assing: fix author can_claim check

* ref: cern-sis/issues-inspire#467

backend/inspirehep/assign/utils.py

@@ -4,6 +4,7 @@
 #
 # inspirehep is free software; you can redistribute it and/or modify it under
 # the terms of the MIT License; see LICENSE file for more details.
+
 import structlog
 from flask import request
 from inspire_dojson.utils import get_recid_from_ref
@@ -14,6 +15,7 @@
 from invenio_pidstore.models import PersistentIdentifier
 from invenio_records.models import RecordMetadata
 from sqlalchemy.orm.exc import NoResultFound
+from unidecode import unidecode
 
 from inspirehep.accounts.api import get_current_user_orcid
 from inspirehep.records.api import AuthorsRecord
@@ -108,30 +110,35 @@ def can_claim(data, author_profile_recid):
     if not lit_record:
         return False
 
-    author_parsed_name = ParsedName.loads(current_author_profile["name"]["value"])
-    author_names = {
-        current_author_profile["name"]["value"],
-        author_parsed_name.last,
-        str(author_parsed_name),  # removes ',' and puts it in normal order
-    }
-    author_names.update(
-        [
-            author_name.split(",")[0]
-            for author_name in get_value(
-                current_author_profile, "name.name_variants", []
-            )
-        ]
-    )
+    def get_last_names(name):
+        parsed_name = ParsedName.loads(name)
+        # corner case for single name (ie. "Smith")
+        if len(parsed_name) == 1:
+            return {unidecode(parsed_name.first)}
+        # corner case for full names without comma,
+        # we are treating them as last names (ie. "Smith Davis")
+        if "," not in name:
+            names = name.split()
+        else:
+            names = parsed_name.last_list
+
+        last_names = set()
+        for last_name in names:
+            last_name = unidecode(last_name)
+            last_names.add(last_name)
+        return last_names
+
+    author_last_names = set()
+    author_last_names.update(get_last_names(current_author_profile["name"]["value"]))
+    for variant in get_value(current_author_profile, "name.name_variants", []):
+        author_last_names.update(get_last_names(variant))
 
     lit_author = get_author_by_recid(lit_record, int(author_profile_recid))
-    lit_author_parsed_name = ParsedName.loads(lit_author.get("full_name", ""))
-    lit_author_names = {
-        lit_author.get("full_name", ""),
-        lit_author_parsed_name.last,
-        str(lit_author_parsed_name),
-    }
+    lit_author_last_names = set()
+    if lit_author:
+        lit_author_last_names.update(get_last_names(lit_author.get("full_name", "")))
 
-    return lit_author_names & author_names
+    return bool(author_last_names & lit_author_last_names)
 
 
 def _check_names_compability(lit_record, author_parsed_name, last_names_only=False):
@@ -141,7 +148,7 @@ def _check_names_compability(lit_record, author_parsed_name, last_names_only=Fal
     author_name_to_compare = (
         author_parsed_name.last
         if last_names_only
-        else f"{ author_parsed_name.last}, {author_parsed_name.first}".strip(", ")
+        else f"{author_parsed_name.last}, {author_parsed_name.first}".strip(", ")
     )
     matched_authors_recids = [
         recid

backend/inspirehep/assign/views.py

@@ -166,7 +166,7 @@ def assign_different_profile(args):
 
     for literature_id in literature_ids:
         record = LiteratureRecord.get_record_by_pid_value(literature_id)
-        if record.get("curated") and not is_from_author_stub:
+        if record.get("curated_relation") and not is_from_author_stub:
             literature_ids_already_claimed.append(literature_id)
         if not can_claim(record, from_author_recid):
             literature_ids_not_compatible_name.append(literature_id)

backend/tests/unit/assign/test_assign_utils.py

@@ -0,0 +1,50 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2019 CERN.
+#
+# inspirehep is free software; you can redistribute it and/or modify it under
+# the terms of the MIT License; see LICENSE file for more details.
+
+import pytest
+from mock import patch
+
+from inspirehep.assign.utils import can_claim
+
+
+@pytest.mark.parametrize(
+    "authors_full_name, profile_name, name_variants, expected",
+    [
+        ("Smith, John", "Smith, J.", [], True),
+        ("Smith, J.", "Smith, John", [], True),
+        ("Smith Davis, J.", "Smith, Robert", [], True),
+        ("Davis, J.", "Smith Davis, P.", [], True),
+        ("Smith, J.", "Davis, Smith", [], False),
+        ("Smïth, J.", "Smith, J.", [], True),
+        ("Smith Davis", "Smith, J.", [], True),
+        ("Smith, J.", "Smith", [], True),
+    ],
+)
+@patch("inspirehep.assign.utils._get_current_user_author_profile")
+@patch("inspirehep.assign.utils._get_lit_record_from_db")
+@patch("inspirehep.assign.utils.get_author_by_recid")
+def test_can_claim(
+    mock_get_author_by_recid,
+    mock_get_lit_record_from_db,
+    mock_get_current_user_author_profile,
+    authors_full_name,
+    profile_name,
+    name_variants,
+    expected,
+):
+    mock_get_current_user_author_profile.return_value = {
+        "name": {"value": profile_name},
+        "name_variants": name_variants,
+    }
+    mock_get_lit_record_from_db.return_value = {"control_number": 123}
+    mock_get_author_by_recid.return_value = {"full_name": authors_full_name}
+
+    data = {"control_number": 123}
+    author_profile_recid = 1
+
+    result = can_claim(data, author_profile_recid)
+    assert result == expected