Merge branch 'main' into obo-compliance-suite

.github/workflows/main.yaml

@@ -80,7 +80,9 @@ jobs:
       #----------------------------------------------
       - name: Install dependencies
         # if: steps.cached-poetry-dependencies.outputs.cache-hit != 'true'
-        run: poetry install --no-interaction --no-root
+        run: |
+          poetry add setuptools@latest
+          poetry install --no-interaction --no-root
 
       #----------------------------------------------
       #    install your root project, if required

README.md

@@ -11,7 +11,7 @@ Python lib for common ontology operations over a variety of backends.
 [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.6456239.svg)](https://doi.org/10.5281/zenodo.6456239)
 [![Contributor Covenant](https://img.shields.io/badge/Contributor%20Covenant-2.1-4baaaa.svg)](.github/CODE_OF_CONDUCT.md) 
 
-OAK provides a collection of [interfaces](https://incatools.github.io/ontology-access-kit/interfaces/index.html) for various ontology operations, including:
+OAK provides a collection of [interfaces](https://incatools.github.io/ontology-access-kit/packages/interfaces/index.html#interfaces) for various ontology operations, including:
 
  - [look up basic features](https://incatools.github.io/ontology-access-kit/guide/basics.html) of an ontology element, such as its label, definition, relationships, or aliases
  - search an ontology for a term
@@ -84,7 +84,7 @@ For more examples, see
 See:
 
  - [CLI docs](https://incatools.github.io/ontology-access-kit/cli.html)
- - [Example notebooks](https://github.com/INCATools/ontology-access-kit/tree/main/notebooks/Commands
+ - [Example notebooks](https://github.com/INCATools/ontology-access-kit/tree/main/notebooks/Commands)
 
 ## Search
 

notebooks/Commands/input/eds-genes-hgnc.tsv

@@ -0,0 +1,20 @@
+id	label
+HGNC:11976	TNXB
+HGNC:1246	C1R
+HGNC:1247	C1S
+HGNC:17978	B3GALT6
+HGNC:18625	FKBP14
+HGNC:20859	SLC39A13
+HGNC:21144	DSE
+HGNC:2188	COL12A1
+HGNC:218	ADAMTS2
+HGNC:2198	COL1A2
+HGNC:2201	COL3A1
+HGNC:2209	COL5A1
+HGNC:2210	COL5A2
+HGNC:23216	ZNF469
+HGNC:24464	CHST14
+HGNC:303	AEBP1
+HGNC:9081	PLOD1
+HGNC:930	B4GALT7
+HGNC:9349	PRDM5

pyproject.toml

@@ -15,13 +15,11 @@ SQLAlchemy = ">=1.4.32"
 linkml-runtime = ">=1.5.3"
 linkml-renderer = ">=0.3.0"
 networkx = ">=2.7.1"
-sssom-schema = ">=0.11.0"
-sssom = ">=0.3.38"
+sssom = "^0.4.0"
 ratelimit = ">=2.2.1"
 appdirs = ">=1.4.4"
 semsql = ">=0.3.1"
-lark = ">=1.1.2"
-kgcl-schema = "0.6.0"
+kgcl-schema = ">=0.6.0"
 funowl = ">=0.2.0"
 gilda = {version = ">=1.0.0", optional = true}
 kgcl-rdflib = "0.5.0"
@@ -37,7 +35,7 @@ pysolr = "^3.9.0"
 eutils = ">=0.6.0"
 requests-cache = "^1.0.1"
 click = "*"
-semsimian = "0.2.1"
+semsimian = "^0.2.1"
 urllib3 = {version = "< 2", optional = true}
 pydantic = "*"
 

src/oaklib/cli.py

@@ -1006,8 +1006,9 @@ def main(
 @main.command()
 @click.argument("terms", nargs=-1)
 @ontological_output_type_option
+@autolabel_option
 @output_option
-def search(terms, output_type: str, output: TextIO):
+def search(terms, output_type: str, autolabel, output: TextIO):
     """
     Searches ontology for entities that have a label, alias, or other property matching a search term.
 
@@ -1056,11 +1057,15 @@ def search(terms, output_type: str, output: TextIO):
     if isinstance(impl, SearchInterface):
         writer = _get_writer(output_type, impl, StreamingInfoWriter)
         writer.output = output
-        for curie_it in chunk(query_terms_iterator(terms, impl)):
-            logging.info("** Next chunk:")
-            # TODO: move chunking logic to writer
-            for curie, label in impl.labels(curie_it):
-                writer.emit(dict(id=curie, label=label))
+        if autolabel:
+            for curie_it in chunk(query_terms_iterator(terms, impl)):
+                logging.info("** Next chunk:")
+                # TODO: move chunking logic to writer
+                for curie, label in impl.labels(curie_it):
+                    writer.emit(dict(id=curie, label=label))
+        else:
+            for curie in query_terms_iterator(terms, impl):
+                writer.emit(dict(id=curie), label_fields=[])
         writer.finish()
     else:
         raise NotImplementedError(f"Cannot execute this using {impl} of type {type(impl)}")

src/oaklib/datamodels/association.py

@@ -1,5 +1,5 @@
-# Auto generated from association.yaml by pythongen.py version: 0.9.0
-# Generation date: 2023-08-14T15:53:04
+# Auto generated from association.yaml by pythongen.py version: 0.0.1
+# Generation date: 2023-11-17T17:56:03
 # Schema: association
 #
 # id: https://w3id.org/oak/association
@@ -69,6 +69,7 @@
 OA = CurieNamespace("oa", "http://www.w3.org/ns/oa#")
 ONTOASSOC = CurieNamespace("ontoassoc", "https://w3id.org/oak/association/")
 RDF = CurieNamespace("rdf", "http://example.org/UNKNOWN/rdf/")
+RDFS = CurieNamespace("rdfs", "http://example.org/UNKNOWN/rdfs/")
 SSSOM = CurieNamespace("sssom", "https://w3id.org/sssom/")
 DEFAULT_ = ONTOASSOC
 
@@ -100,6 +101,10 @@ class PositiveOrNegativeAssociation(YAMLRoot):
     publications: Optional[
         Union[Union[str, URIorCURIE], List[Union[str, URIorCURIE]]]
     ] = empty_list()
+    evidence_type: Optional[Union[str, URIorCURIE]] = None
+    supporting_objects: Optional[
+        Union[Union[str, URIorCURIE], List[Union[str, URIorCURIE]]]
+    ] = empty_list()
     primary_knowledge_source: Optional[Union[str, URIorCURIE]] = None
     aggregator_knowledge_source: Optional[Union[str, URIorCURIE]] = None
     subject_closure: Optional[
@@ -110,6 +115,7 @@ class PositiveOrNegativeAssociation(YAMLRoot):
         Union[Union[str, URIorCURIE], List[Union[str, URIorCURIE]]]
     ] = empty_list()
     object_closure_label: Optional[Union[str, List[str]]] = empty_list()
+    comments: Optional[Union[str, List[str]]] = empty_list()
 
     def __post_init__(self, *_: List[str], **kwargs: Dict[str, Any]):
         if self.subject is not None and not isinstance(self.subject, URIorCURIE):
@@ -148,6 +154,17 @@ def __post_init__(self, *_: List[str], **kwargs: Dict[str, Any]):
             v if isinstance(v, URIorCURIE) else URIorCURIE(v) for v in self.publications
         ]
 
