Skip to main content

BioSage - AI-Powered Medical Diagnostic Assistant TUI with ARGUS Debate Framework

Project description

BioSage Terminal

Version 2.1.1 | Production Release

PyPI version Python 3.11+ License: MIT Status: Production

AI-Powered Medical Diagnostic Assistant - Terminal User Interface

Overview

BioSage Terminal is a sophisticated TUI (Terminal User Interface) application for medical diagnostics powered by AI. It leverages the ARGUS debate-based reasoning framework to provide evidence-based diagnostic suggestions through multiple specialist AI agents.

Features

  • Full ARGUS Debate Framework Integration:

    • Moderator Agent: Orchestrates debate agenda and controls flow
    • Specialist Agents: Infectious Disease, Cardiology, Neurology, Oncology, Autoimmune, and Toxicology experts
    • Refuter Agent: Challenges weak evidence and generates rebuttals
    • Jury Agent: Renders final verdicts via Bayesian aggregation
    • CDAG: Conceptual Debate Graph tracks all propositions, evidence, and rebuttals
  • Evidence-Based Reasoning: Uses Bayesian posterior computation for calibrated confidence scores

  • Patient Management: Complete patient onboarding, vitals tracking, and case management

  • Persistent Storage: Local JSON-based storage for all patient data and cases

  • Beautiful TUI: Rich terminal interface with intuitive navigation using Textual framework

Installation

pip install biosage-terminal

# Install with preferred LLM provider
pip install biosage-terminal[gemini]   # Google Gemini (recommended)
pip install biosage-terminal[openai]   # OpenAI
pip install biosage-terminal[anthropic] # Anthropic Claude
pip install biosage-terminal[all-llms]  # All providers

Quick Start

After installation, simply run:

biosage

Or check your API configuration first:

biosage --check-api

Configuration

API keys and model names are configured via environment variables. BioSage auto-detects available providers in this order: Gemini > OpenAI > Anthropic > Groq > Mistral > Cohere > Ollama

Environment Variables

# API Keys
export GEMINI_API_KEY=your_key_here        # Google Gemini (recommended)
export OPENAI_API_KEY=your_key_here        # OpenAI
export ANTHROPIC_API_KEY=your_key_here     # Anthropic Claude
export GROQ_API_KEY=your_key_here          # Groq
export MISTRAL_API_KEY=your_key_here       # Mistral
export COHERE_API_KEY=your_key_here        # Cohere

# Optional: Custom model names
export GEMINI_MODEL=gemini-1.5-pro
export OPENAI_MODEL=gpt-4o
export ANTHROPIC_MODEL=claude-3-5-sonnet-20241022
export GROQ_MODEL=llama-3.1-70b-versatile

# For local Ollama (no API key needed)
export OLLAMA_MODEL=llama3.1

# Data storage directory (default: ~/.biosage)
export BIOSAGE_DATA_DIR=/path/to/data

You can also create a .env file in your working directory.

Navigation

Key Action
d Dashboard
o Patient Onboarding (New Case)
r Run Diagnosis
s Specialist Grid
e Evidence Explorer
Tab Switch tabs / Next field
Enter Select / Confirm
Esc Back / Cancel
q Quit

Data Storage

All data is stored locally in ~/.biosage/:

  • patients/ - Patient records (JSON)
  • cases/ - Diagnostic cases with ARGUS debate results
  • evidence/ - Evidence and citations
  • reports/ - Generated reports
  • audit/ - Audit trail of all actions

ARGUS Debate Architecture

BioSage Terminal uses the full ARGUS multi-agent debate framework:

  1. Moderator creates debate agenda for the diagnosis
  2. Specialist Agents (6 domains) gather domain-specific evidence
  3. Refuter Agent challenges evidence and generates rebuttals
  4. CDAG (Conceptual Debate Graph) tracks:
    • Propositions (diagnosis hypotheses)
    • Evidence nodes (supporting/attacking)
    • Rebuttal nodes (challenges to evidence)
  5. Bayesian Propagation computes posterior probabilities
  6. Jury Agent renders final verdicts with:
    • Verdict label (supported/rejected/undecided)
    • Posterior confidence
    • Clinical reasoning

Development

git clone https://github.com/biosage/biosage-terminal
cd biosage-terminal
pip install -e .[dev]

Changelog

v2.0.0 (Production Release)

  • Production-ready release with stable APIs
  • Full ARGUS debate framework integration
  • Multi-provider LLM support (Gemini, OpenAI, Anthropic, Groq, Mistral, Cohere, Ollama)
  • Complete patient management and case tracking
  • Evidence-based Bayesian diagnostic reasoning
  • Rich terminal UI with intuitive navigation

License

MIT License

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

biosage_terminal-2.1.1.tar.gz (89.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

biosage_terminal-2.1.1-py3-none-any.whl (104.1 kB view details)

Uploaded Python 3

File details

Details for the file biosage_terminal-2.1.1.tar.gz.

File metadata

  • Download URL: biosage_terminal-2.1.1.tar.gz
  • Upload date:
  • Size: 89.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.11.3

File hashes

Hashes for biosage_terminal-2.1.1.tar.gz
Algorithm Hash digest
SHA256 2bfae52c4be8da9d899d774df31508488dbbab2128068416542f6ccae2e6e56d
MD5 089afb5414c6ea51fcea784a3ae059a2
BLAKE2b-256 6e515e34e30dc281abe38ab91aea498153db1dda70493d6679cedcaf62419973

See more details on using hashes here.

File details

Details for the file biosage_terminal-2.1.1-py3-none-any.whl.

File metadata

File hashes

Hashes for biosage_terminal-2.1.1-py3-none-any.whl
Algorithm Hash digest
SHA256 5a635d0afdbafb5f16b6938178b40411fca3966fd487384f44822c463dbc912e
MD5 bcc4c99b1b7fdf7c51ab3e07a6665097
BLAKE2b-256 2daf920a14d1372e5f05b911f2483f4b27a1fcbbd3a837507453e1b8d0fa4332

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page