Requirements
- Python 3.11 or later
- pip
Standard Install
pip install goldencheck
This installs GoldenCheck with all core dependencies:
| Package | Version | Purpose |
|---|---|---|
| polars | >=1.0 | Data operations |
| typer | >=0.12 | CLI framework |
| rich | >=13.0 | Console output |
| pyyaml | >=6.0 | Config file parsing |
| pydantic | >=2.0 | Config validation |
| openpyxl | >=3.1 | Excel file support |
| textual | >=1.0 | Interactive TUI |
With LLM Boost
To use --llm-boost you need the optional llm extras:
pip install goldencheck[llm]
This adds:
| Package | Version | Purpose |
|---|---|---|
| anthropic | >=0.30 | Anthropic Claude API |
| openai | >=1.30 | OpenAI GPT API |
You only need one provider installed, but both are included in the extras group.
With MCP Server
To use goldencheck mcp-serve for Claude Desktop integration:
pip install goldencheck[mcp]
This adds:
| Package | Version | Purpose |
|---|---|---|
| mcp | >=1.0 | Model Context Protocol server SDK |
See MCP Server for setup instructions.
All Extras
Install everything:
pip install goldencheck[llm,mcp]
Verify Installation
goldencheck --version
# GoldenCheck 0.5.0
Development Setup
Clone the repository and install in editable mode with dev dependencies:
git clone https://github.com/benzsevern/goldencheck.git
cd goldencheck
pip install -e ".[dev,llm,mcp]"
Dev dependencies include:
| Package | Purpose |
|---|---|
| pytest >=8.0 | Test runner |
| pytest-cov >=5.0 | Coverage reporting |
| ruff >=0.4 | Linter and formatter |
Running Tests
pytest
With coverage:
pytest --cov=goldencheck --cov-report=term-missing
Linting
ruff check goldencheck/
ruff format goldencheck/
Project Structure
goldencheck/
├── goldencheck/
│ ├── cli/ # CLI entry points (Typer)
│ ├── config/ # YAML config loader, schema, writer
│ ├── engine/ # Scanner, validator, reader, sampler
│ ├── llm/ # LLM boost: providers, prompts, budget, merger
│ ├── mcp/ # MCP server (6 tools)
│ ├── models/ # Finding and DatasetProfile dataclasses
│ ├── notebook.py # Jupyter/Colab display hooks
│ ├── profilers/ # 10 column profilers
│ ├── relations/ # 2 cross-column profilers
│ ├── semantic/ # Semantic type classification
│ ├── reporters/ # Rich console, JSON, CI reporters
│ └── tui/ # Textual TUI (4 tabs)
├── tests/ # pytest test suite (166 tests)
├── benchmarks/ # Speed and detection benchmarks
├── docs/wiki/ # This documentation
└── pyproject.toml