This repository contains a simple rule-based "meeting transcription agent" for dental SOAP note charting. The agent expects a text transcript of a conversation (for example, a short stand-up between a dentist and a dental assistant while treating a patient) and produces a structured Subjective, Objective, Assessment, and Plan summary.
-
Create and activate a Python 3.9+ virtual environment.
-
Install the project in editable mode with the optional CLI extras:
pip install -e .[cli]
-
Run the agent against one of the sample transcripts:
python -m dentist_agent.transcription_agent transcripts/sample_transcript.txt
The agent will print the generated SOAP note to standard output as formatted Markdown.
├── dentist_agent/
│ ├── __init__.py
│ └── transcription_agent.py
├── transcripts/
│ └── sample_transcript.txt
├── tests/
│ └── test_transcription_agent.py
├── README.md
└── pyproject.toml
dentist_agent/transcription_agent.pyimplements the rule-based summarisation logic.transcripts/sample_transcript.txtprovides a short example conversation.tests/test_transcription_agent.pyexercises the core extraction helper.
Run the automated checks with:
python -m pytest