Merge pull request #275 from aaccomazzi/master

Fix bug with metadata with multiple origins
adsabs · Aug 16, 2022 · 128d535 · 128d535
2 parents 047429d + 7b84ef8
commit 128d535
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 1 deletion.
diff --git a/aip/classic/merger.py b/aip/classic/merger.py
@@ -275,7 +275,7 @@ def _getOriginPriority(self,f,field):
     origins = f[1]['origin'].split('; ')
     o = origins.pop()
     for i in origins:
-      o = i if p.get(i.upper(),0) >= p.get(o1.upper(),0) else o
+      o = i if p.get(i.upper(),0) >= p.get(o.upper(),0) else o
     # if origin not defined, default to 'PUBLISHER'
     P = p.get(o.upper(),p.get('PUBLISHER',0))
     return P

diff --git a/tests/classic/test_merger.py b/tests/classic/test_merger.py
@@ -80,11 +80,13 @@ def test_doiMerger(self):
         B1 = {'tempdata':{'origin':'PUBLISHER','type':'general'}}
         B2 = {'tempdata':{'origin':'ARXIV','type':'general'}}
         B3 = {'tempdata':{'origin':'SIMBAD','type':'general'}}
+        B4 = {'tempdata':{'origin':'JST; CROSSREF','type':'general'}}
 
         # 2022NatAs...6..331D
         B1['doi'] = [ '10.1038/s41550-021-01558-y' ]
         B2['doi'] = [ '10.48550/arXiv.2201.05617' ]
         B3['doi'] = [ '10.1038/s41550-021-01558-y' ]
+        B4['doi'] = [ '10.1038/S41550-021-01558-Y' ] # note different case
 
         blocks = [B1,B2]
         m = merger.Merger(blocks)
@@ -107,6 +109,14 @@ def test_doiMerger(self):
         expectedResults = { 'doi': [ '10.1038/s41550-021-01558-y', '10.48550/arXiv.2201.05617' ], 'altpublications': [] }
         self.assertEqual(results,expectedResults)
 
+        blocks = [B4,B2,B1,B3]
+        m = merger.Merger(blocks)
+        m.merge()
+        results = m.block
+        # the list gets eventually uniqued in solr_adapter
+        expectedResults = { 'doi': [ '10.1038/S41550-021-01558-Y', '10.1038/s41550-021-01558-y', '10.48550/arXiv.2201.05617' ], 'altpublications': [] }
+        self.assertEqual(results,expectedResults)
+
 
     def test_want_datetime(self):
         m = merger.Merger([])