partial MARC21 role code expansion, include $4+$e

internetarchive · Sep 20, 2024 · c4f59fc · c4f59fc
1 parent 2cf44af
commit c4f59fc
Show file tree

Hide file tree

Showing 6 changed files with 26 additions and 5 deletions.
diff --git a/openlibrary/catalog/marc/parse.py b/openlibrary/catalog/marc/parse.py
@@ -30,6 +30,23 @@
 re_bracket_field = re.compile(r'^\s*(\[.*\])\.?\s*$')
 
 
+ROLES = {
+    # MARC 21 realtor codes
+    # https://www.loc.gov/marc/relators/relacode.html
+    'art': 'Artist',
+    'aut': 'Author',
+    'clr': 'Colorist',
+    'com': 'Compiler',
+    'edt': 'Editor',
+    'ill': 'Illustrator',
+    'ltr': 'Letterer',
+    'trl': 'Translator',
+    # Non-standard terms from $e
+    'ed': 'Editor',
+    'comp': 'Compiler',
+}
+
+
 def strip_foc(s: str) -> str:
     foc = '[from old catalog]'
     return s[: -len(foc)].rstrip() if s.endswith(foc) else s
@@ -425,7 +442,7 @@ def read_author_person(field: MarcFieldBase, tag: str = '100') -> dict[str, Any]
     and returns an author import dict.
     """
     author: dict[str, Any] = {}
-    contents = field.get_contents('abcde6')
+    contents = field.get_contents('abcde46')
     if 'a' not in contents and 'c' not in contents:
         # Should have at least a name or title.
         return author
@@ -438,6 +455,7 @@ def read_author_person(field: MarcFieldBase, tag: str = '100') -> dict[str, Any]
         ('b', 'numeration'),
         ('c', 'title'),
         ('e', 'role'),
+        ('4', 'role'),
     ]
     for subfield, field_name in subfields:
         if subfield in contents:
@@ -452,6 +470,8 @@ def read_author_person(field: MarcFieldBase, tag: str = '100') -> dict[str, Any]
         ):
             author['alternate_names'] = [author['name']]
             author['name'] = name_from_list(name)
+    if author.get('role') in ROLES:
+        author['role'] = ROLES[author['role']]
     return author
 
 

diff --git a/openlibrary/catalog/marc/tests/test_data/bin_expect/lesnoirsetlesrou0000garl_meta.json b/openlibrary/catalog/marc/tests/test_data/bin_expect/lesnoirsetlesrou0000garl_meta.json
@@ -13,6 +13,7 @@
     {
       "birth_date": "1969",
       "name": "Garlini, Alberto",
+      "role": "Author",
       "entity_type": "person"
     },
     {

diff --git a/openlibrary/catalog/marc/tests/test_data/bin_expect/memoirsofjosephf00fouc_meta.json b/openlibrary/catalog/marc/tests/test_data/bin_expect/memoirsofjosephf00fouc_meta.json
@@ -36,7 +36,7 @@
       "birth_date": "1767",
       "death_date": "1832",
       "name": "Beauchamp, Alph. de",
-      "role": "ed",
+      "role": "Editor",
       "entity_type": "person"
     }
   ],

diff --git a/openlibrary/catalog/marc/tests/test_data/bin_expect/warofrebellionco1473unit_meta.json b/openlibrary/catalog/marc/tests/test_data/bin_expect/warofrebellionco1473unit_meta.json
@@ -62,7 +62,7 @@
       "birth_date": "1849",
       "name": "Cowles, Calvin D.",
       "entity_type": "person",
-      "role": "comp"
+      "role": "Compiler"
     },
     {
       "entity_type": "org",

diff --git a/openlibrary/catalog/marc/tests/test_data/xml_expect/00schlgoog.json b/openlibrary/catalog/marc/tests/test_data/xml_expect/00schlgoog.json
@@ -27,7 +27,7 @@
       "name": "Schlosberg, Leon",
       "date": "d. 1899",
       "entity_type": "person",
-      "role": "ed"
+      "role": "Editor"
     }
   ],
   "publish_places": [

diff --git a/openlibrary/catalog/marc/tests/test_data/xml_expect/warofrebellionco1473unit.json b/openlibrary/catalog/marc/tests/test_data/xml_expect/warofrebellionco1473unit.json
@@ -68,7 +68,7 @@
       "birth_date": "1849",
       "name": "Cowles, Calvin D.",
       "entity_type": "person",
-      "role": "comp"
+      "role": "Compiler"
     },
     {
       "entity_type": "org",