Adding support for dataset categories (for now optional)

Categories will be added to keywords.

PiperOrigin-RevId: 729274084

checker/node/__init__.py

@@ -11,6 +11,7 @@
 from checker.node import feature_view
 from checker.node import file_path
 from checker.node import gee_bitmask
+from checker.node import gee_categories
 from checker.node import gee_classes
 from checker.node import gee_is_derived
 from checker.node import gee_provider_ids
@@ -69,6 +70,7 @@
     version_extension.Check,
     # gee extensions
     gee_bitmask.Check,
+    gee_categories.Check,
     gee_classes.Check,
     gee_is_derived.Check,
     gee_provider_ids.Check,

checker/node/gee_categories.py

@@ -0,0 +1,62 @@
+"""Checks that the "gee:categories" field is valid."""
+
+from typing import Iterator
+
+from checker import stac
+
+# Use dashes, not underscores, as the HTML catalog converts
+# underscores to dashes anyway.
+CATEGORIES = frozenset([
+    'agriculture',
+    'atmosphere',
+    'climate',
+    'cryosphere',
+    'ecosystems',
+    'elevation-topography',
+    'fire',
+    'forest-biomass',
+    'hydrology',
+    'infrastructure-boundaries',
+    'landuse-landcover',
+    'oceanography',
+    'orthophotos',
+    'plant-productivity',
+    'population',
+    'precipitation',
+    'satellite-imagery',
+    'soil',
+    'vegetation-indices',
+])
+
+GEE_CATEGORIES = 'gee:categories'
+
+
+class Check(stac.NodeCheck):
+  """Checks the gee:categories field."""
+
+  name = 'gee_categories'
+
+  @classmethod
+  def run(cls, node: stac.Node) -> Iterator[stac.Issue]:
+    if node.type == stac.StacType.CATALOG:
+      return
+
+    if GEE_CATEGORIES not in node.stac:
+      # TODO(b/397988701): require categories when all datasets have them
+      # yield cls.new_issue(node, f'Missing {GEE_CATEGORIES}')
+      return
+
+    categories = node.stac[GEE_CATEGORIES]
+    if not categories:
+      yield cls.new_issue(node, f'"{GEE_CATEGORIES}" must not be empty')
+      return
+
+    if not isinstance(categories, list):
+      yield cls.new_issue(node, f'"{GEE_CATEGORIES}" must be a list of strings')
+      return
+
+    extra_categories = set(categories) - CATEGORIES
+    for extra in extra_categories:
+      yield cls.new_issue(
+          node, f'Found unknown category "{extra}" in "{GEE_CATEGORIES}"'
+      )

checker/node/gee_categories_test.py

@@ -0,0 +1,38 @@
+from absl.testing import absltest
+from checker import test_utils
+from checker.node import gee_categories
+
+
+class GeeCategoiresTest(test_utils.NodeTest):
+
+  def setUp(self):
+    super().setUp()
+    self.check = gee_categories.Check
+
+  def test_catalog(self):
+    self.assert_catalog({})
+
+  def test_not_a_list(self):
+    self.assert_collection(
+        {'gee:categories': 77}, '"gee:categories" must be a list of strings'
+    )
+
+  def test_missing_ok(self):
+    self.assert_collection({})
+
+  # TODO(b/397988701): require categories when all datasets have them
+  # def test_empty_categories_is_bad(self):
+  #   self.assert_collection({'gee:categories': []}, 'must not be empty')
+
+  def test_valid_category(self):
+    self.assert_collection({'gee:categories': ['soil']})
+
+  def test_bad_category(self):
+    self.assert_collection(
+        {'gee:categories': ['bogus']},
+        'Found unknown category "bogus" in "gee:categories"',
+    )
+
+
+if __name__ == '__main__':
+  absltest.main()

checker/node/top_level.py

@@ -17,6 +17,7 @@
 COLLECTION_ALLOWED_FIELDS = [
     'description',
     'extent',
+    'gee:categories',
     'gee:interval',
     'gee:is_derived',
     'gee:min_zoom_level',