review / fix issues with updated inputs_v2

Author: jonsch1

docs/inputs-v2-direct-port-strategy.md

@@ -0,0 +1,58 @@
+"""Source-neutral parsing of finite numeric observations and comparison bounds."""
+
+from __future__ import annotations
+
+import math
+import re
+
+from biolink_model.datamodel.model import QuantityValue
+from omnipath_core.measurements import Measurement
+
+_NUMBER = re.compile(
+    r'\s*(<=|>=|<|>|=|~|≈)?\s*'
+    r'([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?)\s*'
+)
+_COMPARATORS = {None, '<', '>', '=', '<=', '>=', '~', '≈'}
+
+
+def measurement(
+    value: object,
+    unit: str | None = None,
+    source_field: str | None = None,
+    comparator: str | None = None,
+) -> Measurement | None:
+    """Parse a scalar without inferring its endpoint, units or biological meaning.
+
+    Invalid, non-finite and contradictory values omit only the measurement.
+    Preserve inclusive/approximate bounds without replacing them by strict ones.
+    """
+    if value is None:
+        return None
+    match = _NUMBER.fullmatch(str(value))
+    if match is None:
+        return None
+    explicit = (
+        (str(comparator).strip() or None) if comparator is not None else None
+    )
+    embedded = match.group(1)
+    if explicit not in _COMPARATORS:
+        return None
+    if explicit and embedded and explicit != embedded:
+        return None
+    comparison = explicit or embedded
+    number = float(match.group(2))
+    if not math.isfinite(number):
+        return None
+    return Measurement(
+        quantity=QuantityValue(
+            has_numeric_value=number,
+            has_unit=unit or None,
+            has_binary_relation={
+                '<': 'less_than',
+                '>': 'greater_than',
+                '=': 'equal_to',
+            }.get(comparison),
+        ),
+        source_field=source_field,
+        comparator=comparison,
+    )

docs/inputs-v2-legacy-adapter-inventory.md

@@ -238,6 +238,19 @@ def __call__(
             yield self.mapper(record)
 
 
+def _ontology_identifier_namespace(identifier, default):
+    """Keep imported CURIE namespaces; a document can contain foreign terms."""
+    prefix, separator, _ = str(identifier).partition(':')
+    if not separator:
+        return default
+    return {
+        'GO': Namespace.GO, 'HP': Namespace.HPO, 'MONDO': Namespace.MONDO,
+        'CHEMONTID': Namespace.CHEMONT, 'MI': Namespace.MI,
+        'EC': Namespace.EC, 'OM': Namespace.OM,
+        'UniProtKB-KW': Namespace.UNIPROT_KEYWORD,
+    }.get(prefix, prefix.lower())
+
+
 def ontology_term_to_entity(
     term: OntologyTerm,
     *,
@@ -254,7 +267,7 @@ def ontology_term_to_entity(
     if not term.id or term.is_obsolete:
         return None
     identifiers = [
-        Identifier(type=identifier_type, value=value)
+        Identifier(type=_ontology_identifier_namespace(value, identifier_type), value=value)
         for value in dict.fromkeys([term.id, *(term.alt_ids or [])])
         if value
     ]
@@ -293,7 +306,7 @@ def ontology_term_to_entity(
         relations.append(
             OntologyRelation(
                 predicate=predicate,
-                object=EntityRef(entity_type, identifier_type, target),
+                object=EntityRef(entity_type, _ontology_identifier_namespace(target, identifier_type), target),
                 ontology_id=ontology_id,
             )
         )

docs/inputs-v2-legacy-adapter-mapping.md

@@ -7,19 +7,15 @@
 
 from __future__ import annotations
 
-import math
-import re
-
 from biolink_model.datamodel.model import (
     ChemicalEntity,
     MacromolecularComplex,
     Protein,
-    QuantityValue,
     slots,
 )
-from omnipath_core.measurements import Measurement
 from omnipath_core.naming import Namespace
 
+from pypath.inputs_v2._measurements import measurement as _measurement
 from pypath.inputs_v2.base import Dataset, Download, Resource, ResourceConfig
 from pypath.inputs_v2.parsers.bindingdb import _raw
 from pypath.internals.cv_terms import LicenseCV, ResourceCv, UpdateCategoryCV
@@ -38,37 +34,6 @@
 )
 
