Skip to content

CONFIGURATION.md

Latest commit

 

History

History
99 lines (77 loc) · 3.08 KB

CONFIGURATION.md

File metadata and controls

99 lines (77 loc) · 3.08 KB
Raw

Configuration Guide

This application uses environment variables for configuration. Create a .env file in the project root with the following variables:

Required Configuration

Mistral API Configuration

  • API_KEY_MISTRAL: Your Mistral API key (required)
  • MODEL: Mistral model to use (default: mistral-large-latest)

LangGraph Configuration

  • LANGSMITH_API_KEY: Your LangSmith API key for tracing and monitoring (optional)
  • LANGSMITH_TRACING: Enable LangSmith tracing (default: true)
  • LANGSMITH_PROJECT: LangSmith project name (default: medical-xray-workflow)
  • LANGSMITH_ENDPOINT: LangSmith API endpoint (default: https://api.smith.langchain.com)

Tavily Configuration

  • TAVILY_API_KEY: Your Tavily API key for web search capabilities (optional)

Optional Configuration

Patient Configuration

  • PATIENT_NAME: Patient's name (default: John Doe)
  • PATIENT_AGE: Patient's age in years (default: 30)
  • PATIENT_WEIGHT: Patient's weight in kg (default: 130)
  • PATIENT_HEIGHT: Patient's height in cm (default: 170)
  • PATIENT_BLOOD_TYPE: Patient's blood type (default: A+)
  • PATIENT_GENDER: Patient's gender (default: Male)
  • PATIENT_HYPERTENSION: Whether patient has hypertension (default: True)

X-Ray Diagnosis Configuration

  • XRAY_MODEL_WEIGHTS: TorchXRayVision model weights (default: densenet121-res224-all)
  • XRAY_IMAGE_SIZE: Image size for processing (default: 224)
  • XRAY_NORMALIZATION_RANGE: Normalization range for images (default: 255)

UI Configuration

  • GRADIO_THEME: Gradio theme color (default: indigo)
  • GRADIO_SHARE: Whether to create a public share link (default: False)
  • GRADIO_DEBUG: Enable debug mode (default: True)

Example .env File

# Mistral API Configuration
API_KEY_MISTRAL=your_mistral_api_key_here
MODEL=mistral-large-latest

# Langgraph Configuration
LANGSMITH_API_KEY=""
LANGSMITH_TRACING=true
LANGSMITH_PROJECT="medical-xray-workflow"
LANGSMITH_ENDPOINT="https://api.smith.langchain.com"

# Tavily Configuration
TAVILY_API_KEY=""

# Patient Configuration
PATIENT_NAME=John Doe
PATIENT_AGE=30
PATIENT_WEIGHT=130
PATIENT_HEIGHT=170
PATIENT_BLOOD_TYPE=A+
PATIENT_GENDER=Male
PATIENT_HYPERTENSION=True

# X-Ray Diagnosis Configuration
XRAY_MODEL_WEIGHTS=densenet121-res224-all
XRAY_IMAGE_SIZE=224
XRAY_NORMALIZATION_RANGE=255

# UI Configuration
GRADIO_THEME=indigo
GRADIO_SHARE=False
GRADIO_DEBUG=True

Available TorchXRayVision Models

You can change the XRAY_MODEL_WEIGHTS to use different pre-trained models:

  • densenet121-res224-all: DenseNet121 trained on all datasets (default)
  • densenet121-res224-chex: DenseNet121 trained on CheXpert
  • densenet121-res224-mimic_nb: DenseNet121 trained on MIMIC-CXR
  • densenet121-res224-nih: DenseNet121 trained on NIH Chest X-ray
  • densenet121-res224-pc: DenseNet121 trained on PadChest
  • densenet121-res224-kaggle: DenseNet121 trained on Kaggle Pneumonia

Available Gradio Themes

You can change the GRADIO_THEME to use different color schemes:

  • indigo (default)
  • blue
  • green
  • orange
  • red
  • purple
  • pink
  • yellow