| Requirement | Why |
|---|---|
| Python 3.9+ | Runs the pipeline |
| ffmpeg | Audio decoding for Whisper |
| Ollama | Local model that writes the summaries |
| Git | Cloning the archive |
git clone https://github.com/willtheorangeguy/Shell-Game-Transcripts.git
cd Shell-Game-Transcripts
pip install -r requirements.txt
pip install git+https://github.com/openai/whisper.git
Whisper runs on the CPU by default, which is slow. For NVIDIA GPUs, install the CUDA build of PyTorch instead:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
python main.py <show> <year>
main.py runs every stage in sequence for one show and one year.
20242025Stages are also runnable individually when only part of the work needs redoing — see pipeline.md.
PowerShell utilities that sit alongside the pipeline:
| Script | Purpose |
|---|---|
archive.ps1 |
Define the current year to skip |
autocommit.ps1 |
Set strict mode to catch errors |
hf.ps1 |
$PrettyName |
hfall.ps1 |
All Shell Game Transcripts |
Each episode produces an audio file, a transcript, and a summary in its show and year folder. Stages write a log file (transcribed.log, summarized.log, cleaned.log) recording what they have already processed, so re-running a stage skips completed work rather than repeating it.