v0.2.15

What's new:

• Enhanced Scaled Metric Computation and Loss Function Support @elephaint, @nasaul (#189)
  • Fix: Scaled metrics are now correctly computed against the intended denominator.
  • Feature: Adds new functionality to loss functions to support the use of cutoff_col.
  • Feature: Introduces built-in functionality for computing scaled metrics, eliminating the need for manual for loops previously required for correct calculation.
• [FEAT] Improved DataFrame Compatibility with Narwhals @nasaul (#195)

Bug fixes:

• [FIX] NaN Support in Linex Loss @janrth @nasaul (#188)

Documentation:

• [DOC] Migrate and update documentation @deven367 @nasaul (#181, #194, #196, #197, #198)

v0.2.14

Bumps version

v0.2.13

New features

• [FEAT] Adds tweedied_deviance, linex, cfe and pis metrics. @janrth (#162, #170, #184)
• [FEAT] Normalized Deviation @elephaint (#175)
• [FEAT] Allow partial horizons in backtesting @bartoshqr (#163)

Changes

• Updated setup for new contributor and uv @ngupta23 (#149)
• New documentation pipeline and enhancements @deven367 (#174, #176, #177, #178, #180, #182)
• remove fastcore based tests from pytest @deven367 (#172)
• tests for tweedie_deviance @deven367 (#171)
• remove nbdev @deven367 (#169)
• Updated setup for new contributor and uv @ngupta23 (#149)

Bug Fixes

• [FIX] support freq='S' in id_time_grid @elephaint (#157)
• [FIX] Fix spis @elephaint (#190)
• [FIX] Add new transparency values @MMenchero (#165)
• [FIX]: removing constraint from nbdev due to deprecations @ngupta23 (#159)
• [FIX]: Updates testing for Windows @nasaul (#185)

v0.2.12

New Features

• Add Scaled QL and Scaled MQL @marcozanotti (#148)

Bug Fixes

• fix: Validate freq parameter with data in fill_gaps @AlexMilanor (#154)

Dependencies

• deps(python): drop 3.8 and support 3.13 @jmoralez (#155)

v0.2.11

Bug Fixes

• fix(pandas): use arrays for values and indices in time_features @elephaint (#143)

v0.2.10

Bug Fixes

• fix: invalid escape sequence in string @jmoralez (#138)

Dependencies

• deps(pandas): address frequency alias deprecations in generate_series @jmoralez (#139)

v0.2.9

New Features

• Add MSSE and RMSSE losses @marcozanotti (#136)

v0.2.8

Enhancement

• enh: ignore pandas perf warnings on unvectorized offsets @jmoralez (#133)

v0.2.7

Bug Fixes

• fix: ensure numpy in fill_gaps bounds @jmoralez (#130)

v0.2.6

New Features

• add preprocessing.id_time_grid @jmoralez (#129)

Bug Fixes

• fix: ensure xticks are within plot limits @jmoralez (#125)