Merge pull request #77 from INCATools/additional-sqlite-tests

additional sqlite tests

src/oaklib/implementations/sqldb/sql_implementation.py

@@ -1,7 +1,7 @@
 import logging
 from abc import ABC
 from collections import defaultdict
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from typing import List, Any, Iterable, Optional, Type, Dict, Union, Tuple, Iterator
 
 import semsql.builder.builder as semsql_builder
@@ -70,7 +70,6 @@ class SqlImplementation(RelationGraphInterface, OboGraphInterface, ValidatorInte
     _session: Any = None
     _connection: Any = None
     _ontology_metadata_model: SchemaView = None
-    autosave : bool = None
 
     def __post_init__(self):
         if self.engine is None:
@@ -169,9 +168,11 @@ def curies_by_subset(self, subset: SUBSET_CURIE) -> Iterable[CURIE]:
             yield self._get_subset_curie(row.subject)
 
     def set_label_for_curie(self, curie: CURIE, label: str) -> bool:
+        stmt = update(Statements).where(Statements.subject == curie and Statements.predicate == LABEL_PREDICATE).values(value=label)
         self.session.execute(
-            update(Statements).where(Statements.subject == curie and Statements.predicate == LABEL_PREDICATE).values(value=label)
+            stmt
         )
+        self.session.flush()
         if self.autosave:
             self.save()
 
@@ -521,7 +522,8 @@ def gap_fill_relationships(self, seed_curies: List[CURIE], predicates: List[PRED
     # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
     def get_information_content(self, curie: CURIE, background: CURIE = None,
                                 predicates: List[PRED_CURIE] = None):
-        raise NotImplementedError
+        # TODO: fix
+        return 1.0
 
     # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
     # Implements: PatcherInterface
@@ -540,5 +542,6 @@ def migrate_curies(self, curie_map: Dict[CURIE, CURIE]) -> None:
             self.save()
 
     def save(self):
+        logging.info(f'Committing and flushing changes')
         self.session.commit()
         self.session.flush()

src/oaklib/interfaces/basic_ontology_interface.py

@@ -1,5 +1,5 @@
 from abc import ABC
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from typing import Dict, List, Iterable, Tuple, Optional, Any, Iterator
 
 from oaklib.interfaces.ontology_interface import OntologyInterface
@@ -66,6 +66,7 @@ class BasicOntologyInterface(OntologyInterface, ABC):
 
     """
     strict: bool = False
+    autosave: bool = field(default_factory= lambda: True)
 
 
     def get_prefix_map(self) -> PREFIX_MAP:
@@ -335,3 +336,5 @@ def create_entity(self, curie: CURIE, label: str = None, relationships: RELATION
         """
         raise NotImplementedError
 
+    def save(self):
+        pass

src/oaklib/interfaces/semsim_interface.py

@@ -6,6 +6,7 @@
 from oaklib.interfaces.basic_ontology_interface import BasicOntologyInterface
 from oaklib.interfaces.obograph_interface import OboGraphInterface
 from oaklib.types import CURIE, LABEL, URI, PRED_CURIE
+from oaklib.utilities.semsim.similarity_utils import setwise_jaccard_similarity
 
 
 class SemanticSimilarityInterface(BasicOntologyInterface, ABC):
@@ -52,6 +53,9 @@ def pairwise_similarity(self, subject: CURIE, object: CURIE = None,
         best_mrcas = [a for a in cas if ics[a] == max_ic]
         sim = TermPairwiseSimilarity(subject_id=subject, object_id=object, ancestor_id=best_mrcas[0])
         sim.ancestor_information_content = max_ic
-        sim.phenodigm_score = math.sqrt(sim.jaccard_similarity * sim.information_content)
+        if isinstance(self, OboGraphInterface):
+            sim.jaccard_similarity = setwise_jaccard_similarity(list(self.ancestors(subject, predicates=predicates)),
+                                                                list(self.ancestors(object, predicates=predicates)))
+        #sim.phenodigm_score = math.sqrt(sim.jaccard_similarity * sim.information_content)
         return sim
 

src/oaklib/utilities/lexical/lexical_indexer.py

@@ -34,16 +34,22 @@ def add_labels_from_uris(oi: BasicOntologyInterface):
     :return:
     """
     logging.info(f'Adding labels from URIs')
-    for curie in oi.all_entity_curies():
+    curies = list(oi.all_entity_curies())
+    for curie in curies:
         if not oi.get_label_by_curie(curie):
             if '#' in curie:
                 sep = '#'
-            elif curie.startswith('http'):
+            elif curie.startswith('http') or curie.startswith('<http'):
                 sep = '/'
             else:
                 sep = ':'
             label = curie.split(sep)[-1]
+            if curie.startswith('<http'):
+                label = label.replace('>', '')
+            label = " ".join(label.split('_'))
+            #print(f'{curie} ==> {label} // {type(oi)}')
             oi.set_label_for_curie(curie, label)
+            #print(oi.get_label_by_curie(curie))
 
 
 def create_lexical_index(oi: BasicOntologyInterface,

src/oaklib/utilities/reasoning/relation_graph.py

@@ -0,0 +1,16 @@
+from collections import Iterator
+from dataclasses import dataclass
+import oaklib.datamodels.obograph as og
+from oaklib.utilities.obograph_utils import index_graph_edges_by_subject, index_graph_edges_by_subject_object
+
+
+@dataclass
+class RelationGraphReasoner:
+    ontology: og.Graph = None
+
+    def get_non_redundant_edges(self) -> Iterator[og.Edge]:
+        ont = self.ontology
+        six = index_graph_edges_by_subject(ont.edges)
+        soix = index_graph_edges_by_subject_object(ont.edges)
+        for e in ont.edges:
+            yield e