fix search fields on compound queries

Author: k-schnickmann

docs/source/api_search/api_search.rst

@@ -13,36 +13,36 @@ Using the Crossref API as an example, the following illustrates how to construct
    from pathlib import Path
    from colrev.packages.crossref.src import crossref_api
    from colrev.writer.write_utils import write_file
-   
+
    from search_query.constants import Fields, Operators
    from search_query.or_query import OrQuery
    from search_query.search_file import SearchFile
-   
+
    def to_crossref_url(query):
        """Translate query into a Crossref-compatible URL."""
        if query.value != Operators.OR:
            raise ValueError("Crossref serializer only supports OR queries.")
-   
+
        query_parts = []
        for child in query.children:
            if child.operator:
                raise ValueError("Nested operators are not supported in Crossref serializer.")
            if child.search_field.value != Fields.TITLE:
                raise ValueError(f"Only title field is supported in Crossref serializer ({child.search_field})")
            query_parts.append(child.value.strip())
-   
+
        # Crossref uses '+' for spaces in query values
        query_string = "+".join(query_parts)
        return f"https://api.crossref.org/works?query.title={query_string}"
-   
+
    if __name__ == "__main__":
-   
+
        query = OrQuery(["microsourcing", "lululemon"], search_field="ti")
-   
+
        url = to_crossref_url(query)
        api_crossref = crossref_api.CrossrefAPI(params={"url": url})
        records = api_crossref.get_records()
-   
+
        sf = SearchFile(
            search_string=query.to_string(),
            platform="crossref",
@@ -53,9 +53,8 @@ Using the Crossref API as an example, the following illustrates how to construct
            description="Search for work authored by Tom Brady",
            tags=["microsourcing", "lululemon", "research"]
        )
-   
+
        sf.save("test/tom_brady_search.json")
-   
+
        records_dict = {record.get_value("doi"): record.get_data() for record in records}
        write_file(records_dict=records_dict, filename=Path("test/crossref_records.bib"))
-

search_query/linter.py

@@ -8,7 +8,7 @@
 import search_query.parser
 from search_query.constants import Colors
 from search_query.constants import ExitCodes
-from search_query.search_file import load_search_file
+from search_query.search_file import SearchFile
 from search_query.utils import format_query_string_pos
 
 

search_query/parser_pubmed.py

@@ -241,7 +241,7 @@ def _parse_compound_query(self, tokens: list) -> Query:
         return Query(
             value=operator_type,
             operator=True,
-            search_field=SearchField(value=Fields.ALL),
+            search_field=None,
             children=list(children),
             position=(query_start_pos, query_end_pos),
         )
@@ -340,7 +340,7 @@ def _expand_combined_fields(self, query: Query, search_fields: list) -> None:
 
         query.value = Operators.OR
         query.operator = True
-        query.search_field = SearchField(value=Fields.ALL)
+        query.search_field = None
         query.children = query_children
 
     def get_query_leaves(self, query: Query) -> list:

test/test_parser_pubmed.py

@@ -61,8 +61,7 @@ def test_tokenization_pubmed(query_str: str, expected_tokens: list) -> None:
     [
         (
             '(eHealth[Title/Abstract] OR "eHealth"[MeSH Terms]) AND Review[Publication Type]',
-            # TODO : should the operators have search_field?
-            'AND[all][OR[all][OR[all][eHealth[ti], eHealth[ab]], "eHealth"[mh]], Review[pt]]',
+            'AND[OR[OR[eHealth[ti], eHealth[ab]], "eHealth"[mh]], Review[pt]]',
         )
     ],
 )