scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-09 03:18:44 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 109

fix(minimax): properly handle M2.x streaming reasoning

Browse source 
- M2.x streaming chunks don't have 'reasoning' field like M3
- Detect </think> delimiter to split reasoning from content
- Emit reasoning chunks as reasoning_content immediately
- Emit answer chunks as content after </think> detected
- Both M3 and M2.x verified working via opencode run
This commit is contained in:
renearaos 2026-06-18 20:30:44 -04:00
parent 8b6dffa619
commit 64a0a56776
1 changed files with 26 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

42
litellm/llms/minimax/chat/transformation.py
View file

@ -137,29 +137,39 @@ class MinimaxChatResponseIterator(OpenAIChatCompletionStreamingHandler):
content = delta.get("content", "")
if content and isinstance(content, str):
reasoning_matches = re.findall(
r"<think>(.*?)</think>", content, re.DOTALL
)
if reasoning_matches:
self.pending_reasoning += "".join(reasoning_matches)
if "<think>" in content and "</think>" in content:
reasonings = re.findall(
r"<think>(.*?)</think>", content, re.DOTALL
)
if reasonings:
self.pending_reasoning += "".join(reasonings)
clean_content = re.sub(
r"<think>.*?</think>", "", content, flags=re.DOTALL
)
delta["content"] = clean_content if clean_content else None
delta["content"] = clean_content if clean_content.strip() else None
if self.pending_reasoning:
delta["reasoning_content"] = self.pending_reasoning
self.started_reasoning_content = True
elif self.started_reasoning_content and not self.finished_reasoning_content:
elif "<think>" in content and "</think>" not in content:
clean_content = content.replace("<think>", "", 1)
delta["content"] = None
delta["reasoning_content"] = clean_content.strip()
self.started_reasoning_content = True
elif "</think>" in content and "<think>" not in content:
parts = content.split("</think>", 1)
before = parts[0].strip()
after = parts[1] if len(parts) > 1 else ""
if before:
delta["reasoning_content"] = before
delta["content"] = after if after.strip() else ("" if not after else None)
self.finished_reasoning_content = True
self.pending_reasoning = ""
clean_content = re.sub(
r"</?think>", "", content
)
delta["content"] = clean_content if clean_content else None
elif not self.started_reasoning_content:
if "</think>" in content:
clean_content = content.replace("</think>", "").replace("<think>", "")
delta["content"] = clean_content if clean_content else None
elif has_reasoning:
pass
elif not self.finished_reasoning_content and content.strip():
if not self.started_reasoning_content:
self.started_reasoning_content = True
delta["reasoning_content"] = content.strip()
delta["content"] = None
return super().chunk_parser(chunk)
except Exception: