Skip to content

README.md

Latest commit

 

History

History
65 lines (50 loc) · 2.23 KB

README.md

File metadata and controls

65 lines (50 loc) · 2.23 KB

Dependencies

The following dependencies are required to build the Python bindings:

Unlike the command-line interface, GDAL is not a requirement, although the Python bindings can work with the GDAL Python bindings if they are available.

Installation

To install the Python bindings, run pip install . from the root of the source tree.

Basic usage

In most cases, the simplest usage is the exact_extract function:

from exactextract import exact_extract

rast = "elevation.tif"
polys = "watersheds.shp"

exact_extract(rast, polys, ["mean", "min", "max"])

This will return a list of GeoJSON-like features with properties containing the values of the requested stats. exact_extract can understand several different types of inputs:

  • raster inputs:

    • a GDAL Dataset
    • a rasterio DatasetReader
    • an xarray DataArray
    • a path to a file that can be opened with one of the above packages
    • a RasterSource object (such as a NumPyRasterSource that wraps a numpy array)

  • vector inputs:

    • a GDAL/OGR Dataset or DataSource
    • a fiona Collection
    • a geopandas GeoDataFrame
    • a path to a file that can be opened with one of the above packages
    • a list of GeoJSON-like features
    • a FeatureSource object
    • a QgsVectorLayer object

The following additional arguments are understood by exact_extract:

  • weights: a raster representing weights to use for stats such as weighted_mean
  • include_cols: a list of columns from the input features to be copied into the output features
  • include_geom: a flag indicating that the input geometry should be copied into the output features
  • strategy: a processing strategy to use, either feature-sequential (iterate over features and read the pixels that intersect each feature) or raster-sequential (iterate over chunks of the raster and process the features that intersect each chunk)
  • max_cells_in_memory: used to control the number of cells that are loaded at one time, for large features
  • output: used to control the output format: either geojson, pandas, qgis or the name of an output file to write to using GDAL.