Releases: pySTEPS/pysteps
pysteps v1.6.2
What's Changed
- Minor exporters adjustment by @RubenImhoff in #272
- Place ensemble member number determination for blending inside forecast loop to prevent out of memory issues by @RubenImhoff in #273
- Ensure clim window has same size as skill today and place forecast_prev update at end of extrapolation loop by @RubenImhoff in #274
Full Changelog: v1.6.1...v1.6.2
pystep v1.6.1
- Fix duplicate link in the README
- Fix imports of optional dependencies in
pysteps.blending.utils
- Skip tests if missing optional dependency
pysteps v1.6.0
What's Changed
New features
- New STEPS blending module by @RubenImhoff @cvelascof @ladc @wdewettin in #233 (including #231, #232, #236, #255)
- New linear blending module by @wdewettin in #229
- New SAL verification module by @EsmailGhaemi in #248
Docs and others
- Update contributing guide by @aperezhortal in #241
- Use new theme for docs by @dnerini in #259
- Replace miniconda by micromamba in CI by @aperezhortal in #257
- Apply black new stable release (22.1.0) by @dnerini in #261
- Set minimum python version to 3.7 by @dnerini in #253
Breaking changes
New Contributors
- @EsmailGhaemi made his first contribution in #248
- @wdewettin made his first contribution in #229 and #233
- @ladc made her first contribution in #233
Full Changelog: v1.5.1...v1.6.0
pysteps v1.5.1
This release includes minor fixes and documentation updates.
What's Changed
- Update contributing guide by @aperezhortal in #241
- document tdating by @feldmann-m in #245
- Fix compatibility with matplotlib>=3.5 by @dnerini in #246
- Add project_urls by @dnerini in #247
New Contributors
- @feldmann-m made their first contribution in #245
Full Changelog: v1.5.0...v1.5.1
pysteps v1.5.0
Pysteps v1.5.0 is now available with some exciting new features! We would like to thank everyone who contributed to this release.
Key updates
- Add new ensemble nowcast model: Lagrangian INtegro-Difference equation model with Autoregression (LINDA) by Pulkkinen et al. (2021). [#221, gallery]
- Implement the local Lagrangian approach for probabilistic nowcasting by Germann and Zawadzki (2004). [#207, gallery]
- Refactor the
utils.interpolate
module. Therbfinterp2d
method now wraps the scipy's Rbf class, while the old version is renamed toidwinterp2d
(inverse distance interpolation). [#210] - Implement the
max_num_features
keyword argument for all methods in thepysteps.feature
module. With this argument, the user can specify the maximum number of detected features. [#225]
Bug fixes
- Fix a bug in
utils.spectral.rapsd
function causing negative frequency for evenly-sized images. [#211] - Fix coordinates of the bounding box in
io.importers.import_mrms_grib
. [#222]
Installation
You can upgrade to the latest release using pip install pysteps --upgrade
or conda update pysteps -c conda-forge
. It is also possible to build from source following these instructions.
Important note
This release represents our last feature release for pysteps v1, while our efforts will now focus on the upcoming v2! See #216 to learn more about this.
Contributors
Thanks to the following developers for their contributions in this release (alphabetical order):
@aperezhortal, @dnerini, @loforest, @pulkkins, @RubenImhoff
pysteps v1.4.1
Summary
- Major refactoring of the
pysteps.visualization
module to improve the code quality, documentation, and coverage. As a consequence, there have been few changes in the interfaces to the plotting functionalities that will produce new deprecation warnings for v1.6 [#199] - Fix incompatibility that appeared with scipy>=1.6 causing
pysteps.nowcasts.steps.forecast
to fail with argumentvel_pert_method="bps"
and a zero velocity field [#203] - Fix bug introduced in v1.4.0 concerning the usage of the netcdf exporter in
pysteps.nowcasts.steps.forecast
[#200] - Fix bug in
pysteps.utils.cleansing.decluster
[#194]
For a more comprehensive list of all changes please have a look at the release v1.4.1 Milestone.
Contributors
pysteps v1.4
Summary
- Introducing the thunderstorm detection and tracking (DATing) modules [example] [doc1] [doc2] [#178]
- Introducing the RainFARM (Rebora et al. 2006) module for downscaling precipitation fields [example] [doc] [#173]
- Introducing the pysteps importers plugins [doc] [#174]
- Improving the semi-lagrangian extrapolation method [doc] [#176] [#191]
- Improving pysteps netCDF exporter [doc] [#175]
- Implementing a more general importer for ODIM-compliant data sources [doc] [#171] [#172]
- Improving plotting of basemaps [doc] [#177] [#180] [#192]
... and plenty of smaller bug fixes and general improvements. For more info please have a look at the release v1.4 Milestone.
Breaking changes
pysteps.extrapolation.semilagrangian
D_prev
argument renamed todisplacement_prev
[#184]
pysteps.utils.images
pysteps.motion.lucaskande
pysteps.motion.lucaskanade.track_features
method migrated topysteps.tracking.lucaskanade.track_features
[#178]
pysteps.visualization
Contributors
@aperezhortal @chiara-arpae @dnerini @edigiacomo @feldmann-m @jleinonen @loforest @pulkkins @RubenImhoff
pysteps v1.3.2
Summary of main changes :
- Improve documentation
- Fix bug in the VET padding
- Fix visualization routines for motion fields
- Update default
pystepsrc
- Minor code adjustments and fixes
Pysteps v1.3.2 changelog:
cascade
- Compute cascade level statistics by default in the
decomposition.decomposition_fft
function.
