Skip to content

v0.1.4 - Signal Suitability Evaluation & Package Distribution

Choose a tag to compare

View all tags
@stabilefrisur stabilefrisur released this 07 Nov 21:45
· 225 commits to master since this release
v0.1.4
81ae24c

Added

Evaluation Layer

  • Signal-product suitability assessment framework with 4-component scoring:
    • Data health (20%): Sample size and missing data quality
    • Predictive association (40%): Statistical significance via OLS regression
    • Economic relevance (20%): Effect size in basis points
    • Temporal stability (20%): Subperiod beta sign consistency
  • SuitabilityConfig for immutable evaluation parameters (frozen dataclass)
  • SuitabilityResult with structured PASS/HOLD/FAIL decisions
  • SuitabilityRegistry for tracking evaluation history with JSON catalog
  • Markdown report generation with component breakdowns and interpretations
  • Statistical tests: correlation, OLS regression, subperiod analysis
  • Configurable scoring thresholds and component weights
  • Comprehensive test suite with 7 test modules
  • Suitability demonstration example (suitability_demo.py)
  • 10 sample evaluation reports in reports/suitability/demo_reports/

Distribution Improvements

  • Documentation now included in PyPI package distribution (src/aponyx/docs/)
  • Examples included in PyPI package distribution (src/aponyx/examples/)
  • Helper functions for locating docs and examples after installation:
    • get_docs_dir() for accessing documentation
    • get_examples_dir() for locating example scripts
  • Examples can now be run via python -m aponyx.examples.<demo_name>

Changed

  • All demonstration examples now use Bloomberg Terminal as primary data source with graceful fallback to synthetic data
  • data_demo.py: BloombergSource with FileSource fallback
  • models_demo.py: Bloomberg fetches for CDX/VIX/ETF data
  • backtest_demo.py: Real market data (2024-01-01 to present)
  • persistence_demo.py: Bloomberg fetch → save → register workflow
  • end_to_end_demo.ipynb: Updated for Bloomberg integration
  • Performance metrics now consolidated into run_metadata.json under performance_metrics key
  • Type hints cleaned up across examples folder to use modern Python syntax

Fixed

  • Bloomberg provider implementation corrected and validated with comprehensive tests
  • .gitignore patterns updated to properly exclude runtime data while preserving static config
  • Type annotations in examples now follow project guidelines consistently

Documentation

  • Updated PROJECT_STATUS.md for accuracy and clarity
  • Added evaluation layer to architecture documentation
  • Research workflow diagram showing PASS/FAIL quality gate branching
  • Complete methodology documentation in signal_suitability_evaluation.md
  • Updated Bloomberg requirements with installation instructions
  • Consolidated dependency sections and repository structure
  • Added agent context hints for evaluation layer in copilot-instructions.md
Assets 2
Loading