Add more details

nationalarchives · Jan 18, 2024 · 3b0fba3 · 3b0fba3
1 parent 0c41e8c
commit 3b0fba3
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 2 deletions.
diff --git a/app/records/schemas/details.py b/app/records/schemas/details.py
@@ -18,6 +18,7 @@ def toJSON(self):
 class Record(Details):
     type: str = "record"
     ref: str | None = None
+    iaid: str | None = None
     former_ref: str | None = None
     title: str | None = None
     description: str | None = None
@@ -36,6 +37,7 @@ def __init__(self, id: str):
 
 class Aggregation(Details):
     type: str = "aggregation"
+    ref: str | None = None
     iaid: str | None = None
     title: str | None = None
     description: str | None = None

diff --git a/app/sources/rosetta/api.py b/app/sources/rosetta/api.py
@@ -89,6 +89,7 @@ def parse_results(self, raw_results, source_url):
         if parsed_data.type() == "record":
             record = Record(parsed_data.id())
             record.ref = parsed_data.identifier()
+            record.iaid = parsed_data.iaid()
             record.former_ref = parsed_data.former_identifier()
             record.title = parsed_data.title()
             record.description = parsed_data.description()
@@ -108,6 +109,7 @@ def parse_results(self, raw_results, source_url):
             return record.toJSON()
         if parsed_data.type() == "aggregation":
             record = Aggregation(parsed_data.id())
+            record.ref = parsed_data.identifier()
             record.iaid = parsed_data.iaid()
             record.title = parsed_data.title()
             record.description = parsed_data.description()

diff --git a/app/sources/rosetta/lib/response_parser.py b/app/sources/rosetta/lib/response_parser.py
@@ -38,7 +38,7 @@ def __init__(self, rosetta_data_source):
         self.source = rosetta_data_source
 
     def strip_scope_and_content(self, markup):
-        document = PyQuery(markup)
+        document = PyQuery(markup.replace("<p/>", ""))
         return str(document("span.scopecontent").contents())
 
     def strip_wrapper_and_split_span(self, markup):
@@ -431,7 +431,7 @@ def administrative_background(self) -> str | None:
                 None,
             ):
                 document = PyQuery(administrative_background)
-                return str(document("span.bioghist").contents().eq(0))
+                return str(document("span.bioghist").contents())
         return None
 
     def functions(self) -> str | None: