cva-setup-interop

Use libpython-clj to enable Python-Clojure interoperability. Call (py/initialize!) at startup, import Python modules with (require-python '[numpy :as np]), and work with NumPy arrays and Pandas DataFrames directly in Clojure code with automatic data conversion between the two languages.

Yes. libpython-clj supports importing and calling Python functions from transformers and scikit-learn within Clojure workflows. The Skill handles Python environment configuration, automatic data type conversion, and GIL management to execute machine learning operations seamlessly in your Clojure application.

Add libpython-clj as a dependency in deps.edn, then initialize it with (py/initialize!) at application startup. The Skill supports auto-detection of your Python environment, manual path configuration, and virtual environment setup—choose the approach that fits your deployment context.

libpython-clj automatically converts Clojure data types to Python equivalents and vice versa, including vectors to NumPy arrays, maps to Python dictionaries, and Clojure sequences to Pandas DataFrames. Complex nested structures convert seamlessly, eliminating manual marshalling overhead.

The Skill includes built-in GIL handling and performance optimizations for production deployments. It manages thread-safe access to Python code, allowing you to safely use Python ML libraries in concurrent Clojure workflows without manual GIL coordination.

libpython-clj supports auto-detection of system Python installations and manual configuration for specific Python versions. It works with virtual environments and supports NumPy, Pandas, Transformers, and Torch dependencies, adapting to your environment setup strategy.