TubeScraper

This document explains how to set up a development environment, architecture of the application as well as how to use.

Overview

The below diagram show the overview of how the different systems are connected.

Getting started with development

• Clone this repo
• Create virtual environment (e.g. establish PyCharm project and set/create interpreter)

Install requirements

pip install -r requirements.txt

Google Cloud Platform

Make a new project in GCP and enable youtube API. Visit the Enabled APIs page. In the list of APIs, make sure the status is ON for the YouTube Data API v3. The Authentication is done using API Keys.

Visit https://developers.google.com/youtube/v3 for more info.

Docs available at https://developers.google.com/youtube/v3/docs

Environment variables

Set the value of following environment variable for local as well as deployment environment

AWS_MYSQL_HOST: Host address for your AWS MySQL database

AWS_MYSQL_MASTER_USER: MySQL User Name

AWS_MYSQL_MASTER_PASSWORD: MySQL Password

BUCKET_NAME: S3 Bucket Name

GCP_TUBESCRAPER_API_KEY: Youtube's API key to access youtube data.

MONGO_CONNECTION_STRING: Connection string of MongoDB client

S3_ACCESS_KEY_ID: User access key which have access to S3 Bucket

SECRET_ACCESS_KEY: User access key secret

ENABLE_CLOUD_DB : Boolean flag to specify whether to use cloud database to save data or not

MAINTENANCE_MODE = Boolean flag temporary disable the functionality

Database Schema

AWS RDS - MySql

1. Youtuber Table

   chanel_id	channel_name

2. Video Table

   chanel_id	video_id	title	youtube_link	s3_link	likes	video_id	comments_count	views	thumbnail_url

Create Table Query

  videos_create_table_query = f'CREATE TABLE `{MY_SQL_DATABASE}`.`{MY_SQL_VIDEOS_TABLE_NAME}` (
                              `index` INT NOT NULL AUTO_INCREMENT,
                              `channel_id` VARCHAR(45) NOT NULL,
                              `video_id` VARCHAR(45) NOT NULL,
                              `title` VARCHAR(200) NOT NULL,
                              `youtube_link` VARCHAR(200) NOT NULL,
                              `s3_link` VARCHAR(200) NULL,
                              `likes` INT NULL,
                              `comments_count` INT NULL,
                              `views` INT NULL,
                              `thumbnail_link` VARCHAR(200) NULL,
                              PRIMARY KEY (`index`),
                              UNIQUE INDEX `video_id_UNIQUE` (`video_id` ASC) VISIBLE);'
  
  
  youtuber_create_table_query = f'CREATE TABLE `{MY_SQL_DATABASE}`.`{MY_SQL_YOUTUBER_TABLE_NAME}` (
                                `channel_id` VARCHAR(45) NOT NULL,
                                `channel_name` VARCHAR(45) NOT NULL,
                                PRIMARY KEY (`channel_id`),
                                UNIQUE INDEX `channel_id_UNIQUE` (`channel_id` ASC) VISIBLE)'

MongoDB

  record = {
     'video_id'  # id from sql video table
     'thumbnail_base64'
     'comments': [
        {
           'comment_author'
           'message'
           'replies': [
              {
                 'comment_author'
                 'message'
              },
           ]
        },
     ]
  }

Run the application

python app.py

Deploy

Deployment is done on 3 platforms -

Heroku (decommissioned)

https://tubescraper.herokuapp.com/

AWS (decommissioned )

http://tubescraper-env.eba-kcbwec25.ap-south-1.elasticbeanstalk.com/

Azure (Live, Takes a while to wake-up from sleep mode.)

https://tubescraperwebapp.azurewebsites.net/

Reset Database Endpoint

As this is a hobby project, deployment is done on free tier of cloud platform. To keep the resources under limit, this endpoint deletes all data from all the databases as well as video files from S3 bucket.

  GET /admin/db/reset

Use Tubescraper

Browse to the home page of the application and enter the video url of any channel for which the data needs to be extracted

1. Video Url: Url of the video of a channel for which the data needs to be fetched. It supports the different url patters like below mention. It should be a youtube.com domain & must be a '/watch' endpoint with parameter 'v'.

/watch Endpoints -

  https://www.youtube.com/watch?v=QXeEoD0pB3E

  https://www.youtube.com/watch?v=FHfZ5X7qn1I&list=PLsyeobzWxl7pGh8x5C2hsu3My4ei-eX1Y

  https://www.youtube.com/watch?v=2fXQvy0kFak&ab_channel=HiteshChoudhary

  https://www.youtube.com/watch?v=EMEvheCVhMk&list=PL7ersPsTyYt2Q-SqZxTA1D-melSfqBRMW&ab_channel=MySirG.com

/user Endpoint

  https://www.youtube.com/user/krishnaik06

  https://www.youtube.com/user/saurabhexponent1/videos

/c Endpoint

  https://www.youtube.com/c/HiteshChoudharydotcom

  https://www.youtube.com/c/Telusko

/channel Endpoint

    https://www.youtube.com/channel/UCjWY5hREA6FFYrthD0rZNIw

2. No. of videos to scrape: Max limit is set to be equal to or less than 50 videos. However there is a maximum limit to fetch the data from Youtube API is 256.

3. Upload to S3: Flag to save the video into S3 bucket. Fetching Stream data from Youtube API takes a lot of time and multiplies as the video count increases and also the process of downloading multiple files from youtube and uploading to S3 is not Asynchronous which causes SESSION TIMEOUT error. Hence it is advisable to set the no. of video count to 1 to test the functionality. Improving this functionality could be the next thing to do.

Output