### 性能 84 | 85 | 在常规速度类性能测试中通常通过提前埋点进行测试，一般会有两个问题： 86 | 87 | - 具有侵入性（需要改动源码） 88 | - 对于界面相关的场景不适用（并不知道界面是否已经被真正渲染出来） 89 | 90 | ## 图像识别在测试中的应用 91 | 92 | 一般来说，通过图像识别来进行测试分为三个步骤： 93 | 94 | - 图像/视频 采集 95 | - 这个部分通常由高速摄像机或稳定帧率的外置相机进行拍摄，得到固定帧率的视频 96 | - 软件录制是不靠谱的，很容易出现帧率不稳定的情况。而如果时间与帧数不能精确对应的话数据会失真 97 | 98 | - 视频处理 99 | - 提取视频中的信息，输出成为我们需要的形式 100 | - 也是整个流程最关键的部分 101 | 102 | - 数据分析 103 | - 将视频处理的结果进行分析，得到结论或生成报告 104 | 105 | 该项目将承载视频处理的部分，将录制好的视频解析成开发者需要的格式。 106 | 107 | # 使用 108 | 109 | 目前支持两种使用方式： 110 | 111 | - docker（推荐） 112 | - python包 113 | 114 | 从 [官方示例](https://github.com/williamfzc/stagesep2-sample) 开始吧。 115 | 116 | # 相关内容 117 | 118 | ## 依赖 119 | 120 | - [opencv](https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_tutorials.html): 图像与视频处理 121 | - [tesseract-ocr](https://github.com/tesseract-ocr/tesseract/wiki/Downloads) & [tesserocr](https://github.com/sirfz/tesserocr): 文本检测 122 | - [skimage](https://github.com/scikit-image/scikit-image): 主要用于图片相似度比较 123 | - [jieba](https://github.com/fxsjy/jieba)：ocr结果的进一步处理 124 | - [pyecharts](https://github.com/pyecharts/pyecharts)：结果图表渲染 125 | 126 | ## 旧版本 127 | 128 | - [stagesep](https://github.com/williamfzc/stagesep) 129 | - [利用图像识别与 OCR 进行速度类测试](https://testerhome.com/topics/16063) 130 | 131 | # Bug与建议 132 | 133 | 欢迎通过 issue 告知，或直接发起 PR 加入 :) 134 | 135 | # 协议 136 | 137 | [MIT](LICENSE) 138 | -------------------------------------------------------------------------------- /demo.py: -------------------------------------------------------------------------------- 1 | # :) 2 | from stagesep2 import VideoManager, AnalysisRunner, NormalConfig, OCRConfig, MatchTemplateConfig 3 | import cv2 4 | 5 | 6 | # 添加待测视频 7 | # 虽然能够支持多视频，但还是建议每次只分析一个 8 | video = VideoManager.add('./temp.avi') 9 | 10 | # 旋转视频 11 | # 因为拍摄设备的差异，拍摄的视频可能方向是不同预期的，容易导致分析器分析不出有效结果 12 | # 设定该参数能够旋转视频，使分析器可以正常生效 13 | # 逆时针旋转 90 * rotate 度 14 | video.rotate = 3 15 | 16 | # 添加match template的样本图 17 | video.template_manager.add('./template.png') 18 | 19 | # 分析器列表 20 | # 默认情况下会全选，使用 OCR/模型匹配/首尾帧相似度 进行处理 21 | NormalConfig.ANALYSER_LIST = ['ocr', 'match_template', 'trend'] 22 | 23 | # 修改OCR的语种 24 | # 默认情况下是英文（tesseract自带了英文） 25 | # 如果使用中文，则需要自行安装tesseract的中文支持包，详见tesseract wiki： 26 | # https://github.com/tesseract-ocr/tesseract/wiki 27 | OCRConfig.lang = 'eng' 28 | # 设置方法与tesseract保持一致，其他语言请参考官方文档 29 | # 这是简体中文的例子 30 | # OCRConfig.lang = 'chi_sim' 31 | 32 | # 修改match template的算法 33 | # 此处直接使用了opencv提供的matchTemplate 34 | # 可参考 https://docs.opencv.org/master/d4/dc6/tutorial_py_template_matching.html 35 | MatchTemplateConfig.cv_method = cv2.TM_SQDIFF_NORMED 36 | 37 | # 启动分析 38 | result = AnalysisRunner.run() 39 | 40 | # 分析的结果（dict） 41 | dict_data = result.data 42 | 43 | # 或者输出到文件内（json） 44 | result.export('./result.json') 45 | 46 | # 绘制结果报告 47 | result.draw('./result_report.html') 48 | -------------------------------------------------------------------------------- /pics/sample_report.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/williamfzc/stagesep2/20c00187c86e8b807bbb3373f58918a575c2ccc9/pics/sample_report.png -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- 1 | from setuptools import setup, find_packages 2 | 3 | 4 | setup( 5 | name='stagesep2', 6 | version='0.2.6', 7 | description='Analyse, and convert video into useful data.', 8 | author='williamfzc', 9 | author_email='fengzc@vip.qq.com', 10 | url='https://github.com/williamfzc/stagesep2', 11 | packages=find_packages(), 12 | license='MIT', 13 | classifiers=[ 14 | 'License :: OSI Approved :: MIT License', 15 | 'Programming Language :: Python', 16 | 'Programming Language :: Python :: 3', 17 | 'Programming Language :: Python :: 3.6', 18 | 'Programming Language :: Python :: 3.7', 19 | ], 20 | python_requires=">=3.6", 21 | install_requires=[ 22 | 'opencv-python', 23 | 'structlog', 24 | 'numpy', 25 | 'jieba', 26 | 'scikit-image', 27 | 'pyecharts==0.5.11', 28 | 'pyecharts_snapshot', 29 | 'findit', 30 | 'tesserocr', 31 | 'Pillow', 32 | ] 33 | ) 34 | -------------------------------------------------------------------------------- /stagesep2/__init__.py: -------------------------------------------------------------------------------- 1 | from stagesep2.loader import VideoManager 2 | from stagesep2.executor import AnalysisRunner 3 | from stagesep2.config import NormalConfig, OCRConfig, MatchTemplateConfig 4 | -------------------------------------------------------------------------------- /stagesep2/analyser/__init__.py: -------------------------------------------------------------------------------- 1 | """ 2 | analyser 3 | 4 | opencv frame -> result 5 | """ 6 | from stagesep2.analyser.ocr import OCRAnalyser 7 | from stagesep2.analyser.match_template import MatchTemplateAnalyser 8 | from stagesep2.analyser.trend import TrendAnalyser 9 | 10 | 11 | ANALYSER_DICT = { 12 | 'ocr': OCRAnalyser, 13 | 'match_template': MatchTemplateAnalyser, 14 | 'trend': TrendAnalyser, 15 | } 16 | -------------------------------------------------------------------------------- /stagesep2/analyser/base.py: -------------------------------------------------------------------------------- 1 | class BaseAnalyser(object): 2 | name = '' 3 | 4 | @classmethod 5 | def run(cls, *args, **kwargs): 6 | """ 7 | normally, 8 | 9 | 1. frame (opencv image object) 10 | 2. ssv (provide necessary info, eg: template pictures) 11 | 12 | :param args: 13 | :param kwargs: 14 | :return: 15 | """ 16 | return '' 17 | 18 | @classmethod 19 | def clean(cls, *args, **kwargs): 20 | """ 21 | clean Analyser config after usage 22 | 23 | :param args: 24 | :param kwargs: 25 | :return: 26 | """ 27 | -------------------------------------------------------------------------------- /stagesep2/analyser/match_template.py: -------------------------------------------------------------------------------- 1 | from stagesep2.analyser.base import BaseAnalyser 2 | from stagesep2.config import MatchTemplateConfig 3 | from findit import FindIt 4 | 5 | fi = FindIt( 6 | engine=['template'], 7 | cv_method_name=MatchTemplateConfig.cv_method, 8 | ) 9 | 10 | 11 | class MatchTemplateAnalyser(BaseAnalyser): 12 | """ match-template analyser """ 13 | name = 'match_template' 14 | 15 | @classmethod 16 | def run(cls, frame, ssv): 17 | template_pic_dict = ssv.template_manager.get_dict() 18 | for each_pic_name, each_pic in template_pic_dict.items(): 19 | fi.load_template(each_pic_name, pic_object=each_pic.cv_object) 20 | 21 | match_result = fi.find( 22 | 'temp', 23 | target_pic_object=frame, 24 | engine_template_scale=(1, 2, 5) 25 | )['data'] 26 | 27 | result_dict = dict() 28 | for each_pic_name, each_pic_result in match_result.items(): 29 | result_dict[each_pic_name] = { 30 | # no min any more 31 | 'min': -1, 32 | 'max': each_pic_result['TemplateEngine']['target_sim'], 33 | } 34 | fi.clear() 35 | return result_dict 36 | -------------------------------------------------------------------------------- /stagesep2/analyser/ocr.py: -------------------------------------------------------------------------------- 1 | import jieba 