Losita/mai-bot
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Ruff Format

Browse Source
This commit is contained in:
DrSmoothl
2026-02-21 16:24:24 +08:00
parent 2cb512120b
commit eaef7f0e98
82 changed files with 1881 additions and 1900 deletions
Download Patch File Download Diff File Expand all files Collapse all files

184
src/webui/api/planner.py
View File

@@ -7,6 +7,7 @@
2. 日志列表使用文件名解析时间戳，只在需要时读取完整内容
3. 详情按需加载
"""
import json
from pathlib import Path
from typing import List, Dict, Optional
@@ -21,6 +22,7 @@ PLAN_LOG_DIR = Path("logs/plan")
class ChatSummary(BaseModel):
"""聊天摘要 - 轻量级，不读取文件内容"""
chat_id: str
plan_count: int
latest_timestamp: float
@@ -29,6 +31,7 @@ class ChatSummary(BaseModel):
class PlanLogSummary(BaseModel):
"""规划日志摘要"""
chat_id: str
timestamp: float
filename: str
@@ -41,6 +44,7 @@ class PlanLogSummary(BaseModel):
class PlanLogDetail(BaseModel):
"""规划日志详情"""
type: str
chat_id: str
timestamp: float
@@ -54,6 +58,7 @@ class PlanLogDetail(BaseModel):
class PlannerOverview(BaseModel):
"""规划器总览 - 轻量级统计"""
total_chats: int
total_plans: int
chats: List[ChatSummary]
@@ -61,6 +66,7 @@ class PlannerOverview(BaseModel):
class PaginatedChatLogs(BaseModel):
"""分页的聊天日志列表"""
data: List[PlanLogSummary]
total: int
page: int
@@ -71,7 +77,7 @@ class PaginatedChatLogs(BaseModel):
def parse_timestamp_from_filename(filename: str) -> float:
"""从文件名解析时间戳: 1766497488220_af92bdb1.json -> 1766497488.220"""
try:
timestamp_str = filename.split('_')[0]
timestamp_str = filename.split("_")[0]
# 时间戳是毫秒级，需要转换为秒
return float(timestamp_str) / 1000
except (ValueError, IndexError):
@@ -86,41 +92,39 @@ async def get_planner_overview():
"""
if not PLAN_LOG_DIR.exists():
return PlannerOverview(total_chats=0, total_plans=0, chats=[])
chats = []
total_plans = 0
for chat_dir in PLAN_LOG_DIR.iterdir():
if not chat_dir.is_dir():
continue
# 只统计json文件数量
json_files = list(chat_dir.glob("*.json"))
plan_count = len(json_files)
total_plans += plan_count
if plan_count == 0:
continue
# 从文件名获取最新时间戳
latest_file = max(json_files, key=lambda f: parse_timestamp_from_filename(f.name))
latest_timestamp = parse_timestamp_from_filename(latest_file.name)
chats.append(ChatSummary(
chat_id=chat_dir.name,
plan_count=plan_count,
latest_timestamp=latest_timestamp,
latest_filename=latest_file.name
))
chats.append(
ChatSummary(
chat_id=chat_dir.name,
plan_count=plan_count,
latest_timestamp=latest_timestamp,
latest_filename=latest_file.name,
)
)
# 按最新时间戳排序
chats.sort(key=lambda x: x.latest_timestamp, reverse=True)
return PlannerOverview(
total_chats=len(chats),
total_plans=total_plans,
chats=chats
)
return PlannerOverview(total_chats=len(chats), total_plans=total_plans, chats=chats)
@router.get("/chat/{chat_id}/logs", response_model=PaginatedChatLogs)
@@ -128,7 +132,7 @@ async def get_chat_plan_logs(
chat_id: str,
page: int = Query(1, ge=1),
page_size: int = Query(20, ge=1, le=100),
search: Optional[str] = Query(None, description="搜索关键词，匹配提示词内容")
search: Optional[str] = Query(None, description="搜索关键词，匹配提示词内容"),
