Nixtla/hierarchicalforecast v1.3.0

on GitHub

What's new

• [FEAT] Probabilistic methods for MiddleOut by @elephaint in #382
• [FEAT] Add python 3.13 by @elephaint in #376
• [FEAT] migrate tests from nbs to pytest by @deven367 in #384

Documentation

• [DOC] Adds example detailing the use of different forecasting methods when doing HF by @nasaul in #378
• [DOC] New documentation pipeline by @deven367 in #386
• [DOC] Documentation enhancements by @nasaul in #396
• [DOC] Correctly links to Multimodel example by @nasaul in #399
• [DOC] fixes for introduction nb by @deven367 in #387
• [DOC] Doc enhancements by @deven367 in #388

Bug fixes

• [FIX] Add error message when y_hat contains missing timestamps by @elephaint in #367
• [FIX] Add utilsforecast to dependencies by @elephaint in #366
• [FIX] TopDown with "forecast_proportions" should error when providing a non-strict hierarchy. by @elephaint in #368
• [FIX] add returning fig to plot methods by @khuyentran1401 in #371
• [FIX] Revert documentation clearing of "add returning fig to plot methods" by @elephaint in #372
• [FIX] Remove conda suggestions by @elephaint in #377
• [FIX] Update CI, tests and deprecations by @elephaint in #373
• [FIX] Improve type hints for probabilistic methods and fix small bugs by @elephaint in #374
• [FIX] Remove bilinearity in Wh calculation by @elephaint in #380
• [FIX] Deprecate S in favor of S_df by @elephaint in #379
• [FIX] use backend= instead of deprecated native_namespace= in nw.from_dict by @MarcoGorelli in #381

New Contributors

• @khuyentran1401 made their first contribution in #371
• @MarcoGorelli made their first contribution in #381
• @deven367 made their first contribution in #384

Full Changelog: v1.2.1...v1.3.0