talk-summarizer

Python library to summarize talks. Turn a video file into a PDF report with a summary of each section of the talk.

summarize_talk.py - operates on a talk/lecture where there is a video of the speaker and slides.

• uses ffmpeg to extract the audio from the video
• uses whisper to transcribe the audio
• uses OpenCV to identify slide transitions and extract slides
• uses OpenAI to summarize each section

Installation

Install this library using pip:

pip install talk-summarizer

Usage

python talk_summarizer/summarize_talk.py \
    --output_dir output \
    --video_file video.mp4 \
    --no-summarize \
    --whisper_model tiny

Development

To contribute to this library, first checkout the code. Then create a new virtual environment:

cd talk-summarizer
python -m venv venv
source venv/bin/activate

Now install the dependencies and test dependencies:

pip install -r requirements.txt

To run the tests:

pytest

Install pre-commit hooks:

pre-commit install

Set API keys

Make sure your OPENAI_API_KEY is set as an environment variable in either your .bashrc or .zshrc file. For example:

export OPENAI_API_KEY=sk-1234...

TODO

• get output directory working (ffmpeg not working with output directory)
• create different tasks like transcribe (video --> audio, transcript, overall summary). Right now summarize_talk works on a video (video with slides --> audio, slide images, transcript, section summaries, overall summary, report)
• add more unit tests
• use pyproject.toml instead of setup.py