From 722e4321bfa6431136b1e03f0c9e4ac11903c4df Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@robocracy.org>
Date: Wed, 12 Sep 2018 11:06:06 -0700
Subject: extid support for crossref importer

---
 python/fatcat/crossref_importer.py     |  31 ++++++++++++++++++++++++++++++-
 python/fatcat_import.py                |   7 +++++--
 python/tests/crossref.py               |   2 +-
 python/tests/files/example_map.sqlite3 | Bin 0 -> 72704 bytes
 4 files changed, 36 insertions(+), 4 deletions(-)
 create mode 100644 python/tests/files/example_map.sqlite3

diff --git a/python/fatcat/crossref_importer.py b/python/fatcat/crossref_importer.py
index a59d0a45..a2e14ed1 100644
--- a/python/fatcat/crossref_importer.py
+++ b/python/fatcat/crossref_importer.py
@@ -1,6 +1,7 @@
 
 import sys
 import json
+import sqlite3
 import itertools
 import fatcat_client
 from fatcat.importer_common import FatcatImporter
@@ -8,10 +9,31 @@ from fatcat.importer_common import FatcatImporter
 
 class FatcatCrossrefImporter(FatcatImporter):
 
-    def __init__(self, host_url, issn_map_file, create_containers=True):
+    def __init__(self, host_url, issn_map_file, extid_map_file=None, create_containers=True):
         super().__init__(host_url, issn_map_file)
+        self.extid_map_db = None
+        if extid_map_file:
+            db_uri = "file:{}?mode=ro".format(extid_map_file)
+            print("Using external ID map: {}".format(db_uri))
+            self.extid_map_db = sqlite3.connect(db_uri, uri=True)
+        else:
+            print("Not using external ID map")
         self.create_containers = create_containers
 
+    def lookup_ext_ids(self, doi):
+        if self.extid_map_db is None:
+            return dict(core_id=None, pmid=None, pmcid=None, wikidata_qid=None)
+        row = self.extid_map_db.execute("SELECT core, pmid, pmcid, wikidata FROM ids WHERE doi=? LIMIT 1",
+            [doi.lower()]).fetchone()
+        if row is None:
+            return dict(core_id=None, pmid=None, pmcid=None, wikidata_qid=None)
+        row = [str((cell or None)) for cell in row]
+        return dict(
+            core_id=row[0],
+            pmid=row[1],
+            pmcid=row[2],
+            wikidata_qid=row[3])
+
     def parse_crossref_dict(self, obj):
         """
         obj is a python dict (parsed from json).
@@ -103,6 +125,9 @@ class FatcatCrossrefImporter(FatcatImporter):
             'license': obj.get('license', [dict(URL=None)])[0]['URL'] or None,
             'alternative-id': obj.get('alternative-id', [])})
 
+        # external identifiers
+        extids = self.lookup_ext_ids(doi=obj['DOI'].lower())
+
         re = fatcat_client.ReleaseEntity(
             work_id=None,
             title=obj['title'][0],
@@ -111,6 +136,10 @@ class FatcatCrossrefImporter(FatcatImporter):
             container_id=container_id,
             release_type=obj['type'],
             doi=obj['DOI'].lower(),
+            core_id=extids['core_id'],
+            pmid=extids['pmid'],
+            pmcid=extids['pmcid'],
+            wikidata_qid=extids['wikidata_qid'],
             release_date=obj['created']['date-time'],
             issue=obj.get('issue'),
             volume=obj.get('volume'),
diff --git a/python/fatcat_import.py b/python/fatcat_import.py
index 2804a210..2f0c746f 100755
--- a/python/fatcat_import.py
+++ b/python/fatcat_import.py
@@ -10,7 +10,7 @@ from fatcat.issn_importer import FatcatIssnImporter
 
 def run_import_crossref(args):
     fci = FatcatCrossrefImporter(args.host_url, args.issn_map_file,
-        create_containers=(not args.no_create_containers))
+        args.extid_map_file, create_containers=(not args.no_create_containers))
     fci.process_batch(args.json_file, size=args.batch_size)
 
 def run_import_orcid(args):
@@ -46,7 +46,10 @@ def main():
         default=sys.stdin, type=argparse.FileType('r'))
     sub_import_crossref.add_argument('issn_map_file',
         help="ISSN to ISSN-L mapping file",
-        default=sys.stdin, type=argparse.FileType('r'))
+        default=None, type=argparse.FileType('r'))
+    sub_import_crossref.add_argument('extid_map_file',
+        help="DOI-to-other-identifiers sqlite3 database",
+        default=None, type=argparse.FileType('r'))
     sub_import_crossref.add_argument('--no-create-containers',
         action='store_true',
         help="skip creation of new container entities based on ISSN")
diff --git a/python/tests/crossref.py b/python/tests/crossref.py
index 402195d7..e9814da2 100644
--- a/python/tests/crossref.py
+++ b/python/tests/crossref.py
@@ -6,7 +6,7 @@ from fatcat.crossref_importer import FatcatCrossrefImporter
 @pytest.fixture(scope="function")
 def crossref_importer():
     with open('tests/files/ISSN-to-ISSN-L.snip.txt', 'r') as issn_file:
-        yield FatcatCrossrefImporter("http://localhost:9411/v0", issn_file)
+        yield FatcatCrossrefImporter("http://localhost:9411/v0", issn_file, 'tests/files/example_map.sqlite3')
 
 def test_crossref_importer_batch(crossref_importer):
     with open('tests/files/crossref-works.2018-01-21.badsample.json', 'r') as f:
diff --git a/python/tests/files/example_map.sqlite3 b/python/tests/files/example_map.sqlite3
new file mode 100644
index 00000000..e9aa1bbe
Binary files /dev/null and b/python/tests/files/example_map.sqlite3 differ
-- 
cgit v1.2.3