MyTube

video explanation

A video server serving local video files over local internet using HLS protocol along with many capabilities like searching using text-search engine etc.

How does this works

1. First step is to edit the config.json file. One needs to add the video directories to include. This file is not there, app should be run atleast one to generate this file.

{
  "version": "1.0.0",
  "vide_directory_absolute": [
    <video directory 1>
    <video directory 2>
  ],
  "max_HLS_folder_size_GB": 1.0
}

There maybe other files other than video like .txt, .mp3 etc, one can determine file type using the mimetype or the extension of the file. The extensions are mentioned in Constants.ACCEPT_FILE_FORMATS 2. Once the paths are added, program calculated the Md5 hash of each file. Whenever the server starts, it also creates a config-lock.json to preserver the last know image of config.json along with hash. Delete config-loc.json if one wants to reset everything. Program then calculates the diff namely

1. which files should be removed from indexing (db & text search engine) -- files removed by user
2. Which files should be inserted as a new file
3. which files should be renamed (Md5 hash same, maybe video renamed or moved by user)

3. For database TinyDB is used and for text search Whoosh is used. The data structure for the database is like

{'id': file_id,
 'file_name': ,
 'abs_path': ,
 'thumbnail_path': ,
 'duration': ,
 'width': ,
 'height': ,
 'format_name': ,
 'size': ,
 'v_bitrate': ,
 'a_bitrate': ,
 'is_favourite': ,
 'hls_processing': ,
 'hls_already_processed': ,
 'hls_process_location': ,
 'cluster_id': ,
 'views': 
 }

4. After updating the indexes, Thumbnails are generated for the required files using FFmpeg. make sure FFmpeg installed on the system. and stored in Constants.THUMBNAIL_FOLDER_PATH
5. Once it is done, server is ready to accept the requests.
6. Once a request is GET, server checks if the <file_name>.m3u8 available inside Constants.HLS_OUTPUT_PATH or not. If not it will start a new process to process the file and deliver the chunks to client accordingly.

Once processed, the next time the buffering will be less to none. Buffering can be also be affected by the resolution during conversion. This app is using deafult Representation(Size(640, 360), Bitrate(276 * 1024, 128 * 1024))

---

Architecture

---

Installing

1. Make sure FFmpeg is installed. I worked with Ubuntu, hope will work fine with windows als well.
2. Create virtual env python -m venv venv
3. Activate it ex: in ubuntu source ./venv/bin/activate
4. Install the required packages pip install -r requirements.txt
5. Start python app.py

I have not used any proper servers, it worked fine with Flask-test server for personal project, one can use gunicorn.