Detect ADS bibcode redirects in parse()

When querying by P819 and ADS returns a different bibcode (e.g. arXiv
preprint → published), set prior_ident/new_ident instead of silently
adding the new bibcode to patch. Also switch P819 query to use
identifier: field for broader matching, and skip bibcode from obtain()
when querying by P819 to avoid redundant claims.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: ghuron

src/wdpy/connectors/ads.json

@@ -1,7 +1,7 @@
 {
   "source": "Q752099",
   "properties": {
-    "P819": "https://api.adsabs.harvard.edu/v1/search/query?q=bibcode:\"{}\"",
+    "P819": "https://api.adsabs.harvard.edu/v1/search/query?q=identifier:\"{}\"",
     "P356": "https://api.adsabs.harvard.edu/v1/search/query?q=doi:\"{}\"",
     "P818": "https://api.adsabs.harvard.edu/v1/search/query?q=doi:\"10.48550%2FarXiv.{}\""
   },

src/wdpy/connectors/ads.py

@@ -2,7 +2,7 @@
 import logging
 from typing import Optional
 from urllib.request import Request
-from wdpy import SourceItem, Statement, build_request
+from wdpy import Snak, SourceItem, Statement, build_request
 
 
 class ADS(SourceItem):
@@ -33,7 +33,17 @@ def parse(self, text: str, ident: Statement) -> None:
                     d['page'] = f'{p} - {int(p) + d["page_count"] - 1}'
                 except (ValueError, KeyError):
                     pass
-        self.obtain(d, self._config.get('fields', {}), self._config.get('translate', {}))
+        fields = self._config.get('fields', {})
+        translate = self._config.get('translate', {})
+
+        if ident.mainsnak.property == 'P819':
+            bibcode = d.get('bibcode')
+            if bibcode and bibcode != ident.mainsnak.value[0]:
+                self.prior_ident = ident
+                self.new_ident = Statement(Snak('P819', (bibcode,)))
+            fields = {k: v for k, v in fields.items() if k != 'bibcode'}
+
+        self.obtain(d, fields, translate)
         for i, name in enumerate(d.get('author', []), 1):
             self.add_author(name, _get_orcid(d, i - 1))
         for identifier in d.get('identifier', []):

tests/integration/test_ads.py

@@ -33,3 +33,11 @@ def test_nonexistent_arxiv_returns_empty(self):
         self.assertIsNone(result.prior_ident)
         self.assertIsNone(result.patch)
         self.assertIsNone(result.new_ident)
+
+    def test_bibcode_redirect(self):
+        ident = Statement(Snak('P819', ('2023arXiv230313424H',)))
+        result = ADS.extract(ident)
+        self.assertIs(result.prior_ident, ident)
+        self.assertEqual(result.new_ident.mainsnak.property, 'P819')
+        self.assertEqual(result.new_ident.mainsnak.value, ('2023A&A...673A.114H',))
+        self.assertTrue(result.patch)