2 | from PIL import Image 3 | from tesserocr import PyTessBaseAPI 4 | 5 | from stagesep2.analyser.base import BaseAnalyser 6 | from stagesep2.config import OCRConfig 7 | 8 | 9 | def content_filter(old_content): 10 | """ remove unused content and rebuild a word list """ 11 | new_content = old_content.replace(' ', '').replace('\n', '').replace('\r', '') 12 | return list(jieba.cut(new_content)) 13 | 14 | 15 | class OCRAnalyser(BaseAnalyser): 16 | """ ocr analyser """ 17 | 18 | name = 'ocr' 19 | TAG = 'OCRAnalyser' 20 | tesserocr_api = PyTessBaseAPI(lang=OCRConfig.lang) 21 | 22 | @classmethod 23 | def run(cls, frame, *args, **kwargs): 24 | """ 25 | run ocr analyser 26 | 27 | :param frame: ndarray 28 | :return: 29 | """ 30 | frame = Image.fromarray(frame) 31 | cls.tesserocr_api.SetImage(frame) 32 | result = cls.tesserocr_api.GetUTF8Text() 33 | 34 | # content filter 35 | result = content_filter(result) 36 | 37 | return result 38 | -------------------------------------------------------------------------------- /stagesep2/analyser/trend.py: -------------------------------------------------------------------------------- 1 | from skimage.measure import compare_ssim 2 | 3 | from stagesep2.analyser.base import BaseAnalyser 4 | 5 | 6 | class TrendAnalyser(BaseAnalyser): 7 | """ trend analyser """ 8 | name = 'trend' 9 | previous_frame = None 10 | 11 | @classmethod 12 | def run(cls, frame, ssv): 13 | """ 14 | return a dict, eg: {'previous': 1.0, 'first': 0.9, 'last': 0.2} 15 | 16 | :param frame: 17 | :param ssv: 18 | :return: 19 | """ 20 | # init previous frame 21 | if cls.previous_frame is None: 22 | cls.previous_frame = frame 23 | 24 | # calculate sim 25 | previous_sim = compare_ssim(cls.previous_frame, frame) 26 | first_sim = compare_ssim(ssv.first_frame, frame) 27 | last_sim = compare_ssim(ssv.last_frame, frame) 28 | 29 | # update previous frame 30 | cls.previous_frame = frame 31 | 32 | result_dict = { 33 | 'previous': previous_sim, 34 | 'first': first_sim, 35 | 'last': last_sim, 36 | } 37 | return result_dict 38 | 39 | @classmethod 40 | def clean(cls, *args, **kwargs): 41 | cls.previous_frame = None 42 | -------------------------------------------------------------------------------- /stagesep2/config.py: -------------------------------------------------------------------------------- 1 | """ 2 | 分析配置存放在此处 3 | 支持读取外部json配置 4 | """ 5 | import os 6 | import cv2 7 | 8 | 9 | class OCRConfig(object): 10 | lang = 'eng+chi_sim' 11 | 12 | 13 | class MatchTemplateConfig(object): 14 | cv_method = 'cv2.TM_CCOEFF_NORMED' 15 | 16 | 17 | class NormalConfig(object): 18 | # project root path 19 | PROJECT_PATH = os.path.dirname(os.path.dirname(__file__)) 20 | 21 | # default analyser list (will run) 22 | ANALYSER_LIST = ['ocr', 'match_template', 'trend'] 23 | 24 | # default encoding 25 | CHARSET = 'utf-8-sig' 26 | -------------------------------------------------------------------------------- /stagesep2/executor.py: -------------------------------------------------------------------------------- 1 | """ 2 | 分析器 3 | 4 | - OCR 5 | - 特征识别 6 | """ 7 | import cv2 8 | 9 | from stagesep2.loader import VideoManager, frame_prepare, SSVideo 10 | from stagesep2.config import NormalConfig 11 | from stagesep2.logger import logger 12 | from stagesep2.reporter import ResultReporter, ResultRow 13 | from stagesep2.utils import * 14 | 15 | 16 | class AnalysisRunner(object): 17 | """ 18 | 主要逻辑 19 | 20 | - 从VideoManager中导入视频对象 21 | - 从config中读取需要使用的Analyser 22 | - 遍历视频列表 23 | - 切割视频，遍历帧 