v0.13.0

What's Changed

• Cache conditioning samples to improve efficiency by @markean in #220
• Refactor keyword argument handling and update CI configuration by @markean in #221
• Update README to change CI badge and correct installation guide links by @markean in #222
• Enhance posterior property documentation and handle empty posterior samples by @markean in #223
• Add shard_axis draw-parallel sharding for disk-backed predictive meth… by @markean in #228
• Fix validation to exclude strings and mappings from array-like checks by @markean in #230
• Add validation for unknown return sites in ImpactModel.predict and co… by @markean in #232
• Update validation and tests to raise ValueError for missing predictive groups in estimate_effect by @markean in #234
• Improve shutdown handling for writer threads to prevent blocking by @markean in #236
• Disk-backed methods allow repeated runs with differnt optional argument passing by @markean in #238
• Fix intervention predictions by removing redundant random key splitting by @markean in #240
• Add warning for dimension size mismatch in estimate_effect and corresponding test by @markean in #242
• Fix background writer-thread failure handling in disk-backed methods by @markean in #244
• Forward the non-array kwargs into the probe trace in prior predictive sampling by @markean in #246
• Support custom parameter inputs and outputs; update tests accordingly by @markean in #248
• Enhance ArrayLoader integration with disk-backed methods by @markean in #250
• Clarify temporary directory management in ImpactModel; update docstrings and user guide for output_dir behavior by @markean in #251
• Require y for log_likelihood() when X is array-like; raise ValueError if not provided. by @markean in #255
• Improve error handling for input type in _setup_inputs; raise TypeError for non-array-like or non-loader X. by @markean in #257
• Ensure output subdirectory is cleaned up on write failure by @markean in #259
• Register cloudpickle-restored models for cleanup in ImpactModel. by @markean in #261
• Improve MLflow integration with bug fixes and tests by @markean in #263
• Update documentation and version for release 0.13.0 by @markean in #264

Full Changelog: v0.12.0...v0.13.0