v1.2.1 - Torun

• [enhancement] Blocks can handle missing geometries (geometry as None or NaN)
• [debug] Blocks default value column is block_values, and PointSupport default value column is values (to avoiding naming collisions during the join operation)
• [debug] Blocks object default index is created when it is not provided by the user
• [enhancement] verbose parameter for PoissonKrigingInput class that controls progress bar when calculating point support distances between neighbors

Torun - 1.2.0

Changes

• updated use cases
• updated badges (code coverage, CI, Python version)
• citation information
• changed API structure - users might pass separate geometry and values for interpolation

Torun

The package was updated to version 1.0 in June 2025. There are breaking API changes, so please, refer to the CHANGELOG to know more about the changes.
Right now, the package in version 1.2.0. There might be some minor bugs, and large swaths of code are not optimized yet. If you find any bugs, please report them in the issue tracker.

v0.5.4

2025-01-04

version 0.5.4

• (python) removed Python 3.8 from the supported versions,
• (debug) fixed DivisionByZeroWarning when the semivariogram range is equal to 0

0.5.3

version 0.5.3

(logic) debugged variance error calculations for Area-to-Point Poisson Kriging

0.5.2

Updated dependencies (GeoPandas without version constrains, numpy < 2, and pylibtiff >= 0.6.1)

v 0.5.1

2024-02-19

version 0.5.1 (pre production development)

• (enhancement) interpolate_raster() function takes allow_approx_solutions parameter, and it protects from LinAlgError that might occur if interpolation points are duplicated (due to the floating point number representation).
• (refactoring) calc_point_to_point_distance function refactored to point_distance, changed input parameters' schema,
• (refactoring) new selection method for unequally spaced bins: select_values_between_lags
• (debug) np.float type casting has been changed to float

v 0.5.0.post1-updated-docs

Release 0.5.0.post1 with updated docs

v0.5.0.post1 - Mykolaiv

version 0.5

• (feature) to_tiff() function which writes kriging output from the interpolate_raster() function to tiff and tfw files,
• (debug) safe theoretical variogram models,
• (enhancement) model_types parameter can be string only (in the future the name of this parameter will be changed),
• (dependencies) fixed dependencies (hdbscan and scikit-learn),
• (enhancement) updated tutorials, we slightly changed their structure,
• (dependencies) End of support for Python 3.7,
• (invalid) Warning when user tries to use .plot() method of the ExperimentalVariogram class,
• (invalid) Default direction and tolerance are None instead of floats,
• (invalid) Removed unnecessary warning from the .autofit() method.

v0.4.2 - Kharkiv

Welcome to the new release of Pyinterpolate!

This release brings multiple new functionalities:

• Cluster detection with DBSCAN,
• Cluster aggregation,
• Gridding algorithm,
• Grid aggregation,
• Theoretical Variogram calculates Spatial Dependence Index,
• Indicator Kriging,
• Blocks and PointSupport read all files supported by GeoPandas.

The package is stable now, and future changes will focus on the speed and the new spatial interpolation algorithms.