24 | - 用不同的Analyser分析帧 25 | - 记录结果 26 | - 将结果传递给reporter进行处理 27 | """ 28 | TAG = 'AnalyserRunner' 29 | 30 | @classmethod 31 | def run(cls) -> (ResultReporter, list): 32 | analyser_list = check_analyser(NormalConfig.ANALYSER_LIST) 33 | video_dict = VideoManager.video_dict 34 | logger.info(cls.TAG, analyser=analyser_list, video=video_dict) 35 | 36 | result_reporter_list = list() 37 | for each_video_name, each_ssv in video_dict.items(): 38 | result_reporter = ResultReporter(each_video_name) 39 | cls.analyse_video(each_ssv, analyser_list, result_reporter) 40 | result_reporter_list.append(result_reporter) 41 | 42 | # export result 43 | # 如果同时分析多个视频，最终结果是一个装着Reporter的list 44 | # 如果只分析一个视频，最终结果是一个Reporter 45 | if len(result_reporter_list) <= 1: 46 | return result_reporter_list[0] 47 | return result_reporter_list 48 | 49 | @classmethod 50 | def analyse_video(cls, ssv_video: SSVideo, analyser_list: list, result_reporter: ResultReporter): 51 | """ analyse ssv video """ 52 | with video_capture(ssv_video) as each_video: 53 | ret, frame = each_video.read() 54 | while ret: 55 | if not ret: 56 | # end of video 57 | break 58 | 59 | # prepare frame 60 | frame = frame_prepare(frame) 61 | # rotate 62 | frame = rotate_pic(frame, ssv_video.rotate) 63 | 64 | # current status 65 | cur_frame_count = int(each_video.get(cv2.CAP_PROP_POS_FRAMES)) 66 | cur_second = each_video.get(cv2.CAP_PROP_POS_MSEC) / 1000 67 | 68 | # new row of result 69 | new_row = ResultRow( 70 | result_reporter.result_id, 71 | ssv_video.video_path, 72 | cur_frame_count, 73 | cur_second, 74 | ) 75 | 76 | for each_analyser in analyser_list: 77 | result = each_analyser.run(frame, ssv_video) 78 | new_row.add_analyser_result(each_analyser.name, result) 79 | 80 | logger.info(cls.TAG, msg='analysing', **new_row.__dict__) 81 | result_reporter.add_row(new_row) 82 | 83 | # read new frame 84 | ret, frame = each_video.read() 85 | 86 | # clean analyser 87 | for each in analyser_list: 88 | each.clean() 89 | -------------------------------------------------------------------------------- /stagesep2/loader.py: -------------------------------------------------------------------------------- 1 | """ 2 | 负责资源导入、模块检查 3 | """ 4 | import os 5 | import cv2 6 | 7 | from stagesep2.logger import logger 8 | from stagesep2.utils import * 9 | 10 | 11 | def path_to_name(file_path: str) -> str: 12 | """ full path -> file name """ 13 | return os.path.splitext(os.path.basename(file_path))[0] 14 | 15 | 16 | def is_path_existed(file_path: str): 17 | """ check if file is existed """ 18 | return os.path.isfile(file_path) 19 | 20 | 21 | def frame_prepare(frame): 22 | gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) 23 | blur_gray_frame = cv2.medianBlur(gray_frame, 3) 24 | return blur_gray_frame 25 | 26 | 27 | class TemplatePicture(object): 28 | def __init__(self, pic_path: str): 29 | if not is_path_existed(pic_path): 30 | raise FileNotFoundError('file not existed: {}'.format(pic_path)) 31 | 32 | self.pic_name: str = path_to_name(pic_path) 33 | self.pic_path: str = pic_path 34 | self.cv_object = frame_prepare(cv2.imread(self.pic_path)) 35 | 36 | 37 | class TemplateManager(object): 38 | TAG = 'TemplateManager' 39 | 40 | def __init__(self, video_name: str): 41 | self.video_name = video_name 42 | 43 | # match template 需要模板图片 44 | # 该视频需要的模板图片会被放置在此处 45 | self._match_template_pic_dict = dict() 46 | # eg: 47 | # { pic_name: TemplatePicture(pic_path), } 48 | 49 | def add(self, pic_path: str): 50 | new_pic = TemplatePicture(pic_path) 