Version 1.4.0-alpha

Pre-release for testing network support. Relevant features:

• Added NetworkModelResult to the public API.
• Enhanced the match function and related helpers to support matching network observations and models.
• Introduced classes and methods for comparing observations and network geometries (e.g. NodeObservation).

Version 1.4.0a2

Fixing dependencies for network support pre-release.

Version 1.3.0

Highlights

• Add string option for metrics by @stkistner in #504
• Make it easy to use NetCDF for point data by @ecomodeller in #500
• ModelSkill.load should load both netcdf and zip by @jsmariegaard in #519
• Merge instead of + for combining comparers and collections, remove concatenation in time. by @ecomodeller in #521

What's Changed

• Add string option for metrics by @stkistner in #504
• example notebook by @daniel-caichac-DHI in #425
• Make it easy to use NetCDF for point data by @ecomodeller in #500
• pr bug fix by @daniel-caichac-DHI in #516
• Only call extract on model result types where it makes sense by @ecomodeller in #520
• Use uv for everything by @ecomodeller in #525
• Pin mypy version by @ecomodeller in #523
• Fix warnings in taylor plot by @ecomodeller in #517
• ModelSkill.load should load both netcdf and zip by @jsmariegaard in #519
• Test common calibration workflow - enforce unique model names by @ecomodeller in #514
• Bug: Score of multiple models fails if observations are not in each comparer by @ecomodeller in #526
• Flexible error handling of empty comparers by @ecomodeller in #512
• Support DummyModel in multimodel comparison by @ecomodeller in #529
• Merge instead of + for combining comparers and collections, remove concatenation in time. by @ecomodeller in #521
• Points and tracks are different by @ecomodeller in #532
• Release 1.3 by @ecomodeller in #534

Full Changelog: v1.2.0...v1.3.0

Version 1.2.0

Highlights

• Feature: add toy data by @jsmariegaard in #493
• New @metric decorator by @ecomodeller in #505
• Imrove skill_table styling/spacing in plt.scatter by @stkistner in #503
• Add spatial_overview and temporal_coverage to ComparerCollectionPlotter by @ecomodeller in

Complete list of changes

• Post-release change version to 1.2.dev0 by @jsmariegaard in #482
• Use Quarto / Quartodoc for documentation by @ecomodeller in #421
• Notebook examples in documentation by @ecomodeller in #341
• Branch protection by @ecomodeller in #486
• Documentation minor tweaks by @jsmariegaard in #488
• Add py.typed marker by @ecomodeller in #484
• Add missing ruff lint default rules by @ecomodeller in #495
• Raw model data is PointModelResult by @ecomodeller in #497
• Feature: add toy data by @jsmariegaard in #493
• Use 3.10 syntax with mypy. by @ecomodeller in #501
• Imrove skill_table styling/spacing in plt.scatter by @stkistner in #503
• Add .plot attributes with links to docs by @ecomodeller in #507
• Preserve order of models from dataframe by @ecomodeller in #510
• New @Metric decorator by @ecomodeller in #505
• Aux overlapping name logic by @ecomodeller in #508
• Add spatial_overview and temporal_coverage to ComparerCollectionPlotter by @ecomodeller in #494
• Release 1.2.0 by @ecomodeller in #511

Full Changelog: v.1.1.0...v1.2.0

v.1.1.0 Remove deprecated filtering, more plots, many fixes

What's Changed

• Remove setitem from ComparerCollection by @ecomodeller in #416
• Refactor matching to be Observation type agnostic by @ecomodeller in #413
• Handle warnings, mainly from pandas. by @ecomodeller in #432
• New Peak Ratio by @daniel-caichac-DHI in #435
• Fix problem with dfsu elements read out of order by @ecomodeller in #438
• Prepare some more for NumPy 2.0 by @ecomodeller in #439
• Including notebook on data wrangling by @jpalm3r in #440
• Errors rather than asserts by @ecomodeller in #379
• Fix comparer wrong gtype by @ecomodeller in #436
• Including notebook on forecast skill analysis and visualization by @jpalm3r in #441
• Check order of args in match by @ecomodeller in #444
• plot() should raise error if no default plot by @jsmariegaard in #419
• Fixing issue with x axis label by @jpalm3r in #446
• Use ruff for formatting instead of black by @ecomodeller in #445
• Observation xyz is not included in Comparer for point-to-point comparisons by @jsmariegaard in #447
• Maintenance/make uv installable with 312 by @ryan-kipawa in #454
• Feature/box plot in comparer collection by @ryan-kipawa in #455
• Support for Python 3.13 by @ecomodeller in #459
• Feature/multiple model support for residual hist on comparer by @ryan-kipawa in #458
• Avoid from_matched with track data explodes on missing x_item by @jsmariegaard in #466
• Basic Support for Dfs2 in GridmodelResult by @ecomodeller in #428
• Update README.md by @daniel-caichac-DHI in #467
• Fix some static type info by @ecomodeller in #470
• Make netcdf mandatory by @ecomodeller in #473
• Fix/Scatter plot density with few points by @jsmariegaard in #478
• Update GitHub actions by @jsmariegaard in #479
• Use strings as values for enum instead of integers by @ecomodeller in #469
• Add qq and residual_hist plots on comparer collection by @ryan-kipawa in #456
• Change version number for release 1.1.0 by @jsmariegaard in #480
• Remove deprecated arguments, functions and classes. by @ecomodeller in #481

New Contributors

• @jpalm3r made their first contribution in #440

Full Changelog: v1.0.0...v.1.1.0

A stable release: Version 1.0

The time has come to release ModelSkill 1.0 after lot's of small tweaks during the beta phase.

🙏 to @daniel-caichac-DHI for your patience.

What's Changed

• New dev version 0.9 by @jsmariegaard in #168
• Observation: rename .df to .data by @jsmariegaard in #166
• Remove multi-item-modelresults by @jsmariegaard in #165
• Comparer reorganization by @jsmariegaard in #169
• minimal style changes by @daniel-caichac-DHI in #171
• add warning scatter plot by @daniel-caichac-DHI in #173
• ModelResult isolated refactor by @PaulDanielML in #170
• Comparer internal data structure as xarray Dataset by @jsmariegaard in #172
• ModelResult - docs by @ecomodeller in #176
• Singleobscomparer by @ecomodeller in #175
• New Comparer method where() by @jsmariegaard in #177
• Comparer further refactoring by @ecomodeller in #179
• Rich compare method by @jsmariegaard in #180
• Style limits qq by @daniel-caichac-DHI in #182
• add degree to short names by @daniel-caichac-DHI in #188
• fix small density-bug and added test by @daniel-caichac-DHI in #189
• Replace ItemInfo with fmskill.Quantity by @ecomodeller in #191
• Save/load comparer(s) by @ecomodeller in #178
• Validate metrics by @ecomodeller in #195
• Multiple authors + pyproject.toml by @ecomodeller in #196
• added option for quantile regression in scatter by @clemenscremer in #197
• Matched data in a single dataframe by @ecomodeller in #192
• new logo by @jsmariegaard in #184
• Rename to ModelSkill by @ecomodeller in #199
• Remove explicit geometry check by @ecomodeller in #190
• Refactor scatter() by @ecomodeller in #201
• Testable skill table by @ecomodeller in #202
• Deprecate filtering in skill, score, scatter by @ecomodeller in #200
• Updated vision in docs by @ecomodeller in #210
• Kernel Density Estimation by @ecomodeller in #205
• Update aesthetics of skill table and scatter plot by @stkistner in #212
• Fixes in collaboration with ruff by @ecomodeller in #206
• Don't issue warnings with no arguments by @ecomodeller in #213
• Check extract method instead of type by @ecomodeller in #207
• move plot methods from Connector to root namespace by @jsmariegaard in #187
• Added skill table with subset metrics by @daniel-caichac-DHI in #216
• custom metrics as series by @daniel-caichac-DHI in #215
• Added_z_comparison by @daniel-caichac-DHI in #204
• Improved model to dataframe by @jsmariegaard in #222
• ComparerCollection deprecation warnings by @jsmariegaard in #225
• Fix add comparer track by @jsmariegaard in #224
• Update notebooks to new api by @jsmariegaard in #218
• Extract ComparerPlotter from Comparer & Comparison module in subfolder by @ecomodeller in #214
• Source dist without notebooks and tests by @ecomodeller in #228
• Boxplot by @jsmariegaard in #229
• devcontainer initial setup by @PaulDanielML in #223
• custom metrics in skill table by @daniel-caichac-DHI in #230
• Q-Q plot for easy comparison across multiple models by @jsmariegaard in #231
• Scatter vs histo density normalization by @stkistner in #233
• Filtering in scatter is deprecated, not removed by @ecomodeller in #239
• fit_to_quantile legend by @daniel-caichac-DHI in #238
• Consistent skill plot methods by @ecomodeller in #232
• Documentation update by @ecomodeller in #237
• Keep user defined model name by @ecomodeller in #244
• new_metrics_validation_services by @daniel-caichac-DHI in #241
• moved Peak Ratio by @daniel-caichac-DHI in #246
• Wind/wave rose plot by @ecomodeller in #240
• Support aux data in from_matched() by @jsmariegaard in #248
• Add missing api docs by @jsmariegaard in #252
• Plots directional axis by @jsmariegaard in #247
• Fix type hints by @ecomodeller in #255
• Change from base class to protocol by @ecomodeller in #259
• Remove reporting functionality by @ecomodeller in #253
• Comparer.rename() by @ecomodeller in #254
• Fix type hints due to latest matplotlib by @ecomodeller in #264
• Persist raw model data by @ecomodeller in #261
• move from_matched and compare to new module by @jsmariegaard in #266
• Move plot methods to plotting package by @rywm-dhi in #263
• Pandas deprecations by @ecomodeller in #265
• Use xarray dataset as internal data container by @jsmariegaard in #262
• Fix loading named styles by @ecomodeller in #270
• Make plotting api more consistent by @rywm-dhi in #267
• Remove entrypoint by @ecomodeller in #276
• Change sum in groupby to avoid warning by @jsmariegaard in #273
• More work on internal data structures by @jsmariegaard in #269
• Slim AggregatedSkill by @ecomodeller in #260
• Rename and improve notebooks by @jsmariegaard in #279
• Move config to separate module by @jsmariegaard in #280
• Update_bins by @daniel-caichac-DHI in #282
• Set quantity ASAP instead of mutating by @ecomodeller in #283
• Add a event-based example by @rpalDHI in #274
• Default number of bins in scatter plot --- too many defaults? by @jsmariegaard in #285
• Better matching and time interpolation by @jsmariegaard in #281
• Check kwargs valid by @jsmariegaard in #292
• Update Vision by @jsmariegaard in #295
• Observation attrs container by @jsmariegaard in #291
• Fix wind rose by @ecomodeller in #277
• Unit handling by @ecomodeller in #300
• Expose kwargs PR by @daniel-caichac-DHI in #299
• Allow aux items on observations and modelresults for advanced filtering by @jsmariegaard in #302
• Deprecate Connector by @ecomodeller in #307
• Time is not essential for skill assessment by @ecomodeller in #293
• Use Mkdocs as documentation engine instead of Sphinx by @ecomodeller in #303
• First water resource test case by @jsmariegaard in #249
• Fixed and improved hydrology notebook by @jsmariegaard in #316
• show slowest notebook timings by @jsmariegaard in #317
• Allow aux items in ModelResults by @jsmariegaard in #304
• Link to metrics page by @ecomodeller in #313
• Factory function(s) by @ecomodeller in #308
• Rename SpatialSkill to GriddedSkill by @ecomodeller in #315
• SkillTable is a collection of SkillArrays by @jsmariegaard in #309
• ms.match by @ecomodeller in #321...

