Releases: DHI/modelskill
Releases · DHI/modelskill
v0.5.1 Better ModelResult itemInfo for pandas and xarray
- Fix bug with missing default itemInfo for pandas ModelResults
- Add itemInfo argument to pandas and xarray ModelResultItem
- DMI api - Better handling of empty datasets
What's Changed
- add max_error to documentation by @jsmariegaard in #107
- DataFrameModelResult missing default itemInfo by @jsmariegaard in #109
- add Python 3.10 to CI by @jsmariegaard in #94
- DMI api - Better handling of empty datasets by @ecomodeller in #108
Full Changelog: v0.5.0...v0.5.1
v0.5.0 Better mean_skill, avoid extrapolation, more plot arguments
What's Changed
- Force the use of keyword arguments by @ecomodeller in #100
- Allow x_item and y_item arguments when creating TrackObservation by @jsmariegaard in #99
- fix mean skill weights by @jsmariegaard in #102
- Title and other argument in plots by @jsmariegaard in #98
- Test notebooks by @jsmariegaard in #105
- Better reporting by @jsmariegaard in #66
- Improved docs with logo by @jsmariegaard in #93
- catch warnings in pytest by @jsmariegaard in #78
- fix failing in_domain validation for extraction from pandas modelresults by @jsmariegaard in #106
- Mean skill weights as dict by @ecomodeller in #103
- pointwise mean skill by @jsmariegaard in #104
- Fix bug when combining comparers by @jsmariegaard in #101
Full Changelog: v0.4.3...v0.5.0
v0.4.3 Minor
- Make file paths relative to configuration file (instead of cwd)
- Check point inside xarray domain
v0.4.2 Horizontal bar plot and more
- Skill: Horizontal bar plot
- Default model name for in simple compare
- Hit ratio assert matching data shapes
- Improved documentation
Same as version 0.4.1
v0.4.1 Horizontal bar plot and more
- Skill: Horizontal bar plot
- Default model name for in simple compare
- Hit ratio assert matching data shapes
- Improved documentation
v0.4.0: New metrics, normalized Taylor plot, other improvements and bugfixes
- Add normalized Taylor plot
- Add aggregate_observations argument to taylor plot
- Add new metrics: max_error and willmott's Index of Agreement
- Improve TrackObservation item argument: now takes both int and string
- Improve TrackObservation input check, now require input to be specified if more than 3 items
- Bugfix: normalize NSE metric with obs
- Bugfix: remove class attributes that caused problems with multiple objects
- Add obs_item to fmskill.compare()
- Improved getitem to handle negative indexes and raise correct errors
- Bugfix: fmskill.compare() select item
- Improved metric documentation
v0.3.6 Minor warning fix
- Fix pandas styler deprecation warning
v0.3.5 Altimetry API updates and load time improvements
- Altimetry API updates
- load time improvements, only import xarray when needed
v0.3.4 Minor
- Add shapely as a dependency
v0.3.3 Track ModelResult as pandas dataframe
- It is now possible to provide a track ModelResult as a pandas DataFrame