):
"""
获取指定聊天的规划日志列表（分页）
@@ -137,73 +141,69 @@ async def get_chat_plan_logs(
"""
chat_dir = PLAN_LOG_DIR / chat_id
if not chat_dir.exists():
return PaginatedChatLogs(
data=[], total=0, page=page, page_size=page_size, chat_id=chat_id
)
return PaginatedChatLogs(data=[], total=0, page=page, page_size=page_size, chat_id=chat_id)
# 先获取所有文件并按时间戳排序
json_files = list(chat_dir.glob("*.json"))
json_files.sort(key=lambda f: parse_timestamp_from_filename(f.name), reverse=True)
# 如果有搜索关键词，需要过滤文件
if search:
search_lower = search.lower()
filtered_files = []
for log_file in json_files:
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
prompt = data.get('prompt', '')
prompt = data.get("prompt", "")
if search_lower in prompt.lower():
filtered_files.append(log_file)
except Exception:
continue
json_files = filtered_files
total = len(json_files)
# 分页 - 只读取当前页的文件
offset = (page - 1) * page_size
page_files = json_files[offset:offset + page_size]
page_files = json_files[offset : offset + page_size]
logs = []
for log_file in page_files:
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
reasoning = data.get('reasoning', '')
actions = data.get('actions', [])
action_types = [a.get('action_type', '') for a in actions if a.get('action_type')]
logs.append(PlanLogSummary(
chat_id=data.get('chat_id', chat_id),
timestamp=data.get('timestamp', parse_timestamp_from_filename(log_file.name)),
filename=log_file.name,
action_count=len(actions),
action_types=action_types,
total_plan_ms=data.get('timing', {}).get('total_plan_ms', 0),
llm_duration_ms=data.get('timing', {}).get('llm_duration_ms', 0),
reasoning_preview=reasoning[:100] if reasoning else ''
))
reasoning = data.get("reasoning", "")
actions = data.get("actions", [])
action_types = [a.get("action_type", "") for a in actions if a.get("action_type")]
logs.append(
PlanLogSummary(
chat_id=data.get("chat_id", chat_id),
timestamp=data.get("timestamp", parse_timestamp_from_filename(log_file.name)),
filename=log_file.name,
action_count=len(actions),
action_types=action_types,
total_plan_ms=data.get("timing", {}).get("total_plan_ms", 0),
llm_duration_ms=data.get("timing", {}).get("llm_duration_ms", 0),
reasoning_preview=reasoning[:100] if reasoning else "",
)
)
except Exception:
# 文件读取失败时使用文件名信息
logs.append(PlanLogSummary(
chat_id=chat_id,
timestamp=parse_timestamp_from_filename(log_file.name),
filename=log_file.name,
action_count=0,
action_types=[],
total_plan_ms=0,
llm_duration_ms=0,
reasoning_preview='[读取失败]'
))
return PaginatedChatLogs(
data=logs,
total=total,
page=page,
page_size=page_size,
chat_id=chat_id
)
logs.append(
PlanLogSummary(
chat_id=chat_id,
timestamp=parse_timestamp_from_filename(log_file.name),
filename=log_file.name,
action_count=0,
action_types=[],
total_plan_ms=0,
llm_duration_ms=0,
reasoning_preview="[读取失败]",
)
)
return PaginatedChatLogs(data=logs, total=total, page=page, page_size=page_size, chat_id=chat_id)
@router.get("/log/{chat_id}/{filename}", response_model=PlanLogDetail)
