[Frontend] Support count_reasoning_tokens in the Streaming Parser Engine (#45802)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
This commit is contained in:
Chauncey
2026-08-15 06:51:14 -07:00
committed by GitHub
co-authored by mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
parent 5cecfc0137
commit ac2ae8798c
16 changed files with 669 additions and 73 deletions
+10
View File
@@ -367,6 +367,10 @@ class Parser:
tool call extraction via internal stream state.
"""
def count_reasoning_tokens(self, token_ids: Sequence[int]) -> int:
"""Return the number of reasoning tokens in generated token IDs."""
return 0
class DelegatingParser(Parser):
"""
@@ -939,6 +943,12 @@ class DelegatingParser(Parser):
return delta_message
def count_reasoning_tokens(self, token_ids: Sequence[int]) -> int:
"""Count reasoning tokens through the configured reasoning parser."""
if self._reasoning_parser is None:
return 0
return self._reasoning_parser.count_reasoning_tokens(token_ids)
def _flush_engine_parsers(
self, delta_message: DeltaMessage | None
) -> DeltaMessage | None: