TUW-GEO / pytesmo

85%

master: 88%

LAST BUILD ON BRANCH fix-comparison-of-other-dataset-combos

branch: fix-comparison-of-other-dataset-combos

CHANGE BRANCH

x

Reset

• fix-comparison-of-other-dataset-combos
• develop
• master
• v0.2.4
• feature/ismn-dateplot
• feature/confidence_intervals
• python3-support
• appveyor-support
• skip-h08-tests
• remove-grid-stuff
• triple-col-fix-24
• tcol-fix-24
• fix_numpy_installation
• automate-documentation-build
• fix-grids-deprecation-warning
• v0.3.0
• test-data-repo
• add-processing-framwork-tests
• fix-temporal-matching
• v0.3.1
• v0.3.2
• filters-fix
• appveyor-conda
• v0.3.4
• refactor-scaling
• v0.3.5
• update-pyscaffold
• improve-climatology
• v0.3.6
• remove-grids
• fix-boxcar-size1
• recursive_r
• H-SAF-pygeobase
• fix-h07-return-values
• update-docs-minimize-pypi-package
• fix-resample-dtypes
• fix-ismn-plotting
• v0.4.0
• fix-RuntimeError-DataManager
• fix-temporal-resampling-series
• improve-ismn-metadata-reading
• validation-framework-refactor
• fix-ascat-segfault
• make-validation-framework-more-modular
• v0.5.1
• fix-empty-dataframes-in-validation
• add-adapters
• improve-ismn-performance
• v0.5.2
• fix-hsaf-ssm-mean
• use-pybufr-ecmwf-conda-in-travis
• add-julian-tests
• add-hamming-window
• ismn-refactor-common-interfaces
• remove-ascat
• add-additional-jd-test
• v0.6.0
• add-tc-metrics-calc
• pybufr-ecmwf-conda-forge
• rtd
• fix-return-clim
• ubrmsd-test
• fix-julian2datetime-bug
• v0.6.1
• v0.6.2
• python-3-6
• add-poets-timedate-stuff
• temporal-matching-performance
• v0.6.3
• refactor-resample-function
• v0.6.4
• add-dekad-functions
• v0.6.5
• v0.6.6
• fix-appveyor-versioning
• v0.6.7
• fix-docs
• scaling-validation-fw
• ERS-reader-adaption
• v0.6.8
• move-ismn-to-package
• improve_doy_testing
• fix-plotting-test
• fix-exponential-filter-initialization
• v0.6.9
• import_ascat0.9
• update-test-data-repo
• fix-exponential-filter-nan
• v0.6.10
• fix-python36-test
• pyscaff_update
• rolmetrics
• remove-dos-line-endings
• advanced_masking_adapter
• v0.9
• v0.9.1
• climatology-less-than-full-year-leapyear
• v0.10.x
• v0.11.3
• make-mse-test-less-sensitive
• v0.11.4
• v0.12.0
• fix_ci
• v0.13.0
• v0.13.1
• v0.13.2
• v0.13.3
• v0.13.4
• v0.14.0
• v0.14.1
• s-scherrer-patch-1
• v0.14.2
• v0.14.3
• v0.14.4
• s-scherrer-patch-2
• v0.15.0
• v0.15.1
• v0.15.2
• v0.16.0
• fix-seas-val
• v0.16.1
• v0.17.0

Build # 978

Build Type

push

travis-ci

Committed by cpaulik

Commit Message

Val Framework compares dataset matches that do not include the temporal reference

This changes the behavior of the validation framework.

Before this a validation run with `(n, k)` where `k < n` only gave results for
combinations of datasets which included the temporal reference.

e.g. If 3 dataset (DS1, DS2 and DS3) where given and a metric calculator was run
with `(3, 2)` with the temporal reference DS1. Then only the combinations of
`DS1 - DS2` and `DS1 - DS3` gave results.

This rectifies this behavior to also return results for the combination `DS2 -
DS3` in the above example.

Tests still need to be adapted but I want to discuss this first to see if this
is what people expect.

There are also some edge cases. e.g. If 4 datasets are given but the
configuration of the metric calculator would be `(3, 2)`. I think we should
raise an error in these cases because that would mean that the temporal
reference is not included in all matches.

Run Details

577 of 743 branches covered (77.66%)

24 of 24 new or added lines in 2 files covered. (100.0%)

1700 of 1990 relevant lines covered (85.43%)

2.56 hits per line