@@ -212,9 +212,9 @@ async def get_log_detail(chat_id: str, filename: str):
log_file = PLAN_LOG_DIR / chat_id / filename
if not log_file.exists():
raise HTTPException(status_code=404, detail="日志文件不存在")
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
return PlanLogDetail(**data)
except Exception as e:
@@ -223,11 +223,12 @@ async def get_log_detail(chat_id: str, filename: str):
# ========== 兼容旧接口 ==========
@router.get("/stats")
async def get_planner_stats():
"""获取规划器统计信息 - 兼容旧接口"""
overview = await get_planner_overview()
# 获取最近10条计划的摘要
recent_plans = []
for chat in overview.chats[:5]: # 从最近5个聊天中获取
@@ -236,17 +237,17 @@ async def get_planner_stats():
recent_plans.extend(chat_logs.data)
except Exception:
continue
# 按时间排序取前10
recent_plans.sort(key=lambda x: x.timestamp, reverse=True)
recent_plans = recent_plans[:10]
return {
"total_chats": overview.total_chats,
"total_plans": overview.total_plans,
"avg_plan_time_ms": 0,
"avg_llm_time_ms": 0,
"recent_plans": recent_plans
"recent_plans": recent_plans,
}
@@ -258,44 +259,43 @@ async def get_chat_list():
@router.get("/all-logs")
async def get_all_logs(
page: int = Query(1, ge=1),
page_size: int = Query(20, ge=1, le=100)
):
async def get_all_logs(page: int = Query(1, ge=1), page_size: int = Query(20, ge=1, le=100)):
"""获取所有规划日志 - 兼容旧接口"""
if not PLAN_LOG_DIR.exists():
return {"data": [], "total": 0, "page": page, "page_size": page_size}
# 收集所有文件
all_files = []
for chat_dir in PLAN_LOG_DIR.iterdir():
if chat_dir.is_dir():
for log_file in chat_dir.glob("*.json"):
all_files.append((chat_dir.name, log_file))
# 按时间戳排序
all_files.sort(key=lambda x: parse_timestamp_from_filename(x[1].name), reverse=True)
total = len(all_files)
offset = (page - 1) * page_size
page_files = all_files[offset:offset + page_size]
page_files = all_files[offset : offset + page_size]
logs = []
for chat_id, log_file in page_files:
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
reasoning = data.get('reasoning', '')
logs.append({
"chat_id": data.get('chat_id', chat_id),
"timestamp": data.get('timestamp', parse_timestamp_from_filename(log_file.name)),
"filename": log_file.name,
"action_count": len(data.get('actions', [])),
"total_plan_ms": data.get('timing', {}).get('total_plan_ms', 0),
"llm_duration_ms": data.get('timing', {}).get('llm_duration_ms', 0),
"reasoning_preview": reasoning[:100] if reasoning else ''
})
reasoning = data.get("reasoning", "")
logs.append(
{
"chat_id": data.get("chat_id", chat_id),
"timestamp": data.get("timestamp", parse_timestamp_from_filename(log_file.name)),
"filename": log_file.name,
"action_count": len(data.get("actions", [])),
"total_plan_ms": data.get("timing", {}).get("total_plan_ms", 0),
"llm_duration_ms": data.get("timing", {}).get("llm_duration_ms", 0),
"reasoning_preview": reasoning[:100] if reasoning else "",
}
)
except Exception:
continue
return {"data": logs, "total": total, "page": page, "page_size": page_size}
return {"data": logs, "total": total, "page": page, "page_size": page_size}