 
-def _measurement(value, unit=None, source_field=None, comparator=None):
-    """Keep numeric source observations, units and comparison bounds together."""
-    if value is None:
-        return None
-    match = re.fullmatch(
-        '\\s*(<=|>=|<|>|=|~)?\\s*([+-]?(?:\\d+(?:\\.\\d*)?|\\.\\d+)(?:[eE][+-]?\\d+)?)\\s*',
-        str(value),
-    )
-    if match is None:
-        return None
-    original_comparator = comparator or match.group(1)
-    if original_comparator not in {None, '<', '>', '=', '<=', '>=', '~', '≈'}:
-        return None
-    relation = {'<': 'less_than', '>': 'greater_than', '=': 'equal_to'}.get(
-        original_comparator
-    )
-    number = float(match.group(2))
-    if not math.isfinite(number):
-        return None
-    quantity = QuantityValue(
-        has_numeric_value=number,
-        has_unit=unit or None,
-        has_binary_relation=relation,
-    )
-    return Measurement(
-        quantity=quantity,
-        source_field=source_field,
-        comparator=original_comparator,
-    )
-
-
 def _bindingdb_url(dataset: str = 'All', **_kwargs: object) -> str:
     return f'https://bindingdb.org/rwd/bind/downloads/BindingDB_{dataset}_202605_tsv.zip'
 
@@ -214,7 +179,7 @@ def _target(row):
             Identifier(type=Namespace.NAME, value=row.get('Target Name'))
         ],
         membership=members,
-        annotations=[],
+        annotations=target_builder.annotations.build(row),
     )
 
 
@@ -300,7 +265,9 @@ def _target(row):
             ),
         ),
     ),
-    identifiers=IdentifiersBuilder(),
+    identifiers=IdentifiersBuilder(
+        CV(term=Namespace.BINDINGDB, value=f('BindingDB Reactant_set_id')),
+    ),
 )
 resource = Resource(
     config,

pypath/inputs_v2/_measurements.py

@@ -226,7 +226,9 @@ def _type_selector(row: dict[str, Any]) -> type:
             ),
         )
     ),
-    annotations=AnnotationsBuilder(),
+    annotations=AnnotationsBuilder(
+        CV(term=slots.in_taxon, value=f'NCBITaxon:{HUMAN_TAXON_ID}'),
+    ),
 )
 
 

pypath/inputs_v2/base.py

@@ -8,7 +8,6 @@
 from __future__ import annotations
 
 from functools import partial
-import math
 from pathlib import Path
 import re
 
@@ -19,21 +18,20 @@
     ChemicalEntity,
     DirectionQualifierEnum,
     Gene,
-    GeneFamily,
+    ProteinFamily,
     MacromolecularComplex,
     MolecularActivity,
     MolecularEntity,
     NamedThing,
     NucleicAcidEntity,
     OrganismTaxon,
     Protein,
-    QuantityValue,
     RNAProduct,
     slots,
 )
-from omnipath_core.measurements import Measurement
 from omnipath_core.naming import Namespace
 
+from pypath.inputs_v2._measurements import measurement as _measurement
 from pypath.inputs_v2.base import Dataset, Download, Resource, ResourceConfig
 from pypath.inputs_v2.parsers.chembl import (
     activities_parser,
@@ -54,37 +52,6 @@
 from pypath.share import cache
 
 
-def _measurement(value, unit=None, source_field=None, comparator=None):
-    """Keep numeric source observations, units and comparison bounds together."""
-    if value is None:
-        return None
-    match = re.fullmatch(
-        '\\s*(<=|>=|<|>|=|~)?\\s*([+-]?(?:\\d+(?:\\.\\d*)?|\\.\\d+)(?:[eE][+-]?\\d+)?)\\s*',
-        str(value),
-    )
-    if match is None:
-        return None
-    original_comparator = comparator or match.group(1)
-    if original_comparator not in {None, '<', '>', '=', '<=', '>=', '~', '≈'}:
-        return None
-    relation = {'<': 'less_than', '>': 'greater_than', '=': 'equal_to'}.get(
-        original_comparator
-    )
-    number = float(match.group(2))
-    if not math.isfinite(number):
-        return None
-    quantity = QuantityValue(
-        has_numeric_value=number,
-        has_unit=unit or None,
-        has_binary_relation=relation,
-    )
-    return Measurement(
-        quantity=quantity,
-        source_field=source_field,
-        comparator=original_comparator,
-    )
-
-
 VERSION = 36
 DB_REL_PATH = f'chembl_{VERSION}/chembl_{VERSION}_sqlite/chembl_{VERSION}.db'
 SQLITE_PATH = Path(cache.get_cachedir()) / f'ChEMBL_SQLite_{VERSION}.sqlite'
@@ -133,7 +100,7 @@ def _files_needed(version: int = VERSION, **_kwargs: object) -> list[str]:
 TARGET_TYPE_MAP = {
     'SINGLE PROTEIN': Protein,
     'PROTEIN COMPLEX': MacromolecularComplex,
-    'PROTEIN FAMILY': GeneFamily,
+    'PROTEIN FAMILY': ProteinFamily,
     'PROTEIN-PROTEIN INTERACTION': MolecularActivity,
     'SELECTIVITY GROUP': NamedThing,
     'NUCLEIC-ACID': NucleicAcidEntity,
@@ -333,6 +300,8 @@ def chembl_predicate(row):
     predicate=chembl_predicate,
     object=target_builder,
     annotations=AnnotationsBuilder(
+        CV(term=slots.source_record_urls, value=f('assay_chembl_id', transform=lambda v: f'https://www.ebi.ac.uk/chembl/explore/assay/{v}')),
+        CV(term=slots.source_record_urls, value=f('document_chembl_id', transform=lambda v: f'https://www.ebi.ac.uk/chembl/explore/document/{v}')),
         CV(
             term=slots.has_quantitative_value,
             value=f(
@@ -388,7 +357,10 @@ def chembl_predicate(row):
         ),
         CV(term=slots.chembl_confidence_score, value=f('confidence_score')),
     ),
-    identifiers=IdentifiersBuilder(),
+    identifiers=IdentifiersBuilder(
+        CV(term=Namespace.CHEMBL_ACTIVITY, value=f('activity_id')),
+        CV(term=Namespace.CHEMBL_MECHANISM, value=f('mec_id')),
+    ),
 )
 resource = Resource(
     config=config,

