clarify the case where multiple captures with the same name exists

Author: kmaehashi

python/llguidance/.nfs000000000077e0330000005c

@@ -430,12 +430,16 @@ class LLMatcher:
     def get_capture(self, name: str) -> Optional[bytes]:
         """
         Get the captured bytes for a named capture group.
-        Returns None if the capture group does not exist.
+        If there are multiple captures with the same name (e.g., due to repetition of rules),
+        returns the last matching capture.
+        Returns None if no capture with the given name exists.
         """
 
     def get_captures(self) -> List[Tuple[str, bytes]]:
         """
         Get all captured groups as a list of (name, bytes) tuples.
+        Includes all captures, even when there are multiple captures
+        with the same name (e.g., due to repetition of rules).
         """
 
 

python/llguidance/_lib.pyi

@@ -668,19 +668,20 @@ def test_grammar_warnings() -> None:
 
 
 def test_get_capture() -> None:
-    grm = r"""start: "hello " group1 group2
+    grm = r"""start: "hello " group1 group2+ group3 " end"
 group1[capture,lazy]: /[a-z]+/
-group2[capture="body"]: /[a-z]{4}/"""
+group2[capture="body"]: /[a-z]{4}/
+group3[capture="body"]: /[0-9]{4}/"""
     m = matcher(grm)
     check_err(m)
 
     assert m.get_capture("group1") is None
     assert m.get_capture("body") is None
     assert m.get_capture("non-existent-group") is None
     assert m.get_captures() == []
-    m.consume_tokens(tokenizer().tokenize_str("hello world"))
+    m.consume_tokens(tokenizer().tokenize_str("hello worldabcd1234 end"))
     assert m.is_stopped()
     assert m.get_capture("group1") == b"w"
-    assert m.get_capture("body") == b"orld"
+    assert m.get_capture("body") == b"1234"
     assert m.get_capture("non-existent-group") is None
-    assert m.get_captures() == [("group1", b"w"), ("body", b"orld")]
+    assert m.get_captures() == [("group1", b"w"), ("body", b"orld"), ("body", b"abcd"), ("body", b"1234")]