163
src/webui/api/replier.py
View File

@@ -7,6 +7,7 @@
2. 日志列表使用文件名解析时间戳，只在需要时读取完整内容
3. 详情按需加载
"""
import json
from pathlib import Path
from typing import List, Dict, Optional
@@ -21,6 +22,7 @@ REPLY_LOG_DIR = Path("logs/reply")
class ReplierChatSummary(BaseModel):
"""聊天摘要 - 轻量级，不读取文件内容"""
chat_id: str
reply_count: int
latest_timestamp: float
@@ -29,6 +31,7 @@ class ReplierChatSummary(BaseModel):
class ReplyLogSummary(BaseModel):
"""回复日志摘要"""
chat_id: str
timestamp: float
filename: str
@@ -41,6 +44,7 @@ class ReplyLogSummary(BaseModel):
class ReplyLogDetail(BaseModel):
"""回复日志详情"""
type: str
chat_id: str
timestamp: float
@@ -57,6 +61,7 @@ class ReplyLogDetail(BaseModel):
class ReplierOverview(BaseModel):
"""回复器总览 - 轻量级统计"""
total_chats: int
total_replies: int
chats: List[ReplierChatSummary]
@@ -64,6 +69,7 @@ class ReplierOverview(BaseModel):
class PaginatedReplyLogs(BaseModel):
"""分页的回复日志列表"""
data: List[ReplyLogSummary]
total: int
page: int
@@ -74,7 +80,7 @@ class PaginatedReplyLogs(BaseModel):
def parse_timestamp_from_filename(filename: str) -> float:
"""从文件名解析时间戳: 1766497488220_af92bdb1.json -> 1766497488.220"""
try:
timestamp_str = filename.split('_')[0]
timestamp_str = filename.split("_")[0]
# 时间戳是毫秒级，需要转换为秒
return float(timestamp_str) / 1000
except (ValueError, IndexError):
@@ -89,41 +95,39 @@ async def get_replier_overview():
"""
if not REPLY_LOG_DIR.exists():
return ReplierOverview(total_chats=0, total_replies=0, chats=[])
chats = []
total_replies = 0
for chat_dir in REPLY_LOG_DIR.iterdir():
if not chat_dir.is_dir():
continue
# 只统计json文件数量
json_files = list(chat_dir.glob("*.json"))
reply_count = len(json_files)
total_replies += reply_count
if reply_count == 0:
continue
# 从文件名获取最新时间戳
latest_file = max(json_files, key=lambda f: parse_timestamp_from_filename(f.name))
latest_timestamp = parse_timestamp_from_filename(latest_file.name)
chats.append(ReplierChatSummary(
chat_id=chat_dir.name,
reply_count=reply_count,
latest_timestamp=latest_timestamp,
latest_filename=latest_file.name
))
chats.append(
ReplierChatSummary(
chat_id=chat_dir.name,
reply_count=reply_count,
latest_timestamp=latest_timestamp,
latest_filename=latest_file.name,
)
)
# 按最新时间戳排序
chats.sort(key=lambda x: x.latest_timestamp, reverse=True)
return ReplierOverview(
total_chats=len(chats),
total_replies=total_replies,
chats=chats
)
return ReplierOverview(total_chats=len(chats), total_replies=total_replies, chats=chats)
@router.get("/chat/{chat_id}/logs", response_model=PaginatedReplyLogs)
@@ -131,7 +135,7 @@ async def get_chat_reply_logs(
chat_id: str,
page: int = Query(1, ge=1),
page_size: int = Query(20, ge=1, le=100),
search: Optional[str] = Query(None, description="搜索关键词，匹配提示词内容")
search: Optional[str] = Query(None, description="搜索关键词，匹配提示词内容"),
):
"""
获取指定聊天的回复日志列表（分页）
@@ -140,71 +144,67 @@ async def get_chat_reply_logs(
"""
chat_dir = REPLY_LOG_DIR / chat_id
if not chat_dir.exists():
return PaginatedReplyLogs(
data=[], total=0, page=page, page_size=page_size, chat_id=chat_id
)
return PaginatedReplyLogs(data=[], total=0, page=page, page_size=page_size, chat_id=chat_id)
# 先获取所有文件并按时间戳排序
json_files = list(chat_dir.glob("*.json"))
json_files.sort(key=lambda f: parse_timestamp_from_filename(f.name), reverse=True)
# 如果有搜索关键词，需要过滤文件
if search:
search_lower = search.lower()
filtered_files = []
for log_file in json_files:
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
prompt = data.get('prompt', '')
prompt = data.get("prompt", "")
if search_lower in prompt.lower():
filtered_files.append(log_file)
except Exception:
continue
json_files = filtered_files
total = len(json_files)
# 分页 - 只读取当前页的文件
offset = (page - 1) * page_size
page_files = json_files[offset:offset + page_size]
page_files = json_files[offset : offset + page_size]
logs = []
for log_file in page_files:
