FuzzyTube

Overview

This Streamlit web app lets you input a YouTube video URL and search for stock mentions or any keywords (in English or Hindi) within the video's transcript. It supports fuzzy keyword matching and displays timestamped YouTube links to each stock mention.

---

Features

• Accepts stock names or tickers in English or Hindi
• Matches stock mentions using fuzzy text search
• Supports YouTube videos with English or Hindi transcripts
• Outputs timestamped links to where the stock is mentioned
• Automatically expands keywords using a company mapping CSV

---

Requirements

• Python 3.7+
• Streamlit
• pandas
• youtube-transcript-api [@jdepoix]
• rapidfuzz

---

Installation

Prerequisites

• Python 3.7 or higher
• Pip (Python package manager)

Steps

1. Clone the Repository:

   git clone https://github.com/Shriya-Guptaa/FuzzyTube
   cd FuzzyTube
   

2. Install Dependencies:

   pip install -r requirements.txt
   

3. Run the Application:

    streamlit run app.py
   

4. Access the Web App: Open your browser and visit http://localhost:8501 to start using FuzzyTube

---

Usages

• Input YouTube Video URL: Paste the link to a video with English or Hindi subtitles.

• Enter Stock Keywords: Add stock names, tickers, or company keywords (comma-separated).

• Search Transcripts: Click the Search button to find mentions using fuzzy matching.

• Explore Results: View timestamped clickable links directing to exact moments stocks are mentioned.

• Customize Keywords: Update the Company.csv file to add or edit company names and aliases in English and Hindi.

📚 Dependencies & Acknowledgements

This app uses the excellent youtube-transcript-api library by @jdepoix for fetching YouTube transcripts.

License

This project is licensed under the MIT License.