Neat 347 extend depth transformer with typing

Makefile

@@ -1,5 +1,5 @@
 .PHONY: run-explorer run-tests run-linters build-ui build-python build-docker run-docker compose-up
-version="0.85.9"
+version="0.85.10"
 run-explorer:
 	@echo "Running explorer API server..."
 	# open "http://localhost:8000/static/index.html" || true

cognite/neat/_version.py

@@ -1 +1 @@
-__version__ = "0.85.9"
+__version__ = "0.85.10"

cognite/neat/graph/queries/_base.py

@@ -2,6 +2,7 @@
 from typing import Literal, cast, overload
 
 from rdflib import RDF, Graph, URIRef
+from rdflib import Literal as RdfLiteral
 from rdflib.query import ResultRow
 
 from cognite.neat.rules.models.entities import ClassEntity
@@ -18,6 +19,24 @@ def __init__(self, graph: Graph, rules: InformationRules | None = None):
         self.graph = graph
         self.rules = rules
 
+    def summarize_instances(self) -> list[tuple]:
+        """Summarize instances in the graph store by class and count"""
+
+        query_statement = """ SELECT ?class (COUNT(?instance) AS ?instanceCount)
+                             WHERE {
+                             ?instance a ?class .
+                             }
+                             GROUP BY ?class
+                             ORDER BY DESC(?instanceCount) """
+
+        return [
+            (
+                remove_namespace_from_uri(cast(URIRef, cast(tuple, res)[0])),
+                cast(RdfLiteral, cast(tuple, res)[1]).value,
+            )
+            for res in list(self.graph.query(query_statement))
+        ]
+
     def list_instances_ids_of_class(self, class_uri: URIRef, limit: int = -1) -> list[URIRef]:
         """Get instances ids for a given class
 
@@ -68,7 +87,10 @@ def triples_of_type_instances(self, rdf_type: str) -> list[tuple[str, str, str]]
             # We cannot include the RDF.type in case there is a neat:type property
             return [remove_namespace_from_uri(*triple) for triple in result if triple[1] != RDF.type]  # type: ignore[misc, index]
         else:
-            warnings.warn("No rules found for the graph store, returning empty list.", stacklevel=2)
+            warnings.warn(
+                "No rules found for the graph store, returning empty list.",
+                stacklevel=2,
+            )
             return []
 
     def construct_instances_of_class(self, class_: str, properties_optional: bool = True) -> list[tuple[str, str, str]]:
@@ -95,7 +117,10 @@ def construct_instances_of_class(self, class_: str, properties_optional: bool =
             # We cannot include the RDF.type in case there is a neat:type property
             return [remove_namespace_from_uri(*triple) for triple in result if triple[1] != RDF.type]  # type: ignore[misc, index]
         else:
-            warnings.warn("No rules found for the graph store, returning empty list.", stacklevel=2)
+            warnings.warn(
+                "No rules found for the graph store, returning empty list.",
+                stacklevel=2,
+            )
             return []
 
     def list_triples(self, limit: int = 25) -> list[ResultRow]:

cognite/neat/graph/stores/_base.py

@@ -5,6 +5,7 @@
 from pathlib import Path
 from typing import cast
 
+import pandas as pd
 from rdflib import Graph, Namespace, URIRef
 from rdflib.plugins.stores.sparqlstore import SPARQLUpdateStore
 
@@ -283,12 +284,23 @@ def transform(self, transformer: Transformers) -> None:
                 )
             )
 
+    @property
+    def summary(self) -> pd.DataFrame:
+        return pd.DataFrame(self.queries.summarize_instances(), columns=["Type", "Occurrence"])
+
     def _repr_html_(self) -> str:
         provenance = self.provenance._repr_html_()
+        summary: pd.DataFrame = self.summary
+        summary_text = (
+            "<br /><strong>Graph is empty</strong><br />"
+            if summary.empty
+            else f"<br /><strong>Graph content</strong><br />{cast(pd.DataFrame, summary)._repr_html_()}"  # type: ignore[operator]
+        )
 
         return (
             f"<strong>{type(self).__name__}</strong> A graph store is a container for storing triples. "
             "It can be queried and transformed to extract information.<br />"
             "<strong>Provenance</strong> Provenance is a record of changes that have occurred in the graph store.<br />"
             f"{provenance}"
+            f"{summary_text}"
         )

