Skip to content

Issue #116: Support <think> tags #117

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 17 additions & 4 deletions src/api/models/bedrock.py
Original file line number Diff line number Diff line change
Expand Up @@ -211,6 +211,7 @@ def chat_stream(self, chat_request: ChatRequest) -> AsyncIterable[bytes]:
response = self._invoke_bedrock(chat_request, stream=True)
message_id = self.generate_message_id()
stream = response.get("stream")
self.think_emitted = False
for chunk in stream:
stream_response = self._create_response_stream(
model_id=chat_request.model, message_id=message_id, chunk=chunk
Expand All @@ -235,6 +236,7 @@ def chat_stream(self, chat_request: ChatRequest) -> AsyncIterable[bytes]:

# return an [DONE] message at the end.
yield self.stream_response_to_bytes()
self.think_emitted = False # Cleanup

def _parse_system_prompts(self, chat_request: ChatRequest) -> list[dict[str, str]]:
"""Create system prompts.
Expand Down Expand Up @@ -498,6 +500,9 @@ def _create_response(
message.content = c["text"]
else:
logger.warning("Unknown tag in message content " + ",".join(c.keys()))
if message.reasoning_content:
message.content = f"<think>{message.reasoning_content}</think>{message.content}"
message.reasoning_content = None

response = ChatResponse(
id=message_id,
Expand Down Expand Up @@ -566,11 +571,19 @@ def _create_response_stream(
content=delta["text"],
)
elif "reasoningContent" in delta:
# ignore "signature" in the delta.
if "text" in delta["reasoningContent"]:
message = ChatResponseMessage(
reasoning_content=delta["reasoningContent"]["text"],
)
content = delta["reasoningContent"]["text"]
if not self.think_emitted:
# Port of "content_block_start" with "thinking"
content = "<think>" + content
self.think_emitted = True
message = ChatResponseMessage(content=content)
elif "signature" in delta["reasoningContent"]:
# Port of "signature_delta"
if self.think_emitted:
message = ChatResponseMessage(content="\n </think> \n\n")
else:
return None # Ignore signature if no <think> started
else:
# tool use
index = chunk["contentBlockDelta"]["contentBlockIndex"] - 1
Expand Down