AIAgents4Pharma Developer Guide

This guide covers the complete development setup, tooling, and workflow for AIAgents4Pharma project.

📋 Table of Contents

Quick Start
Development Environment Setup
Project Structure
Development Tools
Code Quality & Security
Dependency Management
Testing
CI/CD Pipeline
Docker & Deployment
Security Best Practices
Common Development Tasks
Troubleshooting

🚀 Quick Start

Prerequisites

Python 3.12+
Git
uv (modern Python package manager)
libmagic (for file security validation):
- macOS: brew install libmagic
- Linux: sudo apt-get install libmagic1
- Windows: Bundled with python-magic-bin

Installation

# 1. Clone the repository
git clone https://github.com/VirtualPatientEngine/AIAgents4Pharma
cd AIAgents4Pharma

# 2. Install dependencies (creates virtual environment automatically)
uv sync --extra dev --frozen

# 3. Set up pre-commit hooks (optional but recommended)
uv run pre-commit install

# 4. Set up API keys
export OPENAI_API_KEY=sk-...
export NVIDIA_API_KEY=nvapi-...
export ZOTERO_API_KEY=...
export ZOTERO_USER_ID=...

# 5. Test installation
uv run python -c "import aiagents4pharma; print('✅ Installation successful!')"

🛠 Development Environment Setup

Modern Python Stack

This project uses a modern Python development stack:

📦 uv: Ultra-fast Python package manager and dependency resolver
🏗️ hatchling: Modern build backend (PEP 621 compliant)
📝 pyproject.toml: Single source of truth for project configuration
🔒 uv.lock: Reproducible dependency resolution

Why uv over pip/conda?

10-100x faster than pip for dependency resolution
Automatic virtual environment management
Built-in lock file support for reproducible builds
Better dependency conflict resolution
Native pyproject.toml support

📂 Project Structure

AIAgents4Pharma/
├── aiagents4pharma/           # Main package
│   ├── talk2biomodels/        # Systems biology agent
│   ├── talk2knowledgegraphs/  # Knowledge graph agent
│   ├── talk2scholars/         # Scientific literature agent
│   ├── talk2cells/            # Single cell analysis agent
│   └── talk2aiagents4pharma/  # Meta-agent (orchestrator)
├── app/                       # Streamlit applications
├── docs/                      # Documentation
├── pyproject.toml            # Project configuration (dependencies, tools)
├── uv.lock                   # Lock file for reproducible builds
├── .pre-commit-config.yaml   # Pre-commit hooks configuration
└── release_version.txt       # Version file

🔧 Development Tools

Code Quality Tools

All tools are configured in pyproject.toml and run automatically via pre-commit:

⚡ Ruff - Fast Linting & Import Sorting

# Lint and auto-fix issues
uv run ruff check --fix .

# Check only (no fixes)
uv run ruff check .

# Format imports and code style
uv run ruff format .

📋 Pylint - Comprehensive Static Analysis

# Analyze entire codebase (configuration in pyproject.toml)
uv run pylint aiagents4pharma/

# Analyze specific component
uv run pylint aiagents4pharma/talk2scholars/

# Generate JSON report for CI/CD
uv run pylint aiagents4pharma/ --output-format=json --reports=no > pylint-report.json

🔍 MyPy - Static Type Checking

# Type check the main package
uv run mypy aiagents4pharma/

# Type check everything
uv run mypy .

Security Tools

🛡️ Bandit - Security Vulnerability Scanner

# Scan for security issues
uv run bandit -r aiagents4pharma/

# Generate detailed report
uv run bandit -r aiagents4pharma/ -f json -o security-report.json

🔒 Dependency Vulnerability Scanning

# Scan dependencies for known vulnerabilities
uv run pip-audit

# Alternative scanner
uv run safety check

# Scan with detailed output
uv run pip-audit --desc --format=json

🔄 Code Quality & Security

Pre-commit Hooks

Pre-commit runs automatically before every commit to ensure code quality:

# Install hooks (one-time setup)
uv run pre-commit install

# Run hooks manually on all files
uv run pre-commit run --all-files

# Run specific hook
uv run pre-commit run ruff
uv run pre-commit run mypy

What runs on each commit

Ruff - Lints and fixes imports
MyPy - Type checking (configured but currently disabled in pre-commit)
Bandit - Security scanning
pip-audit - Dependency vulnerability check
General checks - Trailing whitespace, YAML validation, etc.

