Convert unofficial test case to official one

[cecileherr]

I created a fake dataset to be able to test niche_vlaanderen more or less comprehensively. The input layers contain all possible values for the categorical variables and a wide range of representative values for the continuous variables (a.o. the groundwater levels). This gives me a raster of 694 x 657 pixels (resolution 5 m).

Since this has become my go-to dataset when it comes to testing new functionalities, it would be handy to be able to share it with other developers (or even users?).

The question is: how do we do that?

More info:

Input layers:

• soil, mhw, mlw, msw, management, inundation, seepage, minerality, nitrogen deposition, organic fertilization (nitrogen_animal), anorganic fertilization (nitrogen_fertilizer), inundation_nutrient, inundation_acidity (3,12 Mo per file => 40 Mo in total)
• different variants of these input layers used for specific questions (e.g. groundwater levels with sign for niche 1.x vs niche 2.x, rasters in different formats such as Int, Float, ..., rasters with a bigger or smaller extent, ...) => seems optional to me (I would not integrate them as default in a test set)

Vegetation layer: to be used with the validation module

• shapefile with irregular vegetation polygons (extracted from the real habitatmap)
• shapefiles with square vegetation polygons (with the most habitats/vegetations represented)
• = 3 Mo for both files (but I would not use a shp format for sharing, so still need to check how big if we convert that to geopackages for instance)

What I do not have yet (but could be created if we want to have a comprehensive dataset): input layers for the Flooding module (depths, frequency, period and duration.)

I also have jupyter notebooks using this dataset:

• run a simple model, a full model, a full model with post-hoc restriction based on inundation/management
• with different options: creation of deviation maps, of a simple vs detailed legend, with a customized niche table (ct_vegetation = ), with a custom acidity layer
• run zonal statistics per shape (zonal_stats), comparison of two models (NicheDelta), validation module