Skip to content

README.md

Latest commit

 

History

History
26 lines (20 loc) · 1.12 KB

README.md

File metadata and controls

26 lines (20 loc) · 1.12 KB
Raw

Taiga Core

Core library for the TAIGA framework - provides base classes and utilities for creating AI problem-solving environments using MCP (Model Context Protocol) servers.

Development and Testing

# From the repository root:
# Create virtual environment and install with development dependencies
uv venv
source .venv/bin/activate
uv pip install "./taiga-core[dev]"

# Run validation tests for an environment
pytest -m validate_env --problems-metadata-path=examples/math-in-python/problems-metadata.json

The validation tests verify that your environment is properly configured by:

  • Checking that the Docker image specified in problems-metadata.json can be spawned
  • Verifying that required tools (setup_problem and grade_problem) are available
  • Ensuring the problem ID format follows naming conventions
  • Validating the metadata schema structure

Graders that call the Anthropic API (e.g. the rubric example) run against a local stub (tests/anthropic_api_stub.py) when ANTHROPIC_API_KEY is not set, so the validation tests work without API access. Export ANTHROPIC_API_KEY to run grading against the live API instead.