- Compute cascade level statistics by default in the
motion
- Fix the issue with the padding keyword in the
motion.vet
function to correctly handle masked and non-masked input arrays.
- Fix the issue with the padding keyword in the
noise
- Add a very small offset to the tapering window to avoid zero values causing some artifacts with the
ssft
andnested
filters.
- Add a very small offset to the tapering window to avoid zero values causing some artifacts with the
nowcasts
- Set out-of-domain values to nan in
sprog.forecast
andsteps.forecast
. - Set negative values in
anvil.forecast
to zero. - Add option to disable localization in
anvil.forecast
by passingar_window_radius=None
. - Change default to
probmatching_method='cdf'
insprog.forecast
.
- Set out-of-domain values to nan in
pystepsrc
- Convert the value of pixelsize to km for
knmi
data.
- Convert the value of pixelsize to km for
verification
- Fix error in
plot.plot_rankhist
caused by a different number of ticks and tick labels.
- Fix error in
visualization
- Fix interfaces in
motionfields
routines with respect tobasemaps.plot_geography
. - Update docstrings in
motionfields
.
- Fix interfaces in
Future Warnings:
v1.4.0
:- Rename keyword argument
map
toplot_map
for all routines in thevisualization
module.
- Rename keyword argument
pysteps v1.3.1
Summary of main changes and new features:
- Add an interface to the
feature.blob_*
methods from scikit-image. A blob is defined as a local maximum of a Gaussian-filtered image. This newutils.images.blob_detection
feature detection method can be used in themotion.lucaskanade
module to track cells in radar data. - Fix installation errors in windows which caused permissions errors in
datasets.download_pysteps_data
anddatasets.download_mrms_data
functions. - Fix bug with the precipitation mask of
nowcasts.anvil
. This caused the method to produce too much rain in non-rainy areas when rain rate was used as input. - Run Black on the whole source code for consistent, reproducible code formatting.
- Add pre-commit hooks as part of the development workflow.
- Improve documentation.
Pysteps v1.3.1 changelog:
dataset
- Fix permissions errors in
download_pysteps_data
anddownload_mrms_data
functions - Update installation instructions for Windows systems
- Fix permissions errors in
motion.lucaskanade.dense_lucaskanade
- Add option to use
blob_detection
as feature detection method
- Add option to use
nowcasts.anvil
- Implement precipitation mask when rainrate is set to None
- Minor changes to default values, masking and thresholding
utils.images
- Add
blob_detection
method
- Add
pysteps v1.3
Summary of main changes and new features:
- Code formatting and bug fixes.
- Improve documentation and test coverage.
- New deterministic nowcasting module
nowcasts.anvil
implementing the ANVIL method published in Pulkkinen et al. (2020): Nowcasting of Convective Rainfall Using Volumetric Radar Observations. Learn how to use it in this example script. - New importer
io.importers.import_mrms_grib
for the for NSSL’s Multi-Radar/Multi-Sensor System (MRMS). - New
dataset
module to facilitate downloading the pysteps example data and create the configuration file. - New interactive notebook
my_first_nowcast.ipynb
available on Google Colab.
Pysteps v1.3.0 changelog:
dataset
- add function
download_pysteps_data
to download the example data from the pysteps-data GitHub repo - add function
create_default_pystepsrc
to create a default configuration file pointing to the local pysteps data directory - add helper functions to easily load the example data for each data source (fmi, mch, opera, etc)
- add function
io.importers
:- add function
import_mrms_grib
to import NSSL’s Multi-Radar/Multi-Sensor System (MRMS) grib data - change spatial units from meters to kilometers in
import_knmi_hdf5
and fix definition ofy1
andy2
parameters in the metadata. - allow all the importers to specify the output array data type (dtype keyword) and fill invalid data with a predefined value (fillna keyword)
- add function to safely and consistently get the rain/no rain threshold
- add function
motion.constant
constant
method returns two-dimensional advection field instead of one motion vector
motion.lucaskanade
- fix bug in
dense_lucaskanade
handling of MaskedArrays inputs. Previously, the method ignored the input mask, unless the masked values where set as nans - require ndarrays or MaskedArrays as input data to
dense_lucaskanade
- fix bug in
motion.vet
- fix bug in
vet
keywords handling. Previously, the method argument passed to the minimizer was ignored
- fix bug in
nowcasts
- add deterministic nowcasting module
nowcasts.anvil
- add deterministic nowcasting module
utils.dimensions
- include option
trim
inaggregate_fields
function to allow aggregation along axes that are not perfectly divisible by a given window
- include option
utils.images
- rename
ShiTomasi_detection
function to lowercaseshitomasi_detection
to comply with PEP8 - change default value for the
buffer_mask
argument in theshitomasi_detection
function from 0 to 5 to provide a minimum buffer to no-data in themotion.lucaskanade.dense_lucaskanade
method
- rename
utils.interpolate
- change default value for the
epsilon
argument in therbfinterp2d
function from 5 to 10 to provide a more localized interpolation of motion vectors in themotion.lucaskanade.dense_lucaskanade
method
- change default value for the
Contributors to this release:
@aperezhortal @dnerini @pulkkins @RubenImhoff