Merge branch 'main' into fix-eyecite-defendants

eyecite/helpers.py

@@ -24,7 +24,6 @@
     POST_SHORT_CITATION_REGEX,
     YEAR_REGEX,
 )
-from eyecite.utils import strip_punct
 
 BACKWARD_SEEK = 28  # Median case name length in the CL db is 28 (2016-02-26)
 
@@ -41,17 +40,26 @@ def get_court_by_paren(paren_string: str) -> Optional[str]:
     Does not work on SCOTUS, since that court lacks parentheticals, and
     needs to be handled after disambiguation has been completed.
     """
-    court_str = strip_punct(paren_string)
+
+    # Remove whitespace and punctuation because citation strings sometimes lack
+    # internal spaces, e.g. "Pa.Super." or "SC" (South Carolina)
+    court_str = re.sub(r"[^\w]", "", paren_string).lower()
 
     court_code = None
     if court_str:
-        # Map the string to a court, if possible.
         for court in courts:
-            # Use startswith because citations are often missing final period,
-            # e.g. "2d Cir"
-            if court["citation_string"].startswith(court_str):
+            s = re.sub(r"[^\w]", "", court["citation_string"]).lower()
+
+            # Check for an exact match first
+            if s == court_str:
+                return str(court["id"])
+
+            # If no exact match, try to record a startswith match for possible
+            # eventual return
+            if s.startswith(court_str):
                 court_code = court["id"]
-                break
+
+        return court_code
 
     return court_code
 

tests/test_FindTest.py

@@ -122,6 +122,19 @@ def test_find_citations(self):
                                       'defendant': 'test',
                                       'court': 'ca4',
                                       'pin_cite': '347-348'})]),
+            # Test with court string without space
+            ('bob lissner v. test 1 U.S. 12, 347-348 (Pa.Super. 1982)',
+             [case_citation(page='12', year=1982,
+                            metadata={'plaintiff': 'lissner',
+                                      'defendant': 'test',
+                                      'court': 'pasuperct',
+                                      '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,
+                            metadata={'plaintiff': 'Commonwealth',
+                                      'defendant': 'Muniz',
+                                      'court': 'pa'})]),
             # Parallel cite with parenthetical
             ('bob lissner v. test 1 U.S. 12, 347-348, 1 S. Ct. 2, 358 (4th Cir. 1982) (overruling foo)',
              [case_citation(page='12', year=1982,
@@ -553,6 +566,12 @@ def test_find_citations(self):
             # Long pin cite -- make sure no catastrophic backtracking in regex
             ('1 U.S. 1, 2277, 2278, 2279, 2280, 2281, 2282, 2283, 2284, 2286, 2287, 2288, 2289, 2290, 2291',
              [case_citation(metadata={'pin_cite': '2277, 2278, 2279, 2280, 2281, 2282, 2283, 2284, 2286, 2287, 2288, 2289, 2290, 2291'})]),
+            ('Commonwealth v. Muniz, 164 A.3d 1189 (Pa. 2017)', [
+                case_citation(volume='164', reporter='A.3d', year=2017,
+                              page='1189',
+                              metadata={'plaintiff': 'Commonwealth', 'defendant': 'Muniz',
+                                        'court': 'pa'})]),
+            ('Foo v. Bar,  1 F.Supp. 1 (SC 1967)', [case_citation(volume='1', reporter='F.Supp.', year=1967, page='1', metadata={'plaintiff': 'Foo', 'defendant': 'Bar', 'court': 'sc'})]),
         )
         # fmt: on
         self.run_test_pairs(test_pairs, "Citation extraction")