From 425da4363beceb97431012d60d205ad8b2107236 Mon Sep 17 00:00:00 2001
From: stuckyb <stuckyb@users.noreply.github.com>
Date: Fri, 6 Oct 2023 09:58:31 -0400
Subject: [PATCH] Added barely a start to a RAP GSDataSet class. I mean,
 barely.

---
 src/library/datasets/rap-annual.py | 137 +++++++++++++++++++++++++++++
 1 file changed, 137 insertions(+)
 create mode 100644 src/library/datasets/rap-annual.py

diff --git a/src/library/datasets/rap-annual.py b/src/library/datasets/rap-annual.py
new file mode 100644
index 0000000..2dad274
--- /dev/null
+++ b/src/library/datasets/rap-annual.py
@@ -0,0 +1,137 @@
+
+from .gsdataset import GSDataSet
+from pyproj.crs import CRS
+import datetime
+import rioxarray
+import api_core.data_request as dr
+from subset_geom import SubsetPolygon, SubsetMultiPoint
+
+
+class RAPVegCover(GSDataSet):
+    def __init__(self, store_path):
+        """
+        store_path (Path): The location of on-disk dataset storage.
+        """
+        super().__init__(store_path, 'rap-annual')
+
+        # Basic dataset information.
+        self._id = 'RAPvegcover'
+        self.name = 'RAP vegetation cover'
+        self.url = 'https://rangelands.app/'
+        self.description = (
+            'The Rangeland Analysis Platform’s vegetation cover product '
+            'provides annual percent cover estimates from 1986 to present of: '
+            'annual forbs and grasses, perennial forbs and grasses, shrubs, '
+            'trees, and bare ground. The estimates were produced by combining '
+            '75,000 field plots collected by BLM, NPS, and NRCS with the '
+            'historical Landsat satellite record. Cover estimates are '
+            'predicted across the United States at 30m resolution, an area '
+            'slightly larger than a baseball diamond.'
+        )
+
+        # Provider information
+        self.provider_name = ''
+        self.provider_url = ''
+
+        # CRS information.
+        self.crs = CRS.from_epsg(4326)
+
+        # The grid size.
+        self.grid_size = 53899.0 / 200000000
+        self.grid_unit = 'degrees'
+
+        # The variables/layers/bands in the dataset.
+        self.vars = {
+            'ppt': 'total precipitation (rain+melted snow)', 
+            'tmean': 'mean temperature (mean of tmin and tmax)',
+            'tmin': 'minimum temperature', 
+            'tmax': 'maximum temperature',
+            'tdmean': 'mean dew point temperature',
+            'vpdmin': 'minimum vapor pressure deficit',
+            'vpdmax': 'maximum vapor pressure deficit'
+        }
+
+        # Temporal coverage of the dataset.
+        self.date_ranges['year'] = [
+            datetime.date(1986, 1, 1), datetime.date(2021, 1, 1)
+        ]
+        self.date_ranges['month'] = [
+            None, None
+        ]
+        self.date_ranges['day'] = [
+            None, None
+        ]
+
+        # Temporal resolution.
+        self.temporal_resolution['year'] = '1 year'
+
+        # File name patterns for each PRISM variable.  Note that for
+        # precipation data, the current version of "M2" for years < 1981 and
+        # "M3" for years >= 1981.  See
+        # https://prism.oregonstate.edu/documents/PRISM_datasets.pdf for
+        # details.
+        self.fpatterns = 'PRISM_{0}_stable_4km{1}_{2}_bil.bil'
+
+    def getData(
+        self, varname, date_grain, request_date, ri_method, subset_geom=None
+    ):
+        """
+        varname: The variable to return.
+        date_grain: The date granularity to return, specified as a constant in
+            data_request.
+        request_date: A data_request.RequestDate instance.
+        ri_method: The resample/interpolation method to use, if needed.
+        subset_geom: An instance of SubsetGeom.  If the CRS does not match the
+            dataset, an exception is raised.
+        """
+        # Get the path to the required data file.
+        if date_grain == dr.ANNUAL:
+            fname = self.fpatterns.format(varname, 'M3', request_date.year)
+        elif date_grain == dr.MONTHLY:
+            datestr = '{0}{1:02}'.format(request_date.year, request_date.month)
+            fname = self.fpatterns.format(varname, 'M3', datestr)
+        elif date_grain == dr.DAILY:
+            datestr = '{0}{1:02}{2:02}'.format(
+                request_date.year, request_date.month, request_date.day
+            )
+            fname = self.fpatterns.format(varname, 'D2', datestr)
+        else:
+            raise ValueError('Invalid date grain specification.')
+
+        # For precipation data, the current version of "M2" for years < 1981
+        # and "M3" for years >= 1981.  See
+        # https://prism.oregonstate.edu/documents/PRISM_datasets.pdf for
+        # details.
+        if (request_date.year < 1981 and date_grain is not dr.DAILY and 
+            varname == 'ppt'
+        ):
+            fname = fname.replace('M3', 'M2')
+
+        fpath = self.ds_path / fname
+
+        # Open data file
+        data = rioxarray.open_rasterio(fpath, masked=True)
+
+        if subset_geom is not None and not(self.crs.equals(subset_geom.crs)):
+            raise ValueError(
+                'Subset geometry CRS does not match dataset CRS.'
+            )
+
+        if isinstance(subset_geom, SubsetPolygon):
+            # Drop unnecessary 'band' dimension because rioxarray
+            # can't handle >3 dimensions in some later operations
+            data = data.squeeze('band')
+
+            data = data.rio.clip([subset_geom.json], all_touched = True)
+        elif isinstance(subset_geom, SubsetMultiPoint):
+            # Interpolate all (x,y) points in the subset geometry.  For more
+            # information about how/why this works, see
+            # https://xarray.pydata.org/en/stable/user-guide/interpolation.html#advanced-interpolation.
+            res = data.interp(
+                x=('z', subset_geom.geom.x), y=('z', subset_geom.geom.y),
+                method=ri_method
+            )
+            data = res.values[0]
+
+        return data
+