+        if self.evidence_type is not None and not isinstance(self.evidence_type, URIorCURIE):
+            self.evidence_type = URIorCURIE(self.evidence_type)
+
+        if not isinstance(self.supporting_objects, list):
+            self.supporting_objects = (
+                [self.supporting_objects] if self.supporting_objects is not None else []
+            )
+        self.supporting_objects = [
+            v if isinstance(v, URIorCURIE) else URIorCURIE(v) for v in self.supporting_objects
+        ]
+
         if self.primary_knowledge_source is not None and not isinstance(
             self.primary_knowledge_source, URIorCURIE
         ):
@@ -188,6 +205,10 @@ def __post_init__(self, *_: List[str], **kwargs: Dict[str, Any]):
             v if isinstance(v, str) else str(v) for v in self.object_closure_label
         ]
 
+        if not isinstance(self.comments, list):
+            self.comments = [self.comments] if self.comments is not None else []
+        self.comments = [v if isinstance(v, str) else str(v) for v in self.comments]
+
         super().__post_init__(**kwargs)
 
 
@@ -711,6 +732,33 @@ class slots:
     range=Optional[Union[str, URIorCURIE]],
 )
 
+slots.evidence_type = Slot(
+    uri=ONTOASSOC.evidence_type,
+    name="evidence_type",
+    curie=ONTOASSOC.curie("evidence_type"),
+    model_uri=ONTOASSOC.evidence_type,
+    domain=None,
+    range=Optional[Union[str, URIorCURIE]],
+)
+
+slots.supporting_objects = Slot(
+    uri=ONTOASSOC.supporting_objects,
+    name="supporting_objects",
+    curie=ONTOASSOC.curie("supporting_objects"),
+    model_uri=ONTOASSOC.supporting_objects,
+    domain=None,
+    range=Optional[Union[Union[str, URIorCURIE], List[Union[str, URIorCURIE]]]],
+)
+
+slots.comments = Slot(
+    uri=RDFS.comment,
+    name="comments",
+    curie=RDFS.curie("comment"),
+    model_uri=ONTOASSOC.comments,
+    domain=None,
+    range=Optional[Union[str, List[str]]],
+)
+
 slots.denormalized_slot = Slot(
     uri=ONTOASSOC.denormalized_slot,
     name="denormalized_slot",

src/oaklib/datamodels/association.yaml

@@ -41,12 +41,15 @@ classes:
       - object_label
       - negated
       - publications
+      - evidence_type
+      - supporting_objects
       - primary_knowledge_source
       - aggregator_knowledge_source
       - subject_closure
       - subject_closure_label
       - object_closure
       - object_closure_label
+      - comments
 
   Association:
     is_a: PositiveOrNegativeAssociation
@@ -242,6 +245,18 @@ slots:
     description: The knowledge source that aggregated the association
     slot_uri: biolink:aggregator_knowledge_source
     range: uriorcurie
+  evidence_type:
+    description: The type of evidence supporting the association
+    range: uriorcurie
+  supporting_objects:
+    description: The objects that support the association
+    range: uriorcurie
+    multivalued: true
+  comments:
+    description: Comments about the association
+    slot_uri: rdfs:comment
+    range: string
+    multivalued: true
   denormalized_slot:
     mixin: true
     description: |-
@@ -387,3 +402,4 @@ slots:
   associations_for_subjects_in_common:
     multivalued: true
     range: Association
+

src/oaklib/implementations/__init__.py

@@ -17,6 +17,7 @@
 from oaklib.implementations.kgx.kgx_implementation import KGXImplementation
 from oaklib.implementations.llm_implementation import LLMImplementation
 from oaklib.implementations.monarch.monarch_implementation import MonarchImplementation
+from oaklib.implementations.ncbi.ncbi_gene_implementation import NCBIGeneImplementation
 from oaklib.implementations.ols import (
     BaseOlsImplementation,
     OlsImplementation,
@@ -77,6 +78,7 @@
     "OlsImplementation",
     "TIBOlsImplementation",
     "MonarchImplementation",
+    "NCBIGeneImplementation",
     "OntobeeImplementation",
     "ProntoImplementation",
     "SimpleOboImplementation",