fix(extraction): Handle sequential citations better

CHANGES.md

@@ -8,7 +8,7 @@ Features:
 -
 
 Changes:
--
+- Handle sequential citations (in close proximity) better
 
 Fixes:
 - Modifies rendering of AhocorasickTokenizer parameter in API docs II

eyecite/helpers.py

@@ -218,6 +218,7 @@ def _scan_for_case_boundaries(
             word_str.endswith(";")
             or word_str.endswith("”")
             or word_str.endswith('"')
+            or (word_str.endswith("),") and state["v_token"] is not None)
         ):
             state["start_index"] = index + 2
             state["candidate_case_name"] = _extract_text(
@@ -227,7 +228,10 @@ def _scan_for_case_boundaries(
 
         # Handle year before citation
         if re.match(r"\(\d{4}\)", word_str):
-            state["title_starting_index"] = index - 1
+            # Don't override title_starting_index if we already found a v. token
+            # and have a case name set, as this would corrupt the extraction
+            if state["v_token"] is None or not state["candidate_case_name"]:
+                state["title_starting_index"] = index - 1
             state["pre_cite_year"] = word_str[1:5]
             continue
 
@@ -264,11 +268,13 @@ def _scan_for_case_boundaries(
 
         # Handle "v" token - store it but don't break yet
         if _is_v_token(word):
-            state["v_token"] = word
-            state["start_index"] = index - 2
-            state["candidate_case_name"] = _extract_text(
-                words, state["start_index"], state["title_starting_index"]
-            )
+            # Don't overwrite a case name we already found from a more recent v. token
+            if state["v_token"] is None:
+                state["v_token"] = word
+                state["start_index"] = index - 2
+                state["candidate_case_name"] = _extract_text(
+                    words, state["start_index"], state["title_starting_index"]
+                )
             continue
 
         # Break on likely new sentence after "v" token

tests/test_FindTest.py

@@ -828,7 +828,13 @@ def test_find_citations(self):
             # Fix for index error when searching for case name
             ("<p>State v. Luna-Benitez (S53965). Alternative writ issued, dismissed, 342 Or 255</p>",
             [case_citation(volume="342", reporter="Or", page="255")],
-            {'clean_steps': ['html', 'inline_whitespace']})
+            {'clean_steps': ['html', 'inline_whitespace']}),
+            # Sequential full case citations.
+            ("West v. Atkins, 487 U.S. 42, 54-58 (1988), Polk Cty. v. Dodson, 454 U.S. 312, 325-26 (1981), and Monell v. Department of Soc. Servs., 436 U.S. 658, 694 (1978)",
+             [case_citation(volume="487", reporter="U.S.", page="42", metadata={"plaintiff": "West", "defendant": "Atkins", "year": "1988", "pin_cite": "54-58"}),
+              case_citation(volume="454", reporter="U.S.", page="312", metadata={"plaintiff": "Polk Cty.", "defendant": "Dodson", "year": "1981", "pin_cite": "325-26"}),
+              case_citation(volume="436", reporter="U.S.", page="658", metadata={"plaintiff": "Monell", "defendant": "Department of Soc. Servs.", "year": "1978", "pin_cite": "694"}),
+             ]),
         )
 
         # fmt: on