xinyin025/LangBot
1
0
Fork 0
mirror of https://github.com/langbot-app/LangBot.git synced 2026-07-22 12:26:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: In the runner, every 8 tokens yield

Browse Source
This commit is contained in:
Dong_master
2025-08-03 23:23:51 +08:00
parent 48d11540ae
commit 44ac8b2b63
4 changed files with 20 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pkg/provider/modelmgr/requesters/chatcmpl.py
+1 -2
View File
@@ -189,8 +189,7 @@ class OpenAIChatCompletions(requester.LLMAPIRequester):
delta_message.is_final = True delta_message.is_final = True
delta_message.content = current_content delta_message.content = current_content
if chunk_idx % 64 == 0 or delta_message.is_final: yield delta_message
yield delta_message
# return # return
async def _closure( async def _closure(
pkg/provider/modelmgr/requesters/giteeaichatcmpl.py
+1 -2
View File
@@ -195,5 +195,4 @@ class GiteeAIChatCompletions(chatcmpl.OpenAIChatCompletions):
delta_message.is_final = True delta_message.is_final = True
delta_message.content = current_content delta_message.content = current_content
if chunk_idx % 64 == 0 or delta_message.is_final: yield delta_message
yield delta_message
pkg/provider/modelmgr/requesters/modelscopechatcmpl.py
+1 -2
View File
@@ -286,8 +286,7 @@ class ModelScopeChatCompletions(requester.LLMAPIRequester):
delta_message.is_final = True delta_message.is_final = True
delta_message.content = current_content delta_message.content = current_content
if chunk_idx % 64 == 0 or delta_message.is_final: yield delta_message
yield delta_message
# return # return
async def invoke_llm( async def invoke_llm(
pkg/provider/runners/localagent.py
+17 -14
View File
@@ -47,6 +47,7 @@ class LocalAgentRunner(runner.RequestRunner):
else: else:
# 流式输出，需要处理工具调用 # 流式输出，需要处理工具调用
tool_calls_map: dict[str, llm_entities.ToolCall] = {} tool_calls_map: dict[str, llm_entities.ToolCall] = {}
msg_idx = 0
async for msg in query.use_llm_model.requester.invoke_llm_stream( async for msg in query.use_llm_model.requester.invoke_llm_stream(
query, query,
query.use_llm_model, query.use_llm_model,
@@ -54,7 +55,9 @@ class LocalAgentRunner(runner.RequestRunner):
query.use_funcs, query.use_funcs,
extra_args=query.use_llm_model.model_entity.extra_args, extra_args=query.use_llm_model.model_entity.extra_args,
): ):
yield msg msg_idx = msg_idx + 1
if msg_idx % 8 == 0 or msg.is_final:
yield msg
if msg.tool_calls: if msg.tool_calls:
for tool_call in msg.tool_calls: for tool_call in msg.tool_calls:
if tool_call.id not in tool_calls_map: if tool_call.id not in tool_calls_map:
@@ -115,19 +118,19 @@ class LocalAgentRunner(runner.RequestRunner):
extra_args=query.use_llm_model.model_entity.extra_args, extra_args=query.use_llm_model.model_entity.extra_args,
): ):
yield msg yield msg
if msg.tool_calls: if msg.tool_calls:
for tool_call in msg.tool_calls: for tool_call in msg.tool_calls:
if tool_call.id not in tool_calls_map: if tool_call.id not in tool_calls_map:
tool_calls_map[tool_call.id] = llm_entities.ToolCall( tool_calls_map[tool_call.id] = llm_entities.ToolCall(
id=tool_call.id, id=tool_call.id,
type=tool_call.type, type=tool_call.type,
function=llm_entities.FunctionCall( function=llm_entities.FunctionCall(
name=tool_call.function.name if tool_call.function else '', arguments='' name=tool_call.function.name if tool_call.function else '', arguments=''
), ),
) )
if tool_call.function and tool_call.function.arguments: if tool_call.function and tool_call.function.arguments:
# 流式处理中，工具调用参数可能分多个chunk返回，需要追加而不是覆盖 # 流式处理中，工具调用参数可能分多个chunk返回，需要追加而不是覆盖
tool_calls_map[tool_call.id].function.arguments += tool_call.function.arguments tool_calls_map[tool_call.id].function.arguments += tool_call.function.arguments
final_msg = llm_entities.Message( final_msg = llm_entities.Message(
role=msg.role, role=msg.role,
content=msg.all_content, content=msg.all_content,