cognite-neat 0.123.38__tar.gz → 0.123.39__tar.gz

{cognite_neat-0.123.38 → cognite_neat-0.123.39}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cognite-neat
-Version: 0.123.38
+Version: 0.123.39
 Summary: Knowledge graph transformation
 Project-URL: Documentation, https://cognite-neat.readthedocs-hosted.com/
 Project-URL: Homepage, https://cognite-neat.readthedocs-hosted.com/

cognite_neat-0.123.39/cognite/neat/_version.py

@@ -0,0 +1,2 @@
+__version__ = "0.123.39"
+__engine__ = "^2.0.4"

{cognite_neat-0.123.38 → cognite_neat-0.123.39}/cognite/neat/v0/core/_data_model/models/data_types.py

@@ -302,7 +302,7 @@ class DateTime(DataType):
     python = datetime
     dms = dms.Timestamp
     graphql = "Timestamp"
-    xsd = "dateTimeStamp"
+    xsd = "dateTime"
     sql = "TIMESTAMP"
 
     name: typing.Literal["dateTime"] = "dateTime"

{cognite_neat-0.123.38 → cognite_neat-0.123.39}/cognite/neat/v0/core/_instances/extractors/_dict.py

@@ -43,9 +43,12 @@ class DictExtractor(BaseExtractor):
         self, key: str, value: Any, unpack_json: bool
     ) -> Iterable[tuple[str, Literal | URIRef]]:
         if key in self.uri_ref_keys and not isinstance(value, dict | list):
+            # exist if key is meant to form a URIRef
             yield key, URIRef(self.namespace[urllib.parse.quote(value)])
-        if isinstance(value, str | float | bool | int):
+        elif isinstance(value, float | bool | int):
             yield key, Literal(value)
+        elif isinstance(value, str):
+            yield key, Literal(string_to_ideal_type(value)) if self.str_to_ideal_type else Literal(value)
         elif isinstance(value, dict) and unpack_json:
             yield from self._unpack_json(value)
         elif isinstance(value, dict):

cognite_neat-0.123.39/last_changelog_entry.md

@@ -0,0 +1,7 @@
+
+### Improved
+
+- Improved RAW extractor to convert all string values (not just nested)
+to their ideal types and to generate only a single triple for
+connection-type keys, eliminating redundant string-typed triples that
+previously caused incorrect multi-type inferences.

cognite_neat-0.123.39/last_git_message.txt

@@ -0,0 +1,34 @@
+Improve raw extractor (#1302)
+
+# Description
+
+Previously RAW extractor was not converting string content of keys to
+ideal type, this was only done for nested values (if value was dict).
+Furthermore, if key (e.g., `asset`) was holding connection (`Asset001`),
+we were creating two triples:
+
+```
+ns:instance_001 ns:asset ns:Asset001 .
+ns:instance_001 ns:asset Asset001'^^xsd:string .
+```
+
+which downstream when we are doing data model inference we would get
+that property `asset` is of multi value type (`string`, `connection`)
+
+In this PR these issues are now fixed, i.e. conversion to the ideal type
+is done for any string value, while if key hold connection only triple
+representing it is created (and not one holding data)
+
+## Bump
+
+- [x] Patch
+- [ ] Minor
+- [ ] Skip
+
+## Changelog
+### Improved
+
+- Improved RAW extractor to convert all string values (not just nested)
+to their ideal types and to generate only a single triple for
+connection-type keys, eliminating redundant string-typed triples that
+previously caused incorrect multi-type inferences.

cognite_neat-0.123.39/last_version.txt

@@ -0,0 +1 @@
+0.123.38

{cognite_neat-0.123.38 → cognite_neat-0.123.39}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "cognite-neat"
-version = "0.123.38"
+version = "0.123.39"
 readme = "README.md"
 description = "Knowledge graph transformation"
 authors = [

cognite_neat-0.123.39/tests/v0/tests_unit/test_graph/test_extractors/test_raw_extractor.py

