What problem does it solve?
This Skill eliminates the manual, error-prone troubleshooting required to manage the CondaPkg and PythonCall Python environment for the CalibrateEmulateSample.jl Julia package, which powers the scikit-learn and scipy dependencies of the SKLPy Gaussian process backend.
Core Features & Use Cases
- Version Pinning and Updates: Pin, bump, or downgrade Python, scikit-learn, and scipy versions via the CondaPkg.toml manifest at the repo root.
- Dependency Conflict Resolution: Fix Julia-side version conflicts between CondaPkg and PythonCall, stale Manifest issues, and Pkg.update failures related to these packages.
- Import and Environment Debugging: Resolve Python import errors and out-of-sync CondaPkg environments that cause failures like "sklearn could not be imported".
- Backend Type Renaming: Safely rename Julia dispatch types that wrap the Python backend, adding proper deprecation warnings for backwards compatibility.
A common use case is fixing a broken Python environment after updating a dependency, or updating the installation documentation to match new pinned Python package versions.
Quick Start
Use the python-dep-manager skill to bump the scikit-learn version, resolve CondaPkg environment conflicts, or debug Python import errors in the CalibrateEmulateSample.jl repository.