51 | new_pic_name = new_pic.pic_name 52 | self._match_template_pic_dict[new_pic_name] = new_pic 53 | logger.info(self.TAG, msg='LOAD PICTURE', path=pic_path, name=new_pic_name, video=self.video_name) 54 | 55 | def remove(self, pic_name: str): 56 | if pic_name in self._match_template_pic_dict: 57 | del self._match_template_pic_dict[pic_name] 58 | return True 59 | logger.warn(self.TAG, msg='no pic named {}'.format(pic_name)) 60 | return False 61 | 62 | def get_dict(self): 63 | return self._match_template_pic_dict 64 | 65 | 66 | class SSVideo(object): 67 | """ video object """ 68 | 69 | def __init__(self, video_path: str): 70 | if not is_path_existed(video_path): 71 | raise FileNotFoundError('file not existed: {}'.format(video_path)) 72 | 73 | self.video_name = path_to_name(video_path) 74 | self.video_path = video_path 75 | 76 | # add template example: 77 | # ssv = SSVideo('some_path/123.mp4') 78 | # ssv.template_manager.add('some_path/123.png') 79 | self.template_manager = TemplateManager(self.video_name) 80 | 81 | # degree = rotate * 90, 逆时针 82 | self._rotate = 0 83 | 84 | # video info 85 | # total frame count 86 | self.total_frame = None 87 | 88 | # first and last frame 89 | self.first_frame = None 90 | self.last_frame = None 91 | 92 | # load video base info 93 | self.load_video_info() 94 | 95 | @property 96 | def rotate(self): 97 | return self._rotate 98 | 99 | @rotate.setter 100 | def rotate(self, value: int): 101 | if not isinstance(value, int): 102 | raise TypeError('rotate should be int') 103 | self._rotate = value 104 | self.load_video_info() 105 | 106 | def load_video_info(self): 107 | # TODO need more info? 108 | # get info from video 109 | with video_capture(self) as video_src: 110 | total_frame = video_src.get(cv2.CAP_PROP_FRAME_COUNT) 111 | 112 | _, first_frame = video_src.read() 113 | video_src.set(1, total_frame - 1) 114 | _, last_frame = video_src.read() 115 | 116 | # prepare, and rotate 117 | first_frame, last_frame = [ 118 | frame_prepare(rotate_pic(each, self._rotate)) 119 | for each in (first_frame, last_frame)] 120 | 121 | # init 122 | self.first_frame = first_frame 123 | self.last_frame = last_frame 124 | self.total_frame = total_frame 125 | 126 | 127 | class VideoManager(object): 128 | """ 129 | Analyser需要的信息都应该在此处被导入 130 | 例如 作为分析主体的 视频 131 | 例如 match template需要的 模板图片 132 | """ 133 | TAG = 'VideoManager' 134 | 135 | # 待测视频会被添加到这里 136 | # 在分析开始时，会遍历此字典 137 | video_dict = dict() 138 | 139 | # eg: 140 | # { video_name: SSVideo(video_path), } 141 | 142 | def __init__(self): 143 | raise NotImplementedError('should not init') 144 | 145 | @classmethod 146 | def add(cls, video_path: str): 147 | new_video = SSVideo(video_path) 148 | new_video_name = new_video.video_name 149 | cls.video_dict[new_video_name] = new_video 150 | logger.info(cls.TAG, msg='LOAD VIDEO', path=video_path, name=new_video_name) 151 | return new_video 152 | 153 | @classmethod 154 | def remove(cls, video_name: str): 155 | if video_name in cls.video_dict: 156 | del cls.video_dict[video_name] 157 | return True 158 | logger.warn(cls.TAG, msg='no video named {}'.format(video_name)) 159 | return False 160 | -------------------------------------------------------------------------------- /stagesep2/logger.py: -------------------------------------------------------------------------------- 1 | from structlog import getLogger 2 | 3 | 4 | logger = getLogger(__name__) 5 | -------------------------------------------------------------------------------- /stagesep2/painter.py: -------------------------------------------------------------------------------- 1 | import json 2 | from pyecharts import Line, Page 3 | 4 | from stagesep2.logger import logger 5 | from stagesep2.config import NormalConfig 6 | 7 | 8 | # 如果后续图表需要更加复杂，这个部分最好用类似pandas的库进行比较规范的数据格式处理。