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
output = data.get('output', '')
logs.append(ReplyLogSummary(
chat_id=data.get('chat_id', chat_id),
timestamp=data.get('timestamp', parse_timestamp_from_filename(log_file.name)),
filename=log_file.name,
model=data.get('model', ''),
success=data.get('success', True),
llm_ms=data.get('timing', {}).get('llm_ms', 0),
overall_ms=data.get('timing', {}).get('overall_ms', 0),
output_preview=output[:100] if output else ''
))
output = data.get("output", "")
logs.append(
ReplyLogSummary(
chat_id=data.get("chat_id", chat_id),
timestamp=data.get("timestamp", parse_timestamp_from_filename(log_file.name)),
filename=log_file.name,
model=data.get("model", ""),
success=data.get("success", True),
llm_ms=data.get("timing", {}).get("llm_ms", 0),
overall_ms=data.get("timing", {}).get("overall_ms", 0),
output_preview=output[:100] if output else "",
)
)
except Exception:
# 文件读取失败时使用文件名信息
logs.append(ReplyLogSummary(
chat_id=chat_id,
timestamp=parse_timestamp_from_filename(log_file.name),
filename=log_file.name,
model='',
success=False,
llm_ms=0,
overall_ms=0,
output_preview='[读取失败]'
))
return PaginatedReplyLogs(
data=logs,
total=total,
page=page,
page_size=page_size,
chat_id=chat_id
)
logs.append(
ReplyLogSummary(
chat_id=chat_id,
timestamp=parse_timestamp_from_filename(log_file.name),
filename=log_file.name,
model="",
success=False,
llm_ms=0,
overall_ms=0,
output_preview="[读取失败]",
)
)
return PaginatedReplyLogs(data=logs, total=total, page=page, page_size=page_size, chat_id=chat_id)
@router.get("/log/{chat_id}/{filename}", response_model=ReplyLogDetail)
@@ -213,23 +213,23 @@ async def get_reply_log_detail(chat_id: str, filename: str):
log_file = REPLY_LOG_DIR / chat_id / filename
if not log_file.exists():
raise HTTPException(status_code=404, detail="日志文件不存在")
try:
with open(log_file, 'r', encoding='utf-8') as f:
with open(log_file, "r", encoding="utf-8") as f:
data = json.load(f)
return ReplyLogDetail(
type=data.get('type', 'reply'),
chat_id=data.get('chat_id', chat_id),
timestamp=data.get('timestamp', 0),
prompt=data.get('prompt', ''),
output=data.get('output', ''),
processed_output=data.get('processed_output', []),
model=data.get('model', ''),
reasoning=data.get('reasoning', ''),
think_level=data.get('think_level', 0),
timing=data.get('timing', {}),
error=data.get('error'),
success=data.get('success', True)
type=data.get("type", "reply"),
chat_id=data.get("chat_id", chat_id),
timestamp=data.get("timestamp", 0),
prompt=data.get("prompt", ""),
output=data.get("output", ""),
processed_output=data.get("processed_output", []),
model=data.get("model", ""),
reasoning=data.get("reasoning", ""),
think_level=data.get("think_level", 0),
timing=data.get("timing", {}),
error=data.get("error"),
success=data.get("success", True),
)
except Exception as e:
raise HTTPException(status_code=500, detail=f"读取日志失败: {str(e)}")
@@ -237,11 +237,12 @@ async def get_reply_log_detail(chat_id: str, filename: str):
# ========== 兼容接口 ==========
@router.get("/stats")
async def get_replier_stats():
"""获取回复器统计信息"""
overview = await get_replier_overview()
# 获取最近10条回复的摘要
recent_replies = []
for chat in overview.chats[:5]: # 从最近5个聊天中获取
@@ -250,15 +251,15 @@ async def get_replier_stats():
recent_replies.extend(chat_logs.data)
except Exception:
continue
# 按时间排序取前10
recent_replies.sort(key=lambda x: x.timestamp, reverse=True)
recent_replies = recent_replies[:10]
return {
"total_chats": overview.total_chats,
"total_replies": overview.total_replies,
"recent_replies": recent_replies
"recent_replies": recent_replies,
}
@@ -266,4 +267,4 @@ async def get_replier_stats():
async def get_replier_chat_list():
"""获取所有聊天ID列表"""
overview = await get_replier_overview()
return [chat.chat_id for chat in overview.chats]
return [chat.chat_id for chat in overview.chats]

