0.15.0

This release:

• Adds a new parameter baseline_theta to the Parameters class that allows the user to specify whether to use the steady-state replacement rate parameters from the baseline solution in a reform run. See PR #1077.

0.14.14

This release:

• Create SS.SS_initial_guesses function to allow more flexible initial guesses for steady state solution (PR #1061)
• Robust steady-state solution used for reform solution (PR #1061)
• Test of SS.solve_for_j function (PR #1061)

0.14.13

This release:

• Fix calculation of consumption tax revenue with differentiated goods (PR #1074)

0.14.12

This release:

• Uses data for pre-time path population distribution (rather than inferring it) (PR #1071)

0.14.11

This release:

• Adds Ethiopia demographic data mapping.

0.14.10

This release:

• Fixes nonconformable matrices in TPI.py introduced in version 0.14.9 (PR #1054)

0.14.9

This release:

• Fixes replacement_rate_adjustment parameter in the steady state (PR #1053)
• Adds some saved output to tpi_vars.pkl object (PR #1054)

0.14.8

This release:

• Adds a complete benchmark suite for measuring and optimizing Dask performance in OG-Core, with particular focus on Windows performance issues.
• New and updated files:
  • tests/test_dask_benchmarks.py: Mock benchmark tests with synthetic workloads
  • tests/test_real_txfunc_benchmarks.py: Real-world tax function benchmarks
  • tests/run_benchmarks.py: Automated benchmark runner with reporting
  • tests/BENCHMARK_README.md: Comprehensive documentation and usage guide
  • pytest.ini: Updated with benchmark test markers
• Key features:
  • Platform-specific optimization tests (Windows, macOS, Linux)
  • Memory usage and compute time benchmarking
  • Baseline establishment and performance regression detection
  • Comparison of different Dask schedulers and client configurations
  • Real tax function estimation performance measurement
  • Automated identification of optimal Dask settings per platform
• Benefits:
  • Establishes performance baselines before optimization work
  • Identifies Windows-specific Dask performance bottlenecks
  • Provides automated regression detection for future changes
  • Enables data-driven optimization decisions
  • Supports continuous performance monitoring
• Usage:
  • python tests/run_benchmarks.py # Run all benchmarks
  • python tests/run_benchmarks.py --quick # Quick benchmarks only
  • python tests/run_benchmarks.py --save-baseline # Save performance baseline
  • python tests/run_benchmarks.py --compare-baseline # Compare against baseline
• 🤖 Generated with help from Claude Code

0.14.7

This release:

• Refactors calls to dask in SS.py and TPI.py. See PR #1048.

0.14.6

• Removes initial_guess_w_SS in default_parameters.json
• Updates environment and testing to cover Python 3.13