只是简单展示就不用了 9 | class ReportPainter(object): 10 | TAG = 'Painter' 11 | 12 | @classmethod 13 | def draw_with_file(cls, file_path: str, dst_path: str): 14 | """ draw with json file path, and export to dst path """ 15 | with open(file_path, encoding=NormalConfig.CHARSET) as f: 16 | cls.draw(json.load(f), dst_path) 17 | 18 | @classmethod 19 | def draw_with_json(cls, json_str: str, dst_path: str): 20 | """ draw with raw json content, and export to dst path """ 21 | content = json.loads(json_str) 22 | cls.draw(content, dst_path) 23 | 24 | @classmethod 25 | def draw(cls, content: list, dst_path: str): 26 | # base page 27 | page = Page() 28 | 29 | # time line，直接用float作为x轴会出现 echarts 兼容问题 30 | time_list = [str(each['current_time']) for each in content] 31 | 32 | # 目前只支持 match_template 与 trend 的图表绘制 33 | draw_type_dict = { 34 | 'match_template': cls.build_match_template_line, 35 | 'trend': cls.build_trend_line, 36 | } 37 | for each_type, each_func in draw_type_dict.items(): 38 | # 以第一项为例检验是否包含该类分析结果 39 | if each_type not in content[0]: 40 | continue 41 | 42 | data_list = [each_data[each_type] for each_data in content] 43 | data_line = draw_type_dict[each_type](time_list, data_list) 44 | page.add(data_line) 45 | 46 | page.render(dst_path) 47 | logger.info(cls.TAG, msg='report built finished: "{}"'.format(dst_path)) 48 | 49 | @classmethod 50 | def build_trend_line(cls, time_list: list, trend_list: list): 51 | trend_line = Line('trend') 52 | for each_attr in ['previous', 'first', 'last']: 53 | each_list = [i[each_attr] for i in trend_list] 54 | trend_line.add( 55 | each_attr, 56 | time_list, 57 | each_list, 58 | yaxis_min='dataMin', 59 | is_more_utils=True, 60 | ) 61 | return trend_line 62 | 63 | @classmethod 64 | def build_match_template_line(cls, time_list: list, match_template_list: list): 65 | """ 66 | 67 | :param time_list: 68 | :param match_template_list: 69 | 70 | looks like: 71 | 72 | [ 73 | { 74 | "pic1": { 75 | "min": -0.4684264361858368, 76 | "max": 0.6224471926689148 77 | }, 78 | "pic2": { 79 | "min": -0.4022962152957916, 80 | "max": 0.7294253706932068 81 | }, 82 | "pic3": { 83 | "min": -0.6132965087890625, 84 | "max": 0.7038567066192627 85 | } 86 | } 87 | 88 | ... 89 | ] 90 | 91 | :return: 92 | """ 93 | 94 | # build data structure for drawing 95 | data_to_draw = { 96 | name: list() 97 | for name in match_template_list[0].keys() 98 | } 99 | 100 | for each_name in data_to_draw.keys(): 101 | # draw only max values 102 | data_to_draw[each_name] = [i[each_name]['max'] for i in match_template_list] 103 | 104 | match_template_line = Line('match_template') 105 | for each_name, each_data in data_to_draw.items(): 106 | match_template_line.add( 107 | each_name, 108 | x_axis=time_list, 109 | y_axis=each_data, 110 | is_more_utils=True, 111 | ) 112 | 113 | return match_template_line 114 | -------------------------------------------------------------------------------- /stagesep2/reporter.py: -------------------------------------------------------------------------------- 1 | """ 2 | 处理分析结果 3 | 4 | 应至少支持： 5 | - 文件（json） 6 | - 对象 7 | - 数据库交互 8 | """ 9 | import uuid 10 | import json 11 | import os 12 | 13 | from stagesep2.logger import logger 14 | from stagesep2.painter import