4
src/webui/dependencies.py
View File

@@ -1,6 +1,6 @@
from typing import Optional
from fastapi import Depends, Cookie, Header, Request, HTTPException
from .core import get_current_token, get_token_manager, check_auth_rate_limit, check_api_rate_limit
from fastapi import Depends, Cookie, Header, Request
from .core import get_current_token, get_token_manager, check_auth_rate_limit
async def require_auth(

21
src/webui/middleware/anti_crawler.py
View File

@@ -124,6 +124,7 @@ SCANNER_SPECIFIC_HEADERS = {
# loose: 宽松模式（较宽松的检测，较高的频率限制）
# basic: 基础模式（只记录恶意访问，不阻止，不限制请求数，不跟踪IP）
# IP白名单配置（从配置文件读取，逗号分隔）
# 支持格式：
# - 精确IP：127.0.0.1, 192.168.1.100
@@ -151,7 +152,7 @@ def _parse_allowed_ips(ip_string: str) -> list:
ip_entry = ip_entry.strip() # 去除空格
if not ip_entry:
continue
# 跳过注释行（以#开头）
if ip_entry.startswith("#"):
continue
@@ -237,19 +238,21 @@ def _convert_wildcard_to_regex(wildcard_pattern: str) -> Optional[str]:
def _get_anti_crawler_config():
"""获取防爬虫配置"""
from src.config.config import global_config
return {
'mode': global_config.webui.anti_crawler_mode,
'allowed_ips': _parse_allowed_ips(global_config.webui.allowed_ips),
'trusted_proxies': _parse_allowed_ips(global_config.webui.trusted_proxies),
'trust_xff': global_config.webui.trust_xff
"mode": global_config.webui.anti_crawler_mode,
"allowed_ips": _parse_allowed_ips(global_config.webui.allowed_ips),
"trusted_proxies": _parse_allowed_ips(global_config.webui.trusted_proxies),
"trust_xff": global_config.webui.trust_xff,
}
# 初始化配置（将在模块加载时执行）
_config = _get_anti_crawler_config()
ANTI_CRAWLER_MODE = _config['mode']
ALLOWED_IPS = _config['allowed_ips']
TRUSTED_PROXIES = _config['trusted_proxies']
TRUST_XFF = _config['trust_xff']
ANTI_CRAWLER_MODE = _config["mode"]
ALLOWED_IPS = _config["allowed_ips"]
TRUSTED_PROXIES = _config["trusted_proxies"]
TRUST_XFF = _config["trust_xff"]
def _get_mode_config(mode: str) -> dict:

