Add NLM Catalog and fix default prefix on object properties and annot…

…ation properties (#263)

src/pyobo/sources/__init__.py

@@ -39,6 +39,7 @@
 from .mirbase_mature import MiRBaseMatureGetter
 from .msigdb import MSigDBGetter
 from .ncbigene import NCBIGeneGetter
+from .nlm_catalog import NLMCatalogGetter
 from .npass import NPASSGetter
 from .omim_ps import OMIMPSGetter
 from .pathbank import PathBankGetter
@@ -101,6 +102,7 @@
     "MiRBaseGetter",
     "MiRBaseMatureGetter",
     "NCBIGeneGetter",
+    "NLMCatalogGetter",
     "NPASSGetter",
     "OMIMPSGetter",
     "PIDGetter",

src/pyobo/sources/nlm_catalog.py

@@ -0,0 +1,82 @@
+"""Converter for NLM Providers."""
+
+from collections.abc import Iterable
+from xml.etree import ElementTree
+
+from pyobo.struct import Obo, Reference, Term, TypeDef, default_reference
+from pyobo.utils.path import ensure_df, ensure_path
+
+__all__ = [
+    "NLMCatalogGetter",
+]
+
+PREFIX = "nlm"
+CATALOG_TO_PUBLISHER = "https://ftp.ncbi.nlm.nih.gov/pubmed/xmlprovidernames.txt"
+JOURNAL_INFO_PATH = "https://ftp.ncbi.nlm.nih.gov/pubmed/jourcache.xml"
+PUBLISHER = TypeDef.default(PREFIX, "has_publisher", name="has publisher")
+START_YEAR = TypeDef.default(PREFIX, "has_start_year", name="has start year")
+END_YEAR = TypeDef.default(PREFIX, "has_end_year", name="has end year")
+
+
+# TODO enrich with context from https://ftp.ncbi.nlm.nih.gov/pubmed/J_Entrez.txt and https://ftp.ncbi.nlm.nih.gov/pubmed/J_Medline.txt
+
+
+class NLMCatalogGetter(Obo):
+    """An ontology representation of NLM Providers."""
+
+    bioversions_key = ontology = PREFIX
+    dynamic_version = True
+    typedefs = [PUBLISHER, START_YEAR, END_YEAR]
+    idspaces = {
+        PREFIX: "https://www.ncbi.nlm.nih.gov/nlmcatalog/",
+    }
+
+    def iter_terms(self, force: bool = False) -> Iterable[Term]:
+        """Iterate over gene terms for NLM Catalog."""
+        yield from get_terms()
+
+
+def get_terms(force: bool = False) -> Iterable[Term]:
+    """Get NLM Catalog terms."""
+    path = ensure_path(PREFIX, url=JOURNAL_INFO_PATH)
+    root = ElementTree.parse(path).getroot()
+
+    journal_to_publisher_df = ensure_df(
+        PREFIX, url=CATALOG_TO_PUBLISHER, sep="|", force=force, dtype=str
+    )
+    journal_id_to_publisher_key: dict[str, Reference] = {
+        # TODO change to external prefix later
+        journal_id: default_reference(PREFIX, key, name)
+        for journal_id, key, name in journal_to_publisher_df.values
+    }
+    for element in root.findall("Journal"):
+        yield _process_journal(element, journal_id_to_publisher_key)
+    for k in sorted(set(journal_id_to_publisher_key.values())):
+        yield Term(reference=k)
+
+
+def _process_journal(element, journal_id_to_publisher_key: dict[str, Reference]) -> Term:
+    nlm_id = element.findtext("NlmUniqueID")
+    name = element.findtext("Name")
+    issns = [(issn.text, issn.attrib["type"]) for issn in element.findall("Issn")]
+    # ActivityFlag is either "0" or "1"
+    term = Term(
+        reference=Reference(prefix=PREFIX, identifier=nlm_id, name=name),
+    )
+    for synonym in element.findall("Alias"):
+        term.append_synonym(synonym.text)
+    for issn, _issn_type in issns:
+        # TODO include ISSN type, this is important
+        #  to determine a "canonical" one
+        term.append_xref(Reference(prefix="issn", identifier=issn))
+    if start_year := element.findtext("StartYear"):
+        term.annotate_integer(START_YEAR, start_year)
+    if end_year := element.findtext("EndYear"):
+        term.annotate_integer(END_YEAR, end_year)
+    if publisher_reference := journal_id_to_publisher_key.get(term.identifier):
+        term.annotate_object(PUBLISHER, publisher_reference)
+    return term
+
+
+if __name__ == "__main__":
+    NLMCatalogGetter().cli()

src/pyobo/struct/struct.py

@@ -494,6 +494,10 @@ def annotate_boolean(self, prop: ReferenceHint, value: bool) -> Self:
             prop, str(value).lower(), Reference(prefix="xsd", identifier="boolean")
         )
 
+    def annotate_integer(self, prop: ReferenceHint, value: str) -> Self:
+        """Append an object annotation."""
+        return self.annotate_literal(prop, value, Reference(prefix="xsd", identifier="integer"))
+
     def _definition_fp(self) -> str:
         definition = obo_escape_slim(self.definition) if self.definition else ""
         return f'"{definition}" [{comma_separate_references(self.provenance)}]'
@@ -570,7 +574,7 @@ def _emit_relations(
         for typedef, reference in self.iterate_relations():
             _typedef_warn(prefix=ontology_prefix, predicate=typedef, typedefs=typedefs)
             predicate_reference = self._reference(typedef, ontology_prefix)
-            s = f"relationship: {predicate_reference} {reference.preferred_curie}"
+            s = f"relationship: {predicate_reference} {self._reference(reference, ontology_prefix)}"
             if typedef.name or reference.name:
                 s += " !"
                 if typedef.name:
@@ -592,7 +596,7 @@ def _emit_object_properties(
             _typedef_warn(prefix=ontology_prefix, predicate=predicate, typedefs=typedefs)
             predicate_curie = self._reference(predicate, ontology_prefix)
             for value in sorted(values):
-                yv = f"{predicate_curie} {value.preferred_curie}"
+                yv = f"{predicate_curie} {self._reference(value, ontology_prefix)}"
                 if predicate.name and value.name:
                     yv += f" ! {predicate.name} {value.name}"
                 yield yv