cognite/neat/graph/transformers/_classic_cdf.py

@@ -1,6 +1,6 @@
 from typing import cast
 
-from rdflib import Graph, Literal, URIRef
+from rdflib import RDF, Graph, Literal, URIRef
 
 from cognite.neat.constants import DEFAULT_NAMESPACE
 from cognite.neat.graph import extractors
@@ -21,11 +21,16 @@ class AddAssetDepth(BaseTransformer):
                              <{asset_id}> <{root_prop}> ?root .}}"""
 
     def __init__(
-        self, asset_type: URIRef | None = None, root_prop: URIRef | None = None, parent_prop: URIRef | None = None
+        self,
+        asset_type: URIRef | None = None,
+        root_prop: URIRef | None = None,
+        parent_prop: URIRef | None = None,
+        depth_typing: dict[int, str] | None = None,
     ):
         self.asset_type = asset_type or DEFAULT_NAMESPACE.Asset
         self.root_prop = root_prop or DEFAULT_NAMESPACE.root
         self.parent_prop = parent_prop or DEFAULT_NAMESPACE.parent
+        self.depth_typing = depth_typing
 
     def transform(self, graph: Graph) -> None:
         """Adds depth of asset in the asset hierarchy to the graph."""
@@ -34,6 +39,13 @@ def transform(self, graph: Graph) -> None:
             if depth := self.get_depth(graph, asset_id, self.root_prop, self.parent_prop):
                 graph.add((asset_id, DEFAULT_NAMESPACE.depth, Literal(depth)))
 
+                if self.depth_typing and (type_ := self.depth_typing.get(depth, None)):
+                    # remove existing type
+                    graph.remove((asset_id, RDF.type, None))
+
+                    # add new type
+                    graph.add((asset_id, RDF.type, DEFAULT_NAMESPACE[type_]))
+
     @classmethod
     def get_depth(
         cls,
@@ -62,7 +74,12 @@ def get_depth(
 class AssetTimeSeriesConnector(BaseTransformer):
     description: str = "Connects assets to timeseries, thus forming bi-directional connection"
     _use_only_once: bool = True
-    _need_changes = frozenset({str(extractors.AssetsExtractor.__name__), str(extractors.TimeSeriesExtractor.__name__)})
+    _need_changes = frozenset(
+        {
+            str(extractors.AssetsExtractor.__name__),
+            str(extractors.TimeSeriesExtractor.__name__),
+        }
+    )
     _asset_template: str = """SELECT ?asset_id WHERE {{
                               <{timeseries_id}> <{asset_prop}> ?asset_id .
                               ?asset_id a <{asset_type}>}}"""
@@ -100,7 +117,12 @@ def transform(self, graph: Graph) -> None:
 class AssetSequenceConnector(BaseTransformer):
     description: str = "Connects assets to sequences, thus forming bi-directional connection"
     _use_only_once: bool = True
-    _need_changes = frozenset({str(extractors.AssetsExtractor.__name__), str(extractors.SequencesExtractor.__name__)})
+    _need_changes = frozenset(
+        {
+            str(extractors.AssetsExtractor.__name__),
+            str(extractors.SequencesExtractor.__name__),
+        }
+    )
     _asset_template: str = """SELECT ?asset_id WHERE {{
                               <{sequence_id}> <{asset_prop}> ?asset_id .
                               ?asset_id a <{asset_type}>}}"""
@@ -138,7 +160,12 @@ def transform(self, graph: Graph) -> None:
 class AssetFileConnector(BaseTransformer):
     description: str = "Connects assets to files, thus forming bi-directional connection"
     _use_only_once: bool = True
-    _need_changes = frozenset({str(extractors.AssetsExtractor.__name__), str(extractors.FilesExtractor.__name__)})
+    _need_changes = frozenset(
+        {
+            str(extractors.AssetsExtractor.__name__),
+            str(extractors.FilesExtractor.__name__),
+        }
+    )
     _asset_template: str = """SELECT ?asset_id WHERE {{
                               <{file_id}> <{asset_prop}> ?asset_id .
                               ?asset_id a <{asset_type}>}}"""
@@ -174,7 +201,12 @@ def transform(self, graph: Graph) -> None:
 class AssetEventConnector(BaseTransformer):
     description: str = "Connects assets to events, thus forming bi-directional connection"
     _use_only_once: bool = True
-    _need_changes = frozenset({str(extractors.AssetsExtractor.__name__), str(extractors.EventsExtractor.__name__)})
+    _need_changes = frozenset(
+        {
+            str(extractors.AssetsExtractor.__name__),
+            str(extractors.EventsExtractor.__name__),
+        }
+    )
     _asset_template: str = """SELECT ?asset_id WHERE {{
                               <{event_id}> <{asset_prop}> ?asset_id .
                               ?asset_id a <{asset_type}>}}"""
@@ -211,7 +243,10 @@ class AssetRelationshipConnector(BaseTransformer):
     description: str = "Connects assets via relationships"
     _use_only_once: bool = True
     _need_changes = frozenset(
-        {str(extractors.AssetsExtractor.__name__), str(extractors.RelationshipsExtractor.__name__)}
+        {
+            str(extractors.AssetsExtractor.__name__),
+            str(extractors.RelationshipsExtractor.__name__),
+        }
     )
     _asset_template: str = """SELECT ?source ?target WHERE {{
                               <{relationship_id}> <{relationship_source_xid_prop}> ?source_xid .
@@ -256,8 +291,20 @@ def transform(self, graph: Graph) -> None:
                 # files can be connected to multiple assets in the graph
                 for source_asset_id, target_asset_id in cast(list[tuple], assets_id_res):
                     # create a relationship between the two assets
-                    graph.add((source_asset_id, DEFAULT_NAMESPACE.relationship, relationship_id))
-                    graph.add((target_asset_id, DEFAULT_NAMESPACE.relationship, relationship_id))
+                    graph.add(
+                        (
+                            source_asset_id,
+                            DEFAULT_NAMESPACE.relationship,
+                            relationship_id,
+                        )
+                    )
+                    graph.add(
+                        (
+                            target_asset_id,
+                            DEFAULT_NAMESPACE.relationship,
+                            relationship_id,
+                        )
+                    )
 