Version 1.0 Beta 2

What's Changed

• Fix skill table in multiple model scatter by @ecomodeller in #362
• Addition of Wastewater notebook example and dataset by @ChrisGaszynski-DHI in #336
• Fix random obs colors by @jsmariegaard in #367
• Rename variable to quantity by @jsmariegaard in #365
• skill_table also for plotting.scatter by @ecomodeller in #369
• from_config() return ComparerCollection by @jsmariegaard in #366
• Scatter consistent args by @ecomodeller in #370
• Rename enhanced by @jsmariegaard in #368
• Skill table in plotly scatter by @ecomodeller in #371
• Remove/deprecate more properties by @jsmariegaard in #373
• SkillTable in root namespace by @jsmariegaard in #372
• Skill groupby attrs by @jsmariegaard in #351
• Observation attrs by @ecomodeller in #375
• Refactor long_dataframe by @ecomodeller in #382
• Don't sort on groupby (e.g. in skill() method) by @jsmariegaard in #376
• mikeio.Dataset is no longer a mapping... by @jsmariegaard in #385
• Long-dataframe with time column by @ecomodeller in #384
• Comparer attrs property by @jsmariegaard in #387
• DummyModelResult by @ecomodeller in #380
• Spatial interp method by @jsmariegaard in #383

New Contributors

• @ChrisGaszynski-DHI made their first contribution in #336

Full Changelog: v1.0b1...v1.0b2

Version 1.0 Beta 1

What's Changed

• Rename SkillGrid.field_names to metrics by @jsmariegaard in #344
• rename observation module to obs by @jsmariegaard in #343
• Index into mikeio.Dataset with int | str by @ecomodeller in #348
• Expand docs by @jsmariegaard in #319
• Preserve comparer order on save/load by @ecomodeller in #346
• Include x, y in SkillTable - alternative way by @jsmariegaard in #329
• Use unit.short_name for MIKEIO by @ecomodeller in #349
• Use observation weights in score by @ecomodeller in #342
• Remove start end properties by @jsmariegaard in #353
• Avoid mismatch by @ecomodeller in #350
• PR when no joint peaks by @daniel-caichac-DHI in #355
• Remove unused weight attr from tests by @ecomodeller in #357
• Slim public API; remove properties by @jsmariegaard in #354
• No metrics property on the comparer/collection by @ecomodeller in #356
• Improved mkdocs.yml settings by @jsmariegaard in #358
• Skill groupby dt by @jsmariegaard in #352
• Twek index / getting started pages by @ecomodeller in #360

