https://github.com/daxrajsinh/healthscribe_doctor-patient
https://github.com/daxrajsinh/healthscribe_doctor-patient
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/daxrajsinh/healthscribe_doctor-patient
- Owner: Daxrajsinh
- Created: 2025-01-28T06:29:38.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-01-28T06:48:17.000Z (4 months ago)
- Last Synced: 2025-01-28T07:28:38.331Z (4 months ago)
- Language: Python
- Size: 0 Bytes
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AWS Healthscribe Doctor-Patient Transcription
This project is a Flask-based application that uses **AWS HealthScribe** and other AWS services to process medical conversations between doctors and patients. It converts audio files into structured text transcriptions and clinical insights, enabling efficient documentation and analysis.
---
## Features
- **AWS HealthScribe Integration**:
- Automatically transcribes doctor-patient conversations.
- Extracts clinical insights, including symptoms and diagnoses.
- Saves structured data as a JSON file.
- **Flask API**:
- Exposes endpoints for querying the transcriptions and summaries.
- Allows question-answering based on transcription summaries using **Claude** (via AWS Bedrock).
- **AWS Services**:
- **Amazon S3**: Stores audio files and transcription outputs.
- **AWS Bedrock**: Queries foundation models like Claude for Q&A.
- **Amazon Transcribe Medical**: Converts audio to text with speaker labels.
---
## Table of Contents
- [Features](#features)
- [Getting Started](#getting-started)
- [Usage](#usage)
- [Endpoints](#endpoints)
- [Technologies Used](#technologies-used)
- [Project Structure](#project-structure)
- [License](#license)
---
## Getting Started
### Prerequisites
1. Python 3.7 or above.
2. AWS account with the following services enabled:
- Amazon S3
- Amazon Transcribe Medical
- AWS Bedrock
3. Install **Git** and **AWS CLI**.
### Installation
1. Clone this repository:
```bash
git clone https://github.com/your-username/Healthscribe_Doctor-Patient.git
cd Healthscribe_Doctor-Patient
2. Create and activate a virtual environment::
```bash
python3 -m venv my_env
source my_env/bin/activate
3. Configure AWS CLI with your credentials:
```bash
aws configure
4. Clone this repository:
```bash
git clone https://github.com/your-username/Healthscribe_Doctor-Patient.git
cd Healthscribe_Doctor-Patient
5. Set up a .env file with your AWS configurations:
```bash
AWS_REGION=your-region
BUCKET_NAME=your-s3-bucket
DATA_ACCESS_ROLE_ARN=your-iam-role-arn
AUDIO_FILE_URL=s3://your-bucket/audio-file.mp3
BEDROCK_MODEL_ID=anthropic.claude-3
### Usage
- **Run the Application**:
- Start the Flask app:
```bash
python app.py
## Upload an Audio File
1. **Upload an audio file to your S3 bucket**:
- Ensure the audio file is in the correct format (e.g., MP3, WAV).
- Upload the file to the S3 bucket specified in your `.env` file.
2. **Set the file's public URL**:
- Update your `.env` file with the public URL of the uploaded audio file:
```plaintext
AUDIO_FILE_URL=https://your-bucket.s3.your-region.amazonaws.com/your-audio-file.mp3
```
---
## Process the Transcription
The application processes the transcription in the following steps:
1. **Start an AWS HealthScribe Job**:
- It uses the **AWS Transcribe Medical Scribe** service to transcribe the uploaded audio file.
2. **Poll the Transcription Job**:
- The app waits for the transcription job to complete.
3. **Fetch the Transcription Summary**:
- Once the job is complete, the transcription output (e.g., `summary.json`) is fetched.
4. **Save the Summary to S3**:
- The transcription summary is stored in the specified S3 bucket.
---
## Question-Answering with Claude
- **Ask Questions About the Transcription**:
- Use the `/question-ans` API endpoint to submit questions based on the transcription summary.
- This functionality is powered by **Claude** (via AWS Bedrock).
---
## Endpoints
### **POST /question-ans**
- **Description**: Submit a question to query insights from the transcription summary.
#### **Request Example**:
```json
{
"question": "What symptoms did the patient describe?"
}
```
#### **Response Example**:
```json
{
"question": "What symptoms did the patient describe?",
"answer": "The patient described headache and dizziness."
}