Bypassing Pre-commit (Emergency Only)

# Skip pre-commit hooks (not recommended)
git commit --no-verify -m "emergency fix"

📦 Dependency Management

Adding Dependencies

# Add runtime dependency
uv add numpy>=1.24.0

# Add development dependency
uv add --group dev pytest>=7.0.0

# Add optional dependency group
uv add --optional ml torch>=2.0.0

Updating Dependencies

# Update all dependencies
uv sync --upgrade

# Update specific package
uv add package_name@latest

# Update dev dependencies
uv sync --extra dev --upgrade

Lock File Management

# Generate/update lock file
uv lock

# Install from lock file (production)
uv sync --frozen

# Install with development tools
uv sync --extra dev --frozen

Dependency Groups

Main: Core runtime dependencies
Dev: Development tools (ruff, mypy, etc.)
Optional: Feature-specific dependencies

🧪 Testing

Running Tests

# Run all tests
uv run pytest

# Run with coverage
uv run pytest --cov=aiagents4pharma

# Run specific test file
uv run pytest aiagents4pharma/talk2biomodels/tests/test_api.py

# Run integration tests only
uv run pytest -m integration

Test Categories

Unit tests: Fast, isolated tests
Integration tests: Cross-component tests (marked with @pytest.mark.integration)

🔄 CI/CD Pipeline

GitHub Actions Workflows

The project uses GitHub Actions for automated testing and deployment:

# .github/workflows/ci.yml (example)
name: CI
on: [push, pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: astral-sh/setup-uv@v1
      - run: uv sync --extra dev
      - run: uv run pytest
      - run: uv run pip-audit

Manual CI Commands

# Run the same checks as CI locally
uv run pytest                    # Tests
uv run pylint aiagents4pharma/   # Static analysis (config in pyproject.toml)
uv run pip-audit                 # Security scan
uv run safety check              # Alternative security scan
uv run bandit -r aiagents4pharma/ # Security scan
uv run mypy aiagents4pharma/     # Type checking

Automated Security Workflows

The project includes comprehensive automated security scanning:

# Weekly security audit (runs automatically)
.github/workflows/security_audit.yml   # pip-audit + safety + bandit

# SonarCloud analysis (self-contained quality gate workflow)
.github/workflows/sonarcloud.yml       # Generates reports and uploads one stable scan

# Manual security audit
uv run pip-audit --desc
uv run safety check --json
uv run bandit -c pyproject.toml -r aiagents4pharma/

🐳 Docker & Deployment

Building Docker Images

Each agent has its own Dockerfile:

# Build specific agent
docker build -f aiagents4pharma/talk2scholars/Dockerfile -t talk2scholars .

# Build all with docker-compose
docker-compose build

Production Deployment

# Install production dependencies only (excludes dev tools)
uv sync --frozen

# Build production package
uv build

# Install built package
pip install dist/aiagents4pharma-*.whl

🚨 Security Best Practices

Regular Security Scans

# Weekly security scan (runs automatically in CI)
uv run pip-audit --desc
uv run safety check --json
uv run bandit -c pyproject.toml -r aiagents4pharma/

# Check for outdated packages with vulnerabilities
uv run pip-audit --desc --format=json

Streamlit File Upload Security

The project implements comprehensive file upload security:

# Use secure file upload wrapper
from app.frontend.utils.streamlit_utils import secure_file_upload

# Secure PDF upload with validation
pdf_file = secure_file_upload(
    "Upload PDF",
    allowed_types=["pdf"],
    help_text="Upload a research paper",
    max_size_mb=50,
    accept_multiple_files=False
)

# Secure data upload with multiple types
data_files = secure_file_upload(
    "Upload Data",
    allowed_types=["spreadsheet", "text"],
    help_text="Upload CSV or Excel files",
    max_size_mb=25,
    accept_multiple_files=True
)

Security Features

File type validation - Only allowed extensions (prevents malware.exe → report.pdf)
MIME type checking - Detects file masquerading attacks
File size limits - Prevents DoS attacks (configurable 25-50MB)
Content scanning - Blocks suspicious patterns and scripts
Filename sanitization - Prevents directory traversal attacks

Dependency Updates

Dependabot automatically creates PRs for security updates (weekly)
Pre-commit hooks catch vulnerabilities before commit
CI pipeline blocks PRs with security issues
Weekly security audits with SARIF uploads to GitHub Security

API Key Management

# Set environment variables (never commit these!)
export OPENAI_API_KEY=sk-...
export NVIDIA_API_KEY=nvapi-...
export ZOTERO_API_KEY=...
export ZOTERO_USER_ID=...