pypath/inputs_v2/bindingdb.py

@@ -3,20 +3,17 @@
 from __future__ import annotations
 
 import csv
-import math
-import re
 
 from biolink_model.datamodel.model import (
     ChemicalEntity,
     DirectionQualifierEnum,
     NamedThing,
     Protein,
-    QuantityValue,
     slots,
 )
-from omnipath_core.measurements import Measurement
 from omnipath_core.naming import Namespace
 
+from pypath.inputs_v2._measurements import measurement as _measurement
 from pypath.inputs_v2.base import Dataset, Download, Resource, ResourceConfig
 from pypath.inputs_v2.parsers.base import iter_tsv
 from pypath.internals.cv_terms import LicenseCV, ResourceCv, UpdateCategoryCV
@@ -35,38 +32,6 @@
 )
 from pypath.share.downloads import download_and_open
 
-
-def _measurement(value, unit=None, source_field=None, comparator=None):
-    """Keep numeric source observations, units and comparison bounds together."""
-    if value is None:
-        return None
-    match = re.fullmatch(
-        '\\s*(<=|>=|<|>|=|~)?\\s*([+-]?(?:\\d+(?:\\.\\d*)?|\\.\\d+)(?:[eE][+-]?\\d+)?)\\s*',
-        str(value),
-    )
-    if match is None:
-        return None
-    original_comparator = comparator or match.group(1)
-    if original_comparator not in {None, '<', '>', '=', '<=', '>=', '~', '≈'}:
-        return None
-    relation = {'<': 'less_than', '>': 'greater_than', '=': 'equal_to'}.get(
-        original_comparator
-    )
-    number = float(match.group(2))
-    if not math.isfinite(number):
-        return None
-    quantity = QuantityValue(
-        has_numeric_value=number,
-        has_unit=unit or None,
-        has_binary_relation=relation,
-    )
-    return Measurement(
-        quantity=quantity,
-        source_field=source_field,
-        comparator=original_comparator,
-    )
-
-
 DRUGCENTRAL_STRUCTURES_URL = 'https://unmtid-shinyapps.net/download/DrugCentral/2021_09_01/structures.smiles.tsv'
 config = ResourceConfig(
     id=ResourceCv.DRUGCENTRAL,
@@ -101,7 +66,8 @@ def _clean(value):
 
 
 def _values(value):
-    return [v.strip() for v in _clean(value).split('|') if v.strip()]
+    # These lists are positionally aligned; dropping blanks changes identity.
+    return [v.strip() for v in _clean(value).split('|')]
 
 
 def _taxon(row):
@@ -165,7 +131,7 @@ def _target_entity(row):
                 type=Namespace.NAME, value=_clean(row.get('TARGET_NAME'))
             )
         ],
-        annotations=[],
+        annotations=protein_builder.annotations.build(row),
         membership=[
             Membership(member=p, predicate=slots.has_member) for p in proteins
         ],
@@ -204,6 +170,10 @@ def _predicate(row):
             ),
         ),
         CV(term=slots.description, value=f('ACT_COMMENT')),
+        CV(term=slots.supporting_data_source, value=f('ACT_SOURCE')),
+        CV(term=slots.supporting_data_source, value=f('MOA_SOURCE')),
+        CV(term=slots.source_record_urls, value=f('ACT_SOURCE_URL')),
+        CV(term=slots.source_record_urls, value=f('MOA_SOURCE_URL')),
     ),
     identifiers=IdentifiersBuilder(),
 )