another case

2 files changed, 19 insertions, 0 deletions

diff --git a/fuzzycat/common.py b/fuzzycat/common.py
index f03b7cc..d87bef4 100644
--- a/fuzzycat/common.py
+++ b/fuzzycat/common.py
@@ -24,6 +24,7 @@ class OK(str, Enum):
     DOI = 'ok.doi'
     DUMMY = 'ok.dummy'
     FIGSHARE_VERSION = 'ok.figshare_version'
+    PMID_DOI_PAIR = 'ok.pmid_doi_pair'
     PREPRINT_PUBLISHED = 'ok.preprint_published'
     SLUG_TITLE_AUTHOR_MATCH = 'ok.slug_title_author_match'
     TITLE_AUTHOR_MATCH = 'ok.title_author_match'
diff --git a/fuzzycat/verify.py b/fuzzycat/verify.py
index a223e48..e945a15 100644
--- a/fuzzycat/verify.py
+++ b/fuzzycat/verify.py
@@ -135,6 +135,18 @@ class GroupVerifier:
         self.counter["total"] = sum(v for _, v in self.counter.items())
 
 
+def dict_key_exists(doc, path):
+    """
+    Return true, if a value at a given path exists. XXX: probably in glom, too.
+    """
+    try:
+        _ = glom(doc, path)
+    except PathAccessError:
+        return False
+    else:
+        return True
+
+
 def compare(a, b):
     """
     Compare two entities, return match status and reason.
@@ -419,6 +431,12 @@ def compare(a, b):
         if len(a_slug_authors & b_slug_authors) > 0:
             return (Status.STRONG, OK.SLUG_TITLE_AUTHOR_MATCH)
 
+    if any([a_authors, b_authors]) and not (a_authors and b_authors):
+        if a_release_year == b_release_year and a_title_lower == b_title_lower:
+            if ((dict_key_exists(a, "ext_ids.pmid") and not dict_key_exists(a, "ext_ids.doi")) or
+                (dict_key_exists(b, "ext_ids.pmid") and not dict_key_exists(b, "ext_ids.doi"))):
+                return (Status.STRONG, OK.PMID_DOI_PAIR)
+
     if a_authors and len(a_slug_authors & b_slug_authors) == 0:
         # Before we bail out, run an authors similarity check. TODO: This is
         # not the right place, but lives here now, since these cases popped up