                     # add source and target to the relationship
                     graph.add((relationship_id, DEFAULT_NAMESPACE.source, source_asset_id))

docs/CHANGELOG.md

@@ -15,6 +15,11 @@ Changes are grouped as follows:
 - `Fixed` for any bug fixes.
 - `Security` in case of vulnerabilities.
 
+## [0.85.10] - 10-07-24
+### Added
+- Depth-based typing in `AddAssetDepth` transformer
+- Graph summary repr
+
 ## [0.85.9] - 09-07-24
 ### Added
 - Option for checking for JSON value type when inferring data model

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "cognite-neat"
-version = "0.85.9"
+version = "0.85.10"
 readme = "README.md"
 description = "Knowledge graph transformation"
 authors = [

tests/tests_unit/graph/test_transformers/test_asset_depth_transformer.py

@@ -6,7 +6,7 @@
 from tests.config import CLASSIC_CDF_EXTRACTOR_DATA
 
 
-def test_asset_depth_transformer():
+def test_asset_depth_transformer_without_typing():
     store = NeatGraphStore.from_memory_store()
     extractor = extractors.AssetsExtractor.from_file(CLASSIC_CDF_EXTRACTOR_DATA / "assets.yaml")
     store.write(extractor)
@@ -24,6 +24,30 @@ def test_asset_depth_transformer():
     )
 
 
+def test_asset_depth_transformer_with_typing():
+    store = NeatGraphStore.from_memory_store()
+    extractor = extractors.AssetsExtractor.from_file(CLASSIC_CDF_EXTRACTOR_DATA / "assets.yaml")
+    store.write(extractor)
+
+    transformer = transformers.AddAssetDepth(
+        depth_typing={
+            1: "RootCimNode",
+            2: "GeographicalRegion",
+            3: "SubGeographicalRegion",
+            4: "Substation",
+        }
+    )
+
+    store.transform(transformer)
+
+    assert set(store.queries.summarize_instances()) == {
+        ("GeographicalRegion", 1),
+        ("RootCimNode", 1),
+        ("SubGeographicalRegion", 1),
+        ("Substation", 1),
+    }
+
+
 def test_asset_depth_transformer_warning():
     store = NeatGraphStore.from_memory_store()
 
@@ -38,5 +62,8 @@ def test_asset_depth_transformer_warning():
     store.write(extractor)
     store.transform(transformer)
 
-    with pytest.warns(UserWarning, match="Cannot transform graph store with AddAssetDepth, already applied"):
+    with pytest.warns(
+        UserWarning,
+        match="Cannot transform graph store with AddAssetDepth, already applied",
+    ):
         store.transform(transformer)