Flask Video to Text Application

Explanation Video

https://youtu.be/SDG8TWBiLbc?si=E0YTwIlI5t0L-kOw

Flowchart

This Flask application converts a video file to text and generates tags based on the extracted text. It uses MoviePy to extract audio from the video, Whisper for speech-to-text transcription, and Gemini API for generating tags.

Overview

1. MoviePy: A Python library used for video editing, including extracting audio from video files.
2. Whisper: An automatic speech recognition (ASR) model by OpenAI, used to transcribe audio to text.
3. Gemini API: Google's generative AI service used for generating text, including extracting keywords from the transcribed text.

Prerequisites

Before running the application, make sure you have the following installed:

• Python 3.x
• Flask
• MoviePy
• Whisper
• Google Generative AI (Gemini API)
• Python dotenv

You can install the required packages using pip:

pip install flask moviepy openai google-generativeai python-dotenv

Configuration

1. Create a .env file in the project directory with the following content:

   GOOGLE_API_KEY=your_google_api_key

   Replace your_google_api_key with your actual Google API key.

2. Set up Whisper: Ensure that the Whisper model is available for transcription. You might need to install it if it’s not already available.

Application Structure

• app.py: The main Flask application script.
• templates/index.html: The HTML form for uploading video files.

How It Works

1. Upload a Video: Use the web form to upload a video file.
2. Video to Audio: The application extracts the audio from the video using MoviePy.
3. Audio to Text: The audio file is transcribed into text using Whisper.
4. Generate Tags: The transcribed text is processed to extract keywords using Gemini API.
5. Display Results: The generated tags are displayed on the webpage.

Usage

1. Run the Application:

   python app.py

2. Access the Web Interface:

   Open a web browser and go to http://127.0.0.1:5000/.

3. Upload a Video File:

   • Choose a video file and submit it through the form.
   • The application processes the file and displays the generated tags.

Code Explanation

• read_text_from_file(file_path): Reads the content of a text file.
• video_to_audio(video_path, audio_path): Converts video to audio using MoviePy.
• audio_to_text(audio_path): Transcribes audio to text using Whisper.
• generate_tags_chat(prompt_text): Uses Gemini API to generate keywords from the transcribed text.