Full Changelog: v1.0b0...v1.0b1

Version 1.0 Beta

What's Changed

• Make plotting api more consistent by @rywm-dhi in #267
• Remove entrypoint by @ecomodeller in #276
• Change sum in groupby to avoid warning by @jsmariegaard in #273
• More work on internal data structures by @jsmariegaard in #269
• Slim AggregatedSkill by @ecomodeller in #260
• Rename and improve notebooks by @jsmariegaard in #279
• Move config to separate module by @jsmariegaard in #280
• Update_bins by @daniel-caichac-DHI in #282
• Set quantity ASAP instead of mutating by @ecomodeller in #283
• Add a event-based example by @rpalDHI in #274
• Default number of bins in scatter plot --- too many defaults? by @jsmariegaard in #285
• Better matching and time interpolation by @jsmariegaard in #281
• Check kwargs valid by @jsmariegaard in #292
• Update Vision by @jsmariegaard in #295
• Observation attrs container by @jsmariegaard in #291
• Fix wind rose by @ecomodeller in #277
• Unit handling by @ecomodeller in #300
• Expose kwargs PR by @daniel-caichac-DHI in #299
• Allow aux items on observations and modelresults for advanced filtering by @jsmariegaard in #302
• Deprecate Connector by @ecomodeller in #307
• Time is not essential for skill assessment by @ecomodeller in #293
• Use Mkdocs as documentation engine instead of Sphinx by @ecomodeller in #303
• First water resource test case by @jsmariegaard in #249
• Fixed and improved hydrology notebook by @jsmariegaard in #316
• show slowest notebook timings by @jsmariegaard in #317
• Allow aux items in ModelResults by @jsmariegaard in #304
• Link to metrics page by @ecomodeller in #313
• Factory function(s) by @ecomodeller in #308
• Rename SpatialSkill to GriddedSkill by @ecomodeller in #315
• SkillTable is a collection of SkillArrays by @jsmariegaard in #309
• ms.match by @ecomodeller in #321
• Aux items must be unique by @ecomodeller in #318
• Remove markdown library by @ecomodeller in #314
• Use SkillTable to_dataframe() instead of df attribute by @jsmariegaard in #322
• Remove protocols module by @ecomodeller in #323
• Create ComparerCollection in init by @ecomodeller in #320
• Change categorical index to float by @ecomodeller in #325
• Circular metrics (again) by @ecomodeller in #284
• Further improvements to SkillTable and SkillArray by @jsmariegaard in #324
• Kde empty title by @ecomodeller in #306
• Updated_Peak_Ratio by @daniel-caichac-DHI in #333
• Observation factory function by @jsmariegaard in #328
• Improve item specification in from_matched by @jsmariegaard in #327
• Ensure Comparer / ComparerCollection consistency by @ecomodeller in #330
• No warnings by @ecomodeller in #334
• Model type hints by @ecomodeller in #332
• Hide mean_skill_points for now by @ecomodeller in #339
• remove_bias returns new Comparer by @ecomodeller in #338
• Skillgrid.sel by @ecomodeller in #340
• Scatter plot return multiple plots for multiple models by @jsmariegaard in #337

New Contributors

• @rpalDHI made their first contribution in #274

Full Changelog: v1.0.a2...v1.0b0

Third alpha release of 1.0

Many changes, but we are not 100% certain about all details of the public API.

What's Changed

• Source dist without notebooks and tests by @ecomodeller in #228
• Boxplot by @jsmariegaard in #229
• devcontainer initial setup by @PaulDanielML in #223
• custom metrics in skill table by @daniel-caichac-DHI in #230
• Q-Q plot for easy comparison across multiple models by @jsmariegaard in #231
• Scatter vs histo density normalization by @stkistner in #233
• Filtering in scatter is deprecated, not removed by @ecomodeller in #239
• fit_to_quantile legend by @daniel-caichac-DHI in #238
• Consistent skill plot methods by @ecomodeller in #232
• Documentation update by @ecomodeller in #237
• Keep user defined model name by @ecomodeller in #244
• new_metrics_validation_services by @daniel-caichac-DHI in #241
• moved Peak Ratio by @daniel-caichac-DHI in #246
• Wind/wave rose plot by @ecomodeller in #240
• Support aux data in from_matched() by @jsmariegaard in #248
• Add missing api docs by @jsmariegaard in #252
• Plots directional axis by @jsmariegaard in #247
• Fix type hints by @ecomodeller in #255
• Change from base class to protocol by @ecomodeller in #259
• Remove reporting functionality by @ecomodeller in #253
• Comparer.rename() by @ecomodeller in #254
• Fix type hints due to latest matplotlib by @ecomodeller in #264
• Persist raw model data by @ecomodeller in #261
• move from_matched and compare to new module by @jsmariegaard in #266
• Move plot methods to plotting package by @rywm-dhi in #263
• Pandas deprecations by @ecomodeller in #265
• Use xarray dataset as internal data container by @jsmariegaard in #262
• Fix loading named styles by @ecomodeller in #270

New Contributors

• @rywm-dhi made their first contribution in #263

Full Changelog: v1.0.a1...v1.0.a2