Journal quality fixes

src/paperqa/clients/client_data/journal_quality.csv

@@ -5143,6 +5143,8 @@ annual review of neuroscience,2
 annual review of nuclear and particle science,1
 annual review of nutrition,3
 annual review of pathology-mechanisms of disease,2
+annual review of pathology: mechanisms of disease,2
+annual review of pathology,2
 annual review of pharmacology and toxicology,3
 annual review of physical chemistry,1
 annual review of physiology,3
@@ -6057,6 +6059,15 @@ biochimica et biophysica acta: molecular basis of disease,1
 biochimica et biophysica acta: molecular cell research,1
 biochimica et biophysica acta: proteins and proteomics,1
 biochimica et biophysica acta: reviews on cancer,1
+biochimica et biophysica acta (bba) - bioenergetics,1
+biochimica et biophysica acta (bba) - biomembranes,1
+biochimica et biophysica acta (bba) - gene regulatory mechanisms,1
+biochimica et biophysica acta (bba) - general subjects,1
+biochimica et biophysica acta (bba) - molecular and cell biology of lipids,1
+biochimica et biophysica acta (bba) - molecular basis of disease,1
+biochimica et biophysica acta (bba) - molecular cell research,1
+biochimica et biophysica acta (bba) - proteins and proteomics,1
+biochimica et biophysica acta (bba) - reviews on cancer,1
 biochimie,1
 biochip journal,1
 bioconjugate chemistry,1
@@ -17343,6 +17354,8 @@ proceedings of the linnean society of new south wales,1
 proceedings of the london mathematical society,3
 proceedings of the national academy of sciences india section b: biologicalsciences,1
 proceedings of the national academy of sciences of the united states of america,3
+proceedings of the national academy of sciences,3
+pnas,3
 proceedings of the nutrition society,1
 proceedings of the prehistoric society,2
 proceedings of the risø international symposium on materials science,1
@@ -31738,7 +31751,6 @@ proceedings of international conference on the advancement of steam,0
 selected papers of internet research,0
 bat research news,1
 imerides endymasiologias - praktika,1
-scientific reports,0
 ecaade proceedings,0
 traficomin tutkimuksia ja selvityksiä,0
 esignals research,0
@@ -32339,6 +32351,7 @@ radical philosophy review,1
 psychology of popular media,1
 electronic research archive,1
 bmc ecology and evolution,2
+bmc evolutionary biology,2
 annales fennici mathematici,2
 minerva surgery,1
 forces in mechanics,0

src/paperqa/clients/journal_quality.py

@@ -3,7 +3,7 @@
 import csv
 import logging
 import os
-from typing import Any
+from typing import Any, ClassVar
 
 from pydantic import ValidationError
 
@@ -18,6 +18,10 @@
 
 
 class JournalQualityPostProcessor(MetadataPostProcessor[JournalQuery]):
+
+    # these will be deleted from any journal names before querying
+    CASEFOLD_PHRASES_TO_REMOVE: ClassVar[list[str]] = ["amp;"]
+
     def __init__(self, journal_quality_path: os.PathLike | str | None = None) -> None:
         if journal_quality_path is None:
             # Construct the path relative to module
@@ -41,17 +45,22 @@ async def _process(
     ) -> DocDetails:
         if not self.data:
             self.load_data()
+
+        # TODO: not super scalable, but unless we need more than this we can just grugbrain
+        journal_query = query.journal.casefold()
+        for phrase in self.CASEFOLD_PHRASES_TO_REMOVE:
+            journal_query = journal_query.replace(phrase, "")
+
         # docname can be blank since the validation will add it
         # remember, if both have docnames (i.e. key) they are
         # wiped and re-generated with resultant data
         return doc_details + DocDetails(
             doc_id=doc_details.doc_id,  # ensure doc_id is preserved
             dockey=doc_details.dockey,  # ensure dockey is preserved
             source_quality=max(
-                [
-                    self.data.get(query.journal.casefold(), DocDetails.UNDEFINED_JOURNAL_QUALITY),  # type: ignore[union-attr]
-                    self.data.get("the " + query.journal.casefold(), DocDetails.UNDEFINED_JOURNAL_QUALITY),  # type: ignore[union-attr]
-                ]
+                self.data.get(journal_query, DocDetails.UNDEFINED_JOURNAL_QUALITY),  # type: ignore[union-attr]
+                self.data.get("the " + journal_query, DocDetails.UNDEFINED_JOURNAL_QUALITY),  # type: ignore[union-attr]
+                self.data.get(journal_query.replace("&", "and"), DocDetails.UNDEFINED_JOURNAL_QUALITY),  # type: ignore[union-attr]
             ),
         )