ReportPainter 15 | 16 | 17 | class ResultRow(object): 18 | def __init__( 19 | self, 20 | # must 21 | result_id, 22 | video_name, 23 | frame_id, 24 | current_time, 25 | ): 26 | # task id 27 | self.result_id = result_id 28 | # video name (id), a task may contains more than one video 29 | self.video_name = video_name 30 | # current frame number 31 | # 当前的结果对应的帧编号 32 | self.frame_id = frame_id 33 | # current time 34 | # 当前的结果对应的帧，在视频中对应的时间 35 | self.current_time = current_time 36 | 37 | def add_analyser_result(self, name, result): 38 | """ 39 | add result of analyser 40 | 41 | :param name: analyser name 42 | :param result: result of analyser 43 | :return: 44 | """ 45 | # analyse result 46 | # ocr: list, eg: ['some_word', 'other_word'] 47 | # match_template: dict, eg: {'pic1': 0.85, 'pic2': 0.90} 48 | 49 | self.__dict__[name] = result 50 | 51 | def __str__(self): 52 | return json.dumps(self.__dict__) 53 | 54 | def to_dict(self): 55 | return self.__dict__ 56 | 57 | __repr__ = __str__ 58 | 59 | 60 | class ResultReporter(object): 61 | TAG = 'ResultReporter' 62 | 63 | def __init__(self, video_name): 64 | self.video_name = video_name 65 | self.result_id = str(uuid.uuid1()) 66 | self._row_list = list() 67 | 68 | def add_row(self, new_row): 69 | self._row_list.append(new_row) 70 | 71 | def export(self, file_path): 72 | """ export result to json file. Path can be file, dir. """ 73 | 74 | # check file path 75 | if os.path.isfile(file_path): 76 | logger.warn(self.TAG, msg='File "{}" already existed'.format(file_path)) 77 | file_path = os.path.join(os.path.dirname(file_path), self.result_id + '.json') 78 | elif os.path.isdir(file_path): 79 | logger.warn(self.TAG, msg='Path "{}" is a directory'.format(file_path)) 80 | file_path = os.path.join(file_path, self.result_id + '.json') 81 | 82 | # write file 83 | with open(file_path, 'w+') as json_file: 84 | json_file.write(str(self.data)) 85 | logger.info(self.TAG, msg='Result saved in "{}"'.format(file_path)) 86 | 87 | @property 88 | def data(self): 89 | """ return data, consisted by pyobject """ 90 | return self._row_list 91 | 92 | def draw(self, dst): 93 | """ draw analysis report to file named dst """ 94 | ReportPainter.draw_with_json(str(self.data), dst) 95 | -------------------------------------------------------------------------------- /stagesep2/utils.py: -------------------------------------------------------------------------------- 1 | """ 2 | public functions 3 | """ 4 | import numpy as np 5 | import cv2 6 | import contextlib 7 | 8 | from stagesep2.analyser import ANALYSER_DICT 9 | 10 | 11 | def check_analyser(analyser_list: list) -> list: 12 | """ check if analyser existed, and return list of runnable analyser """ 13 | new_analyser_list = list() 14 | for each in analyser_list: 15 | if each not in ANALYSER_DICT: 16 | raise NotImplementedError('analyser {} not found'.format(each)) 17 | new_analyser_list.append(ANALYSER_DICT[each]) 18 | return new_analyser_list 19 | 20 | 21 | def rotate_pic(old_pic: np.ndarray, rotate_time: int) -> np.ndarray: 22 | """ 帧逆时针旋转 90*rotate_time 度 """ 23 | new_pic = np.rot90(old_pic, rotate_time) 24 | return new_pic 25 | 26 | 27 | @contextlib.contextmanager 28 | def video_capture(ssv): 29 | """ 打开视频的上下文控制 """ 30 | video_cap = cv2.VideoCapture(ssv.video_path) 31 | yield video_cap 32 | video_cap.release() 33 | 34 | 35 | __all__ = [ 36 | 'check_analyser', 37 | 'rotate_pic', 38 | 'video_capture', 39 | ] 40 | --------------------------------------------------------------------------------