修复小bug
This commit is contained in:
h88782481
parent
ae059ffcb8
commit
406a36af89
5 changed files with 92 additions and 13 deletions
|
|
@ -165,6 +165,9 @@ def _handle_openai_stream(
|
|||
for chunk in iter_openai_sse(resp):
|
||||
if chunk is None:
|
||||
_dbg(f'流式响应结束，共 {chunk_count} 个数据片段')
|
||||
close_chunk = think_extractor.finalize()
|
||||
if close_chunk:
|
||||
yield sse_data_message(close_chunk)
|
||||
yield sse_data_message('[DONE]')
|
||||
return
|
||||
|
||||
|
|
|
|||
|
|
@ -93,9 +93,14 @@ def _inject_thinking(data):
|
|||
|
||||
|
||||
def _process_stream(resp):
|
||||
"""处理 /v1/messages 流式响应，检测并注入 thinking 事件"""
|
||||
"""处理 /v1/messages 流式响应，检测并注入 thinking 事件
|
||||
|
||||
追踪上游 content block 的 index，在注入 thinking blocks 时使用独立的 index，
|
||||
并将后续上游 block 的 index 偏移，避免冲突。
|
||||
"""
|
||||
reasoning_buf = ''
|
||||
injected = False
|
||||
index_offset = 0
|
||||
|
||||
for line in resp.iter_lines():
|
||||
if not line:
|
||||
|
|
@ -119,7 +124,6 @@ def _process_stream(resp):
|
|||
|
||||
modified = False
|
||||
|
||||
# 提取 reasoning_content
|
||||
for container_key in ('message', 'delta'):
|
||||
container = event_data.get(container_key)
|
||||
if not container:
|
||||
|
|
@ -129,13 +133,17 @@ def _process_stream(resp):
|
|||
reasoning_buf += rc
|
||||
modified = True
|
||||
|
||||
# 在首个 text_delta 前注入 thinking blocks
|
||||
if reasoning_buf and not injected:
|
||||
if event_data.get('delta', {}).get('type') == 'text_delta':
|
||||
injected = True
|
||||
yield from _emit_thinking_blocks(reasoning_buf)
|
||||
index_offset = 1
|
||||
reasoning_buf = ''
|
||||
|
||||
if index_offset and 'index' in event_data:
|
||||
event_data['index'] = event_data['index'] + index_offset
|
||||
modified = True
|
||||
|
||||
yield f'data: {json.dumps(event_data)}\n\n' if modified else decoded + '\n\n'
|
||||
|
||||
|
||||
|
|
|
|||
19
settings.py
19
settings.py
|
|
@ -5,13 +5,13 @@
|
|||
- model_mappings: Cursor 模型名 → {upstream_model, backend, target_url, api_key, custom_instructions}
|
||||
"""
|
||||
|
||||
import copy
|
||||
import json
|
||||
import os
|
||||
import threading
|
||||
|
||||
from config import Config
|
||||
|
||||
# 数据目录放在项目根目录下，便于 Docker 卷挂载
|
||||
_ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
|
||||
DATA_DIR = os.path.join(_ROOT_DIR, 'data')
|
||||
SETTINGS_FILE = os.path.join(DATA_DIR, 'settings.json')
|
||||
|
|
@ -38,10 +38,10 @@ def load():
|
|||
with open(SETTINGS_FILE, 'r', encoding='utf-8') as f:
|
||||
_cache = {**_DEFAULTS, **json.load(f)}
|
||||
except (json.JSONDecodeError, OSError):
|
||||
_cache = dict(_DEFAULTS)
|
||||
_cache = copy.deepcopy(_DEFAULTS)
|
||||
else:
|
||||
_cache = dict(_DEFAULTS)
|
||||
return dict(_cache)
|
||||
_cache = copy.deepcopy(_DEFAULTS)
|
||||
return copy.deepcopy(_cache)
|
||||
|
||||
|
||||
def save(data):
|
||||
|
|
@ -58,10 +58,13 @@ def save(data):
|
|||
|
||||
|
||||
def get():
|
||||
"""获取当前配置快照，优先返回内存缓存中的结果。"""
|
||||
if _cache is None:
|
||||
return load()
|
||||
return dict(_cache)
|
||||
"""获取当前配置的深拷贝快照，保证调用方修改不影响缓存。"""
|
||||
with _lock:
|
||||
if _cache is None:
|
||||
pass
|
||||
else:
|
||||
return copy.deepcopy(_cache)
|
||||
return load()
|
||||
|
||||
|
||||
def get_url():
|
||||
|
|
|
|||
|
|
@ -126,10 +126,15 @@ def iter_responses_sse(response) -> Iterator[tuple[str, dict[str, Any]]]:
|
|||
|
||||
|
||||
def _iter_event_sse(response) -> Iterator[tuple[str, dict[str, Any]]]:
|
||||
