gbaeke / build_summarizer

Build summarizer

Description

This is a Python script that:

• takes an mp4 video as input
• separates the audio from the video (mp3 files in the chunks folder; chunks of 25MB)
• transcribes the audio to text with a whisper model in Azure
• summarizes the text with a GPT-4o model in Azure
• saves the summary as a markdown file in the same folder as the video and the same name as the video but with a .md extension

Note: the creation of the summary is basic and can be improved.

Add a .env file

AZURE_OPENAI_API_KEY=your_api_key_here
AZURE_OPENAI_ENDPOINT=your_endpoint_here
AZURE_OPENAI_API_KEY_4o=your_api_key_4o_here
AZURE_OPENAI_ENDPOINT_4o=your_endpoint_4o_here

Requirements

Create a Python virtual environment and install the requirements. Some requirements might require additional dependencies to be installed on your system.

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Add mp4 videos to a videos folder.

Run the app with:

python main.py

You might need to modify the main.py file with the name of your whisper and GPT-4o model deployments.

Select the video to start processing. When a video is processed, there will be a file with the same name as the video but with a .json extension. This file will contain the transcript of the video as JSON. Each entry is a part of the transcription.

When such a file exists, the video will not be processed again. The summary, however, will be generated again. This makes it easier to experiment with other prompts.

Assistant

The file assistant.py uses the Azure Assistant API v2 to add all *.md files to a vector search so you can chat with the assistant about the content of the files.