Reorganize tests into unit/integration, add EuropePMC/Crossref/ORCID tests, fix parse() P698 handling

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

Author: ghuron

CLAUDE.md

@@ -2,8 +2,11 @@
 
 ## Testing
 
-Unit tests use Python's built-in `unittest` framework. Run with:
+Unit tests use Python's built-in `unittest` framework.
 
 ```
-python -m unittest discover tests/
+python -m unittest discover tests/unit/       # unit tests
+python -m unittest discover tests/integration/ # integration tests (hit real APIs)
 ```
+
+Connector tests (`tests/integration/`) make real network calls and are treated as integration tests.

src/wdpy/connectors/__init__.py

@@ -1,4 +1,5 @@
 from .astro import *
+from .arxiv import *
 from .exoplanet import *
 from .simbad import *
 from .ads import *

src/wdpy/connectors/arxiv.py

@@ -8,6 +8,8 @@ def parse(self, text: str) -> bool:
             self.patch = []
             return True
         id_text = getattr(entry.find('w3:id', ns), 'text', '') or ''
+        if self.patch and self.patch[0].mainsnak.property == 'P818' and '/abs/' not in id_text:
+            return False
         arxiv_id = id_text.split('/')[-1].split('v')[0]
         self.add_claim('P31', 'Q13442814')
         if arxiv_id: self.add_claim('P953', 'https://arxiv.org/pdf/' + arxiv_id)

src/wdpy/connectors/europepmc.py

@@ -6,6 +6,8 @@ class EuropePMC(SourceItem):
     def parse(self, text: str) -> bool:
         results = json.loads(text).get('resultList', {}).get('result', [])
         if len(results) != 1:
+            if self.patch and self.patch[0].mainsnak.property == 'P698':
+                return False
             self.patch = []
             return True
         d = results[0]

src/wdpy/source_item.py

@@ -156,7 +156,12 @@ def extract(cls, ident: Statement) -> Optional[SourceItem]:
         if not (req := cls.make_request(ident)):
             return None
         url = req.full_url
-        if not (resp := build_opener().open(req, timeout=30)):
+        try:
+            resp = build_opener().open(req, timeout=30)
+        except urllib.error.HTTPError as e:
+            resp = e
+        except Exception as e:
+            logging.error('Request failed for %s: %s', url, e)
             return None
         handled = cls._config.get('extract', [])
         with resp:

tests/integration/__init__.py

@@ -3,7 +3,7 @@
 from wdpy.connectors import ADS
 
 
-class TestADS(TestCase):
+class TestExtract(TestCase):
     def test_doi_returns_result(self):
         result = ADS.extract(Statement(Snak('P356', ('10.1088/2041-8205/763/1/L1',))))
         self.assertEqual(result.patch[0].mainsnak.property, 'P356')

tests/test_ads.py tests/integration/test_ads.pytests/test_ads.py renamed to tests/integration/test_ads.py

@@ -0,0 +1,21 @@
+from unittest import TestCase
+
+from wdpy import Snak, Statement
+from wdpy.connectors import ArxivItem
+
+
+class TestExtract(TestCase):
+    def test_arxiv_id_returns_result(self):
+        result = ArxivItem.extract(Statement(Snak('P818', ('1309.0951',))))
+        self.assertEqual(result.patch[0].mainsnak.property, 'P818')
+
+    def test_doi_returns_result(self):
+        result = ArxivItem.extract(Statement(Snak('P356', ('10.4171/161',))))
+        self.assertEqual(result.patch[0].mainsnak.property, 'P356')
+
+    def test_nonexistent_arxiv_id_returns_none(self):
+        self.assertIsNone(ArxivItem.extract(Statement(Snak('P818', ('X',)))))
+
+    def test_nonexistent_doi_returns_empty_patch(self):
+        result = ArxivItem.extract(Statement(Snak('P356', ('10.99999/nonexistent',))))
+        self.assertEqual(result.patch, [])

tests/integration/test_arxiv.py

@@ -0,0 +1,13 @@
+from unittest import TestCase
+from wdpy import Snak, Statement
+from wdpy.connectors.crossref import Crossref
+
+
+class TestExtract(TestCase):
+    def test_doi_returns_result(self):
+        result = Crossref.extract(Statement(Snak('P356', ('10.1088/2041-8205/763/1/L1',))))
+        self.assertEqual(result.patch[0].mainsnak.property, 'P356')
+
+    def test_nonexistent_doi_returns_empty_patch(self):
+        result = Crossref.extract(Statement(Snak('P356', ('X',))))
+        self.assertEqual(result.patch, [])