"""解析带 event/data 的通用 SSE 流。"""
|
||||
"""解析带 event/data 的通用 SSE 流。
|
||||
|
||||
SSE 规范中空行是事件分隔符，遇到空行时重置 event_type，
|
||||
避免前一个事件的类型泄漏到下一个事件。
|
||||
"""
|
||||
event_type = ''
|
||||
for line in response.iter_lines():
|
||||
if not line:
|
||||
event_type = ''
|
||||
continue
|
||||
decoded = line.decode('utf-8', errors='replace')
|
||||
if decoded.startswith('event:'):
|
||||
|
|
|
|||
|
|
@ -30,24 +30,84 @@ class ThinkTagExtractor:
|
|||
|
||||
处理跨 chunk 的 <think>...</think> 标签，将标签内的文本
|
||||
转为 reasoning_content delta，标签外的文本保持为 content delta。
|
||||
|
||||
额外处理：
|
||||
- content 和 tool_calls 同时出现时拆分为两个独立 chunk（Cursor 会丢弃同时包含两者的 content）
|
||||
- tool_calls 首次出现时在前面插入换行，确保文本以换行结束
|
||||
- 流结束时如果 think 标签仍未关闭，自动合成关闭 chunk
|
||||
"""
|
||||
|
||||
def __init__(self):
|
||||
"""初始化跨 chunk 的 thinking 状态跟踪。"""
|
||||
self._in_thinking = False
|
||||
self._tool_calls_seen = False
|
||||
|
||||
def process_chunk(self, chunk):
|
||||
"""处理一个流式 chunk，返回转换后的 chunk 列表"""
|
||||
for choice in (chunk.get('choices') or []):
|
||||
delta = choice.get('delta') or {}
|
||||
|
||||
has_tool_calls = bool(delta.get('tool_calls'))
|
||||
has_content = delta.get('content') is not None and delta.get('content') != ''
|
||||
|
||||
# content 和 tool_calls 同时出现：拆分为两个独立事件
|
||||
if has_content and has_tool_calls:
|
||||
results = []
|
||||
content_chunk = self._make(chunk, content=delta['content'])
|
||||
results.extend(self._process_content(content_chunk, delta['content']))
|
||||
tc_chunk = {
|
||||
'id': chunk.get('id', ''),
|
||||
'object': 'chat.completion.chunk',
|
||||
'model': chunk.get('model', ''),
|
||||
'choices': [{'index': 0, 'delta': {'tool_calls': delta['tool_calls']},
|
||||
'finish_reason': choice.get('finish_reason')}],
|
||||
}
|
||||
results.extend(self._handle_tool_calls_chunk(tc_chunk))
|
||||
return results
|
||||
|
||||
if has_tool_calls:
|
||||
return self._handle_tool_calls_chunk(chunk)
|
||||
|
||||
if delta.get('reasoning_content'):
|
||||
return [chunk]
|
||||
content = delta.get('content')
|
||||
if content is None or content == '':
|
||||
return [chunk]
|
||||
return self._split(chunk, content)
|
||||
return self._process_content(chunk, content)
|
||||
return [chunk]
|
||||
|
||||
def finalize(self):
|
||||
"""流结束时调用，如果 think 标签仍未关闭则返回关闭 chunk"""
|
||||
if not self._in_thinking:
|
||||
return None
|
||||
self._in_thinking = False
|
||||
return {
|
||||
'id': '',
|
||||
'object': 'chat.completion.chunk',
|
||||
'model': '',
|
||||
'choices': [{'index': 0, 'delta': {'content': '\n</think>\n\n'}, 'finish_reason': None}],
|
||||
}
|
||||
|
||||
def _process_content(self, chunk, content):
|
||||
"""处理包含 content 的 chunk"""
|
||||
return self._split(chunk, content)
|
||||
|
||||
def _handle_tool_calls_chunk(self, chunk):
|
||||
"""处理包含 tool_calls 的 chunk，首次出现时在前面插入换行"""
|
||||
results = []
|
||||
if not self._tool_calls_seen:
|
||||
self._tool_calls_seen = True
|
||||
if self._in_thinking:
|
||||
self._in_thinking = False
|
||||
results.append(self._make(chunk, content='\n</think>\n\n'))
|
||||
else:
|
||||
results.append(self._make(chunk, content='\n'))
|
||||
elif self._in_thinking:
|
||||
self._in_thinking = False
|
||||
results.append(self._make(chunk, content='\n</think>\n\n'))
|
||||
results.append(chunk)
|
||||
return results
|
||||
|
||||
def _split(self, chunk, text):
|
||||
"""根据 <think> 标签拆分文本为多个 chunk"""
|
||||
results = []
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue