Recognize en-dash and em-dash as pin-cite range separators

CHANGES.md

@@ -11,7 +11,7 @@ Changes:
 -
 
 Fixes:
--
+- Recognize Unicode en-dash (U+2013) and em-dash (U+2014) as pin-cite page-range separators, so ranges like `241–242` are no longer dropped
 
 ## Current
 

eyecite/regexes.py

@@ -209,10 +209,11 @@ def reference_pin_cite_re(regexes):
         )\ ?  # optional space after label
     )?
     (?:
-        # page:paragraph cite, like 123:24-25 or 123:24-124:25:
-        \d+:\d+(?:-\d+(?::\d+)?)?|
-        # page range, like 12 or 12-13:
-        [*]?\d+(?:-\d+)?
+        # page:paragraph cite, like 123:24-25 or 123:24-124:25
+        # (the range separator may also be an en- or em-dash):
+        \d+:\d+(?:[-–—]\d+(?::\d+)?)?|
+        # page range, like 12 or 12-13 (or 12–13 / 12—13):
+        [*]?\d+(?:[-–—]\d+)?
     )
 """
 

tests/test_FindTest.py

@@ -146,6 +146,22 @@ def test_find_citations(self):
                                       'defendant': 'Test',
                                       'court': 'pasuperct',
                                       'pin_cite': '347-348'})]),
+            # Pin-cite page range with a Unicode en-dash (U+2013); common
+            # in typeset opinions and OCR'd text. The full range must be
+            # captured, not truncated to the first page.
+            ('bob Lissner v. Test 1 U.S. 12, 347–348 (4th Cir. 1982)',
+             [case_citation(page='12', year=1982,
+                            metadata={'plaintiff': 'Lissner',
+                                      'defendant': 'Test',
+                                      'court': 'ca4',
+                                      'pin_cite': '347–348'})]),
+            # Pin-cite page range with a Unicode em-dash (U+2014).
+            ('bob Lissner v. Test 1 U.S. 12, 347—348 (4th Cir. 1982)',
+             [case_citation(page='12', year=1982,
+                            metadata={'plaintiff': 'Lissner',
+                                      'defendant': 'Test',
+                                      'court': 'ca4',
+                                      'pin_cite': '347—348'})]),
             # Test with court string exact match
             ('Commonwealth v. Muniz, 164 A.3d 1189 (Pa. 2017)',
              [case_citation(page='1189', reporter='A.3d', volume='164', year=2017,