# Use .env file for local development (add to .gitignore!)
echo "OPENAI_API_KEY=sk-..." >> .env

🛠 Common Development Tasks

Starting Development

# 1. Activate environment and install dependencies
uv sync --extra dev --frozen

# 2. Run pre-commit setup
uv run pre-commit install

# 3. Start coding!

Before Committing

# 1. Run quality checks
uv run ruff check --fix .
uv run pylint aiagents4pharma/
uv run mypy aiagents4pharma/

# 2. Run tests
uv run pytest

# 3. Security scan
uv run pip-audit

# 4. Commit (pre-commit will run automatically)
git add .
git commit -m "your message"

Adding a New Agent

Create new directory: aiagents4pharma/talk2newagent/
Add dependencies to pyproject.toml
Update package configuration
Add tests and documentation
Update Docker configuration

🐛 Troubleshooting

Common Issues

Dependency Conflicts

# Clear cache and reinstall
rm -rf .venv uv.lock
uv sync --extra dev

Pre-commit Issues

# Reinstall hooks
uv run pre-commit uninstall
uv run pre-commit install

# Update hook versions
uv run pre-commit autoupdate

Import Errors

# Verify installation
uv run python -c "import aiagents4pharma; print('OK')"

# Check Python path
uv run python -c "import sys; print(sys.path)"

Type Checking Errors

# Install missing type stubs
uv add --group dev types-requests types-PyYAML

# Run with verbose output
uv run mypy --verbose aiagents4pharma/

Performance Issues

# Profile dependency resolution
uv sync --extra dev --verbose

# Check lock file
uv lock --verbose

📚 Additional Resources

Core Tools

uv Documentation - Modern Python package manager
Hatchling - Modern build backend
Ruff Rules - Fast Python linter
MyPy Configuration - Static type checking
Pre-commit Hooks - Git hook framework

Security Tools

Bandit - Security linter for Python
pip-audit - Dependency vulnerability scanner
Safety - Dependency vulnerability checker
python-magic - File type detection
Streamlit Security Guide - File upload security implementation

CI/CD & Quality

SonarCloud Setup Guide - Complete SonarCloud integration guide
SonarCloud - Code quality and security analysis
GitHub Actions - CI/CD workflows
Dependabot - Automated dependency updates

🤝 Contributing

Fork the repository
Create feature branch: git checkout -b feature/amazing-feature
Setup development environment: uv sync --extra dev --frozen
Install pre-commit: uv run pre-commit install
Make changes and ensure all checks pass
Commit with descriptive message
Push to your fork and create Pull Request

All contributions are automatically scanned for:

Code formatting and style (Ruff)
Type safety (MyPy - configured, ready to enable)
Security vulnerabilities (Bandit + pip-audit + Safety)
Test coverage (pytest with coverage reporting)
Code quality (SonarCloud analysis)
Dependency security (Automated weekly scans)

Happy coding! 🚀

FilesExpand file tree

README.md

Latest commit

History

README.md

File metadata and controls

AIAgents4Pharma Developer Guide

📋 Table of Contents

🚀 Quick Start

Prerequisites

Installation

🛠 Development Environment Setup

Modern Python Stack

Why uv over pip/conda?

📂 Project Structure

🔧 Development Tools

Code Quality Tools

⚡ Ruff - Fast Linting & Import Sorting

📋 Pylint - Comprehensive Static Analysis

🔍 MyPy - Static Type Checking

Security Tools

🛡️ Bandit - Security Vulnerability Scanner

🔒 Dependency Vulnerability Scanning

🔄 Code Quality & Security

Pre-commit Hooks

What runs on each commit

Bypassing Pre-commit (Emergency Only)

📦 Dependency Management

Adding Dependencies

Updating Dependencies

Lock File Management

Dependency Groups

🧪 Testing

Running Tests

Test Categories

🔄 CI/CD Pipeline

GitHub Actions Workflows

Manual CI Commands

Automated Security Workflows

🐳 Docker & Deployment

Building Docker Images

Production Deployment

🚨 Security Best Practices

Regular Security Scans

Streamlit File Upload Security

Security Features

Dependency Updates

API Key Management

🛠 Common Development Tasks

Starting Development

Before Committing

Adding a New Agent

🐛 Troubleshooting

Common Issues

Dependency Conflicts

Pre-commit Issues

Import Errors

Type Checking Errors

Performance Issues

📚 Additional Resources

Core Tools

Security Tools

CI/CD & Quality

🤝 Contributing