video-search-engine's Introduction

Video Search Engine

An offline search engine for local media files

Features

Add a local Video Directory
Let the application index the videos
Use search feature in the Web application to find the videos

Installation on Docker

docker compose up

After successfully creating the containers

Refer Getting Started After Installation

Installation on Local Machine

Tools Dependencies

Setup Virtual environment and Install the dependencies

python -m venv venv
source venv/Scripts/activate # windows
source venv/bin/activate # linux
pip install -r SETUP/requirements.txt

mkdir Models
cd Models
git clone https://huggingface.co/google/vit-base-patch16-224

Load MySQL dump

Modify the MySQL login credentials in the following section of mysql_DB.py file

db = mysql.connector.connect(
    host="db", # for docker
   #host="localhost", for local machin installation 
    user="groot",
    password="iamgroot",
    database = "search_engine"
)

Load Table structure into the database

mysql -u username -p search_engine < SETUP/DB_Dump.sql

Run

Web application

python API.py &
npm run dev -- --host

Command Line application

python main.py

Getting Started After Installation

Step 1: Video Indexing

Videos must be indexed using
http://localhost:5000/add_videos/<vid_dir_path>
Replace <vid_dir_path> with custom video directory path with the project directory (move the custom video directory into the project directory)

Currently only the sample videos present in the project directory can be used with Docker installation. Custom video support is available in Local Machine Installation and will be introduced later in Docker version.

Sample videos are available in the project directory: