video-subtitle-process

A website on which video(s) can be uploaded, processed in some manner (in the background) and then searched using the subtitles in that video as keywords.

For instance, if a 2 minute clip of a music video was uploaded, the application will parse it, apply subtitles to it and ensure that searching for a particular word or phrase returns the time segment within which the video has those phrases being mentioned.

1. Used the ccextractor binary for extracting subtitles from video. Note that using any API, etc for subtitle extraction is not allowed.

2. Further, after processing, the videos will be stored in S3 and the search keywords (subtitles) in DynamoDB.

3. HTTP requests should has a maximum latency of approx 1 second.

4. The backend is written in Django. Celery for background tasks.

"Use this sample video file for running ccextractor”: https://drive.google.com/file/d/1gLi5ho33TIRNVZkSE-gD4S24zs6Xy1ci/view?usp=sharing

How to start

Demo video

• Clone the repo with git clone https://github.com/aghosh0605/video-subtitle-process.git
• cd into the folder cd video-subtitle-process
• Install the python dependencies pip install -r requirements.txt or python3 -m pip install -r requirements.txt
• cd into the Django project root directory cd video_sub_parser
• Run the below commands

sudo apt-get install ccextractor - Tested on ubuntu 20.04 LTS

python3 manage.py makemigrations #Need if you use SQL DB
python3 manage.py migrate  #Need if you use SQL DB
python3 manage.py runserver #TO start the server

For Nginx Reverse Proxy increase the upload file size and get rid of CORS errors, follow the below method

1. Open /etc/nginx/nginx.conf and modify the below parameter within http
2. client_max_body_size SIZE; SIZE Refers to max body size
3. sudo systemctl restart nginx To restart nginx server

Useful command snippet for Developers

source env/bin/activate #Enable python virtual environment
sudo apt install ccextractor #To install ccextractor binaries
pip install -r /path/to/requirements.txt #Install python requirements packages
pip3 freeze > requirements.txt #To install the python3 packages
ccextractor video_test.mp4 -out=ttxt #To extract subtitles from the video
sudo systemctl start redis-server.service # To start the redis service
celery -A video_sub_parser worker --loglevel=debug --concurrency=4  #To start celery instance

Important Links

• How to setup the Django Project in python virtual environment https://realpython.com/django-setup/
• Install Redis Server on Ubuntu https://redis.io/docs/getting-started/installation/install-redis-on-linux/

API Endpoints

Request

POST /api/v1/upload/file

Body	Type	Description
file	File Object	Required. File to Upload
remark	string	Required. Remark for File

Response

{
    "message":string,
    "filename":string,
    "task_id":string
}

Request

GET /api/v1/parse?file=filename

Body	Type	Description
filename	string	Required. Name of Uploaded File

Response

{
    "message":string
}

Request

GET /api/v1/subtitle/find?query=TO SUCCESS

Body	Type	Description
query	string	Required. Query to search

Response

{
    {
    "Items": [
        {
            "subtitle_url": string,
            "media_path": string,
            "SubtitleID": string,
            "subtitle": string,
            "start_time": string,
            "end_time": string
        }
    ],
    "Count": number,
    "ScannedCount": number,
    "ResponseMetadata": {
        "RequestId": string,
        "HTTPStatusCode": number,
        "HTTPHeaders": {
            "server": string,
            "date": string,
            "content-type": string,
            "content-length": string,
            "connection": string,
            "x-amzn-requestid": string,
            "x-amz-crc32": string
        },
        "RetryAttempts": number
    }
}
}

Request

PATCH /api/v1/status/file/task_id

Body	Type	Description
task_id	string	Required. Background Task ID from Celery

Response

boolean

Request

GET /progress-track/task_id

Body	Type	Description
task_id	string	Required. Background Task ID from Celery

Response

   {
    "state": string,
    "complete": boolean,
    "success": boolean,
    "progress": {
        "pending": boolean,
        "current": number,
        "total": number,
        "percent": number
    },
    "result": string | null
}