@@ -0,0 +1,37 @@
+from cognite.client.data_classes import Row
+from rdflib import RDF, XSD, Literal
+
+from cognite.neat.v0.core._client.testing import monkeypatch_neat_client
+from cognite.neat.v0.core._instances.extractors import RAWExtractor
+
+
+class TestRAWExtractor:
+    def test_extract_triples(self) -> None:
+        with monkeypatch_neat_client() as client:
+            client.raw.rows.return_value = [
+                Row(
+                    key="key1",
+                    columns={"column1": "value1", "column2": "100", "column3": "1983-01-22", "column4": "key2"},
+                ),
+                Row(
+                    key="key2",
+                    columns={"column1": "value3", "column2": "200", "column3": "1983-01-22", "column4": "key1"},
+                ),
+            ]
+        extractor = RAWExtractor(client, "my_db", "my_table", str_to_ideal_type=True, foreign_keys={"column4"})
+        ns = extractor.namespace
+
+        triples = set(extractor.extract())
+
+        assert triples == {
+            (ns["key1"], RDF.type, ns["my_table"]),
+            (ns["key1"], ns["column1"], Literal("value1")),
+            (ns["key1"], ns["column2"], Literal("100", datatype=XSD.integer)),
+            (ns["key1"], ns["column3"], Literal("1983-01-22T00:00:00", datatype=XSD.dateTime)),
+            (ns["key1"], ns["column4"], ns["key2"]),
+            (ns["key2"], RDF.type, ns["my_table"]),
+            (ns["key2"], ns["column1"], Literal("value3")),
+            (ns["key2"], ns["column2"], Literal("200", datatype=XSD.integer)),
+            (ns["key2"], ns["column3"], Literal("1983-01-22T00:00:00", datatype=XSD.dateTime)),
+            (ns["key2"], ns["column4"], ns["key1"]),
+        }

{cognite_neat-0.123.38 → cognite_neat-0.123.39}/uv.lock

@@ -346,7 +346,7 @@ wheels = [
 
 [[package]]
 name = "cognite-neat"
-version = "0.123.38"
+version = "0.123.39"
 source = { editable = "." }
 dependencies = [
     { name = "backports-strenum", marker = "python_full_version < '3.11'" },

cognite_neat-0.123.38/cognite/neat/_version.py

@@ -1,2 +0,0 @@
-__version__ = "0.123.38"
-__engine__ = "^2.0.4"

cognite_neat-0.123.38/last_changelog_entry.md

@@ -1,7 +0,0 @@
-
-### Improved
-
-- Node shape target class and property shape node are now set to
-instance source (if exists) or class uri
-- Property shape path is now set to instance source (if exists) or class
-uri when property is object property

cognite_neat-0.123.38/last_git_message.txt

@@ -1,30 +0,0 @@
-THIS 752: Correct target class and value type setting (#1304)
-
-# Description
-
-This is stacked PR on top of #1303 
-
-During the testing of SHACL engine, which was taking shacl rules
-generated by neat, it has beed detected that we were setting
-`NodeShape`'s `targetClass` and `PropertyShape`'s `node` to URI of SHACL
-shapes. The correct way is to set those properties to class URI, which
-in our case is stored as `instance_source` (if conceptual data model is
-linked to the instances) or to generate such URI from concept entity
-suffix and data model namespace. Furthermore, `PropertyShape`'s `path`
-was not leveraging `instance_source` of property (if it exists). The
-above has been implemented in this PR.
-
-
-## Bump
-
-- [x] Patch
-- [ ] Minor
-- [ ] Skip
-
-## Changelog
-### Improved
-
-- Node shape target class and property shape node are now set to
-instance source (if exists) or class uri
-- Property shape path is now set to instance source (if exists) or class
-uri when property is object property

cognite_neat-0.123.38/last_version.txt

@@ -1 +0,0 @@
-0.123.37

cognite_neat-0.123.38/tests/v0/tests_unit/test_graph/test_extractors/test_raw_extractor.py

@@ -1,27 +0,0 @@
-from cognite.client.data_classes import Row
-from rdflib import RDF, Literal
-
-from cognite.neat.v0.core._client.testing import monkeypatch_neat_client
-from cognite.neat.v0.core._instances.extractors import RAWExtractor
-
-
-class TestRAWExtractor:
-    def test_extract_triples(self) -> None:
-        with monkeypatch_neat_client() as client:
-            client.raw.rows.return_value = [
-                Row(key="key1", columns={"column1": "value1", "column2": "value2"}),
-                Row(key="key2", columns={"column1": "value3", "column2": "value4"}),
-            ]
-        extractor = RAWExtractor(client, "my_db", "my_table")
-        ns = extractor.namespace
-
-        triples = set(extractor.extract())
-
-        assert triples == {
-            (ns["key1"], RDF.type, ns["my_table"]),
-            (ns["key1"], ns["column1"], Literal("value1")),
-            (ns["key1"], ns["column2"], Literal("value2")),
-            (ns["key2"], RDF.type, ns["my_table"]),
-            (ns["key2"], ns["column1"], Literal("value3")),
-            (ns["key2"], ns["column2"], Literal("value4")),
-        }