Screening Task 1: Video Processing

This is a Django application that allows a user to upload a video and its .srt file containing subtitles.The video is then broken into chunks,based on the timing information specified in the .srt file. The corresponding audio is extracted from each video chunk and stored separately.For the uploaded video the web app displays the start time, end time, video chunk sequence number,subtitles and its respective audio(.mp3). For each audio the use can reupload the edited .mp3 file for that video chunk.The app combines all the audio and video chunks into one a single video(.mp4) file when the use clicks on the compile button at the bottom of the page and provides a download link.

Table of Content:

• Installation
• Usage
• API
• videoProcessing module
• miscFunctions module
• Database
• Credits

Watch the video

Installation:

• It is recommended to use Linux/Unix based systems. windows based system can be used but the mysqlclient python package causes problems in windows.

• You can either clone the project or download as zip

• To clone the project you must have git installed on your computer.

• click on the Clone or download button and copy the url.

• open the command prompt(windows) or terminal(mac/linux) and type git clone [copied_url] or simply copy and past in terminal git clone https://github.com/rohitgeddam/fosseeVideoProcessing.git

You need to install few things before you can run the project.

• Python 3.x
• FFmpeg
• MySQL
• MySQL Workbench

After you have installed the above you can continue

• Install virtualenv by typing sudo pip3 install virtualenv in terminal.
• It is recommended to create a virtual environment to install all python packages.
• To create a virtual environment type virtualenv -p python3 venv in terminal
• To activate your virtual environment type source venv/bin/activate
• Move into the downloaded directory cd fosseeVideoProcessing/videoProcessingAPI/
• Install required python packages by typing pip install -r requirments.txt into the terminal
• NOTE-> installation of mysqlclient in windows systemss requires Microsoft Visual C++ 14.0. so install that first.

You need to change some settings to run the database

• Create a new schema in MySQL with any name you like e.g. fossee-task-1
• Open MySQL Workbench

• navigate to fosseeVideoProcessing/videoProcessingAPI/ and open settings.py file.
• make the following changes to the settings.py file
  1. Change DB_NAME to the name of the database you created (around line number 80)
  2. Change DB_PASSWORD to your database password
  3. likewise change DB_HOST and DB_PORT
• If you dont want to make the above changes you should create the database with the name fossee-task-1 and everything will work without any changes in settings.py file.
• Navigating to fosseeVideoProcessing/videoProcessingAPI/ and run command python manage.py migrate in terminal.
• Before running the project you can run the tests by navigating to fosseeVideoProcessing/videoProcessingAPI/ and typing python manage.py test
• Before you start the server make sure the MySQL server is running at port 3306 or the port that you decided while creating the database.
• You can now run the project by navigating to fosseeVideoProcessing/videoProcessingAPI/ and typing python manage.py runserver
• If you want to run the django server on some other port than 8000. then you need to change the LOCAL_HOST_URL_WITH_PORT constant in the settings.js file of the frontend application at location /fosseeVideoProcessing/videoProcessingFrontend/scripts/controllers/settings.js
• If everything goes well the server will be up and running

Usage:

• NOTE This app seperates Backend from the Frontend
• LOCATION OF FRONTEND APP fosseeVideoProcessing/videoProcessingFrontend/
• LOCATION OF BACKEND APP fosseeVideoProcessing/videoProcessingAPI/
• Make sure the Django server and MySQL database is up and running. if not follow the installation section properly.
• Now navigate to fosseeVideoProcessing/videoProcessingFrontend/
• open index.html in a browser (you need to have internet connection as it need to download some scripts from cdn)
• Screenshot below

• Select any video file (.mp4) and srt file (.srt) from your drive or you can use the ones in the /fosseeVideoProcessing/videoProcessingAPI/test_files/ (do not delete any files in test_files directory)
• Click the Upload Files button to upload your files to the server.
• After the File Upload is successfull Process Video button appears click on that button to process the video.

• After the processing is done. The details are listed.

• If you want to re-upload modified audio for a particular video chunk. Under AUDIO RE-UPLOAD column select the audio file and click upload button

• Click on the COMPILE button at the bottom of the page.

• When you click on the COMPILE button a DOWNLOAD TUTORIAL button is available to download the tutorial.

• You can click on the download button to download the tutorial or right-click and select save link as.

Api

Api Endpoints:

• api/upload/ [POST]

  This endpoint is used to upload files to the server.It only accepts (.mp4) and (.srt) files the uploaded files are stored temporarily before its processed in the media directory. It requires 2 files one is video file (.mp4) and the other is (.srt) file.Pass these using form data with the name 'video' and 'srt' respectively.

returns: JSON

operationId
message
operation_url

• api/process/<int:id> [GET]

  This endpoint is used to process the uploaded video.Requires (id) paramater which is the operation_id of the video

returns: JSON

message
time(sec)
path
downloadUrl
chunks

• api/getdetails/<int:id> [GET]

  This endpoint fetches the details of processed video.The (id) parameter is required which is the operation_id of the processed video.

  We can also get the details of previously processed video by specifying the operation_id of that video in parameter.

returns: JSON

message
downloadUrl
chunks

• api/reupload/<int:chunk_id> [POST]

This endpoint is used to reupload the audio file (.mp3).The file is uploaded using form field with the name 'file'. chunk_id is required as parameter.

This route is redirected to the getdetails route which returns JSON

message
downloadUrl
chunks

• api/download/<int:operationId> [GET]

  This endpoint compiles all the video and audio chunk into one file and provide a download link. requires operation_id of the video as parameter

returns:

message
name_of_file
download

API Testing:

To test the api run the following command.

• navigate to fosseeVideoProcessing/videoProcessingAPI/ and run command python manage.py test in the terminal.

VideoProcessing_module

• This is a local module which contains classes and methods usefull for videoprocessing
• location of module fosseeVideoProcessing/videoProcessingAPI/api/videoProcessing

Classes:

• SRT
• VideoClip

Methods in SRT class:

• numberOfChunksInSrt
• extractSrtData

Methods in VideoClip class:

• extractAudioFromVideoAndSave
• removeAudioFromVideoAndSave

Methods:

• mergeVideoAndAudioToGetDownloadFile
• mergeAudiosForDownload
• mergeVideoForDownload
• trimAudioClipAndSave
• getVideoLengthInSeconds
• convertTimeToSeconds

MiscFunctions_module

• This is a local module which contains utility or helper functions

• location of module fosseeVideoProcessing/videoProcessingAPI/api/miscFunctions

Functions in this module

• checkExtensionOfFileFromRequestObject
• handleUploadedFilesAndSave
• serializeObject
• removeAndCreateDirectoryInSamePath
• createDirectoryIfNotExists
• splitAudioAndVideoIntoChunk
• writePathsToTxtFileToUseWithFFMPEG

Database

The database used is MySQL server with MySql workbench.To interact with database within django app djangoORM is used.

Database Schema

Credits

• FFmpeg (License: GNU General Public License (GPL) (GPLv2))
• pysrt (License: GNU General Public License (GPL) (GPLv3))
• mysqlclient (License: GNU General Public License (GPL))
• django-rest-framework (License: BSD License (BSD))
• moviepy (License: MIT License (MIT License))
• pydub (License: MIT License (MIT))
• Django (License: BSD License (BSD))

Some websites to download videos and srt files

• youtube
• downsub
• savesubs