3.14.2

This release includes the following fixes and features:

• General

  • Validation now covers file paths contained in CSVs. CSV validation
    will fail if the files listed in a CSV fail to validate.
    #327

  • We have updated validation in several ways that will improve the
    developer experience of working with InVEST models, and we hope will also
    improve the user experience:

    • Symbols in conditional requirement expressions now represent the values
      of parameters instead of whether the value of the parameter is
      sufficient. If a symbol is not present in args as a key, the
      symbol will have a value of False. This allows for value-based
      comparisons, which is useful in models that have overlapping modes of
      operation. #1509
    • Vector fields, CSV rows/columns and the 1st level of directory
      contents may now all be conditionally required based on a python
      expression.
    • Under certain circumstances, validation may return more warnings than
      before. This specifically applies to model inputs that have conditional
      requirement expressions where their expression evaluates to False,
      and the user has provided a value for this parameter. Previous
      versions of InVEST would skip these parameters' type-specific
      validation. Now, these parameters will be validated with their
      type-specific validation checks.

  • Add support for latest GDAL versions; remove test-specific constraint on
    GDAL versions from invest requirements.
    #916

  • Updated to Cython 3 (#556)

• Annual Water Yield
  • Added the results_suffix to a few intermediate files where it was
    missing. #1517
• Coastal Blue Carbon
  • Updated model validation to prevent the case where a user provides only
    one snapshot year and no analysis year
    (#1534 <https://github.com/natcap/invest/issues/1534>_).
    Also enforces that the analysis year, if provided, is greater than the
    latest snapshot year. An analysis year equal to the latest snapshot year
    is no longer allowed.
• Coastal Vulnerability
  • Fixed a bug in handling nan as the nodata value of the bathymetry
    raster. nan pixels will now be propertly ignored before calculating
    mean depths along fetch rays.
    #1528
• HRA
  • Fixed a bug where habitat and stressor vectors were not being rasterized
    with the ALL_TOUCHED=TRUE setting.
• Scenic Quality
  • Fixed an issue with viewshed calculations where some slight numerical
    error was introduced on M1 Macs, but not on x86-based computers. This
    numerical error was leading to slightly different visibility results.
    #1562
• SDR
  • Fixed an issue encountered in the sediment deposition function where
    rasters with more than 2^32 pixels would raise a cryptic error relating
    to negative dimensions. #1431
  • Optimized the creation of the summary vector by minimizing the number of
    times the target vector needs to be rasterized.
• Seasonal Water Yield
  • Fixed an issue with the precip directory units. Units for these input
    rasters are now correctly stated as mm/month.
    #1571
  • Fixed an issue where the monthly quickflow values were being summed over
    a block area and not summed pixelwise. This caused the quickflow
    output QF.tif to have malformed values.
    #1541
• Wind Energy
  • Fixed a bug where some number inputs were not being properly cast to
    float or int types. If the inputs happened to be passed as
    a str this caused unintended side effects such as a concatenation
    error. (#1498)
• Urban Nature Access
  • Fixed a NameError that occurred when running the model using
    search radii defined per population group with an exponential search
    kernel. #1502
  • Fixed an issue where Urban Nature Access would crash if an administrative
    boundary geometry did not overlap any people in the population raster.
    #1503
  • Fixed an issue where validation was failing to catch missing values in
    the uniform search radius args key when using uniform search radii.
    #1509
  • Fixed an issue where the output administrative units vector's
    Pund_adm and Povr_adm fields representing undersupplied and
    oversupplied populations, respectively, had values of 0 when running the
    model with search radii defined per population group. The output
    administrative units vector now has the correct values for these fields,
    consistent with the user's guide chapter.
    #1512
  • Fixed an issue where certain nodata values were not being handled
    correctly, leading to pixel values of +/- infinity in the urban nature
    balance output raster. #1519
  • Fixed an issue where an LULC raster without a nodata value would
    always raise in exception during reclassification.
    #1539