50
src/webui/routers/knowledge.py
View File

@@ -17,36 +17,36 @@ _paragraph_store_cache = None
def _get_paragraph_store():
"""延迟加载段落 embedding store（只读模式，轻量级）
Returns:
EmbeddingStore | None: 如果配置启用则返回store，否则返回None
"""
# 检查配置是否启用
if not global_config.webui.enable_paragraph_content:
return None
global _paragraph_store_cache
if _paragraph_store_cache is not None:
return _paragraph_store_cache
try:
from src.chat.knowledge.embedding_store import EmbeddingStore
import os
# 获取数据路径
current_dir = os.path.dirname(os.path.abspath(__file__))
root_path = os.path.abspath(os.path.join(current_dir, "..", ".."))
embedding_dir = os.path.join(root_path, "data/embedding")
# 只加载段落 embedding store（轻量级）
paragraph_store = EmbeddingStore(
namespace="paragraph",
dir_path=embedding_dir,
max_workers=1, # 只读不需要多线程
chunk_size=100
chunk_size=100,
)
paragraph_store.load_from_file()
_paragraph_store_cache = paragraph_store
logger.info(f"成功加载段落 embedding store，包含 {len(paragraph_store.store)} 个段落")
return paragraph_store
@@ -57,10 +57,10 @@ def _get_paragraph_store():
def _get_paragraph_content(node_id: str) -> tuple[Optional[str], bool]:
"""从 embedding store 获取段落完整内容
Args:
node_id: 段落节点ID，格式为 'paragraph-{hash}'
Returns:
tuple[str | None, bool]: (段落完整内容或None, 是否启用了功能)
"""
@@ -69,12 +69,12 @@ def _get_paragraph_content(node_id: str) -> tuple[Optional[str], bool]:
if paragraph_store is None:
# 功能未启用
return None, False
# 从 store 中获取完整内容
paragraph_item = paragraph_store.store.get(node_id)
if paragraph_item is not None:
# paragraph_item 是 EmbeddingStoreItem，其 str 属性包含完整文本
content: str = getattr(paragraph_item, 'str', '')
content: str = getattr(paragraph_item, "str", "")
if content:
return content, True
return None, True
@@ -156,14 +156,18 @@ def _convert_graph_to_json(kg_manager) -> KnowledgeGraph:
node_data = graph[node_id]
# 节点类型: "ent" -> "entity", "pg" -> "paragraph"
node_type = "entity" if ("type" in node_data and node_data["type"] == "ent") else "paragraph"
# 对于段落节点，尝试从 embedding store 获取完整内容
if node_type == "paragraph":
full_content, _ = _get_paragraph_content(node_id)
content = full_content if full_content is not None else (node_data["content"] if "content" in node_data else node_id)
content = (
full_content
if full_content is not None
else (node_data["content"] if "content" in node_data else node_id)
)
else:
content = node_data["content"] if "content" in node_data else node_id
create_time = node_data["create_time"] if "create_time" in node_data else None
nodes.append(KnowledgeNode(id=node_id, type=node_type, content=content, create_time=create_time))
@@ -245,14 +249,18 @@ async def get_knowledge_graph(
try:
node_data = graph[node_id]
node_type_val = "entity" if ("type" in node_data and node_data["type"] == "ent") else "paragraph"
# 对于段落节点，尝试从 embedding store 获取完整内容
if node_type_val == "paragraph":
full_content, _ = _get_paragraph_content(node_id)
content = full_content if full_content is not None else (node_data["content"] if "content" in node_data else node_id)
content = (
full_content
if full_content is not None
else (node_data["content"] if "content" in node_data else node_id)
)
else:
content = node_data["content"] if "content" in node_data else node_id
create_time = node_data["create_time"] if "create_time" in node_data else None
nodes.append(KnowledgeNode(id=node_id, type=node_type_val, content=content, create_time=create_time))
@@ -368,11 +376,15 @@ async def search_knowledge_node(query: str = Query(..., min_length=1), _auth: bo
try:
node_data = graph[node_id]
node_type = "entity" if ("type" in node_data and node_data["type"] == "ent") else "paragraph"
# 对于段落节点，尝试从 embedding store 获取完整内容
if node_type == "paragraph":
full_content, _ = _get_paragraph_content(node_id)
content = full_content if full_content is not None else (node_data["content"] if "content" in node_data else node_id)
content = (
full_content
if full_content is not None
else (node_data["content"] if "content" in node_data else node_id)
)
else:
content = node_data["content"] if "content" in node_data else node_id

2
src/webui/services/git_mirror_service.py
View File

@@ -659,4 +659,4 @@ def get_git_mirror_service() -> GitMirrorService:
global _git_mirror_service
if _git_mirror_service is None:
_git_mirror_service = GitMirrorService()
return _git_mirror_service
return _git_mirror_service