gardel/Mofox-Core
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

style: 格式化代码

Browse Source
This commit is contained in:
John Richard
2025-10-02 19:38:39 +08:00
parent d5627b0661
commit ecb02cae31
111 changed files with 2344 additions and 2316 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
scripts/lpmm_learning_tool.py
View File

@@ -40,6 +40,7 @@ file_lock = Lock()
# --- 缓存清理 ---
def clear_cache():
"""清理 lpmm_learning_tool.py 生成的缓存文件"""
logger.info("--- 开始清理缓存 ---")
@@ -53,6 +54,7 @@ def clear_cache():
logger.info("缓存目录不存在，无需清理。")
logger.info("--- 缓存清理完成 ---")
# --- 模块一：数据预处理 ---
@@ -108,7 +110,7 @@ def _parse_and_repair_json(json_string: str) -> Optional[dict]:
cleaned_string = cleaned_string[7:].strip()
elif cleaned_string.startswith("```"):
cleaned_string = cleaned_string[3:].strip()
if cleaned_string.endswith("```"):
cleaned_string = cleaned_string[:-3].strip()
@@ -117,7 +119,7 @@ def _parse_and_repair_json(json_string: str) -> Optional[dict]:
return orjson.loads(cleaned_string)
except orjson.JSONDecodeError:
logger.warning("直接解析JSON失败，将尝试修复...")
# 3. 修复与最终解析
repaired_json_str = ""
try:
@@ -164,10 +166,10 @@ async def extract_info_async(pg_hash, paragraph, llm_api):
content = None
try:
content, (_, _, _) = await llm_api.generate_response_async(prompt)
# 改进点：调用封装好的函数处理JSON解析和修复
extracted_data = _parse_and_repair_json(content)
if extracted_data is None:
# 如果解析失败，抛出异常以触发统一的错误处理逻辑
raise ValueError("无法从LLM输出中解析有效的JSON数据")

64
scripts/rebuild_metadata_index.py
View File

@@ -3,6 +3,7 @@
"""
从现有ChromaDB数据重建JSON元数据索引
"""
import asyncio
import sys
import os
@@ -15,53 +16,53 @@ from src.common.logger import get_logger
logger = get_logger(__name__)
async def rebuild_metadata_index():
"""从ChromaDB重建元数据索引"""
print("="*80)
print("=" * 80)
print("重建JSON元数据索引")
print("="*80)
print("=" * 80)
# 初始化记忆系统
print("\n🔧 初始化记忆系统...")
ms = MemorySystem()
await ms.initialize()
print("✅ 记忆系统已初始化")
if not hasattr(ms.unified_storage, 'metadata_index'):
if not hasattr(ms.unified_storage, "metadata_index"):
print("❌ 元数据索引管理器未初始化")
return
# 获取所有记忆
print("\n📥 从ChromaDB获取所有记忆...")
from src.common.vector_db import vector_db_service
try:
# 获取集合中的所有记忆ID
collection_name = ms.unified_storage.config.memory_collection
result = vector_db_service.get(
collection_name=collection_name,
include=["documents", "metadatas", "embeddings"]
collection_name=collection_name, include=["documents", "metadatas", "embeddings"]
)
if not result or not result.get("ids"):
print("❌ ChromaDB中没有找到记忆数据")
return
ids = result["ids"]
metadatas = result.get("metadatas", [])
print(f"✅ 找到 {len(ids)} 条记忆")
# 重建元数据索引
print("\n🔨 开始重建元数据索引...")
entries = []
success_count = 0
for i, (memory_id, metadata) in enumerate(zip(ids, metadatas), 1):
for i, (memory_id, metadata) in enumerate(zip(ids, metadatas, strict=False), 1):
try:
# 从ChromaDB元数据重建索引条目
import orjson
entry = MemoryMetadataIndexEntry(
memory_id=memory_id,
user_id=metadata.get("user_id", "unknown"),
@@ -75,9 +76,9 @@ async def rebuild_metadata_index():
created_at=metadata.get("created_at", 0.0),
access_count=metadata.get("access_count", 0),
chat_id=metadata.get("chat_id"),
content_preview=None
content_preview=None,
)
# 尝试解析importance和confidence的枚举名称
if "importance" in metadata:
imp_str = metadata["importance"]
@@ -89,7 +90,7 @@ async def rebuild_metadata_index():
entry.importance = 3
elif imp_str == "CRITICAL":
entry.importance = 4
if "confidence" in metadata:
conf_str = metadata["confidence"]
if conf_str == "LOW":
@@ -100,40 +101,41 @@ async def rebuild_metadata_index():
entry.confidence = 3
elif conf_str == "VERIFIED":
entry.confidence = 4
entries.append(entry)
success_count += 1
if i % 100 == 0:
print(f" 处理进度: {i}/{len(ids)} ({success_count} 成功)")
except Exception as e:
logger.warning(f"处理记忆 {memory_id} 失败: {e}")
continue
print(f"\n✅ 成功解析 {success_count}/{len(ids)} 条记忆元数据")
# 批量更新索引
print("\n💾 保存元数据索引...")
ms.unified_storage.metadata_index.batch_add_or_update(entries)
ms.unified_storage.metadata_index.save()
# 显示统计信息
stats = ms.unified_storage.metadata_index.get_stats()
print(f"\n📊 重建后的索引统计:")
print("\n📊 重建后的索引统计:")
print(f" - 总记忆数: {stats['total_memories']}")
print(f" - 主语数量: {stats['subjects_count']}")
print(f" - 关键词数量: {stats['keywords_count']}")
print(f" - 标签数量: {stats['tags_count']}")
print(f" - 类型分布:")
for mtype, count in stats['types'].items():
print(" - 类型分布:")
for mtype, count in stats["types"].items():
print(f" - {mtype}: {count}")
print("\n✅ 元数据索引重建完成！")
except Exception as e:
logger.error(f"重建索引失败: {e}", exc_info=True)
print(f"❌ 重建索引失败: {e}")
if __name__ == '__main__':
if __name__ == "__main__":
asyncio.run(rebuild_metadata_index())

5
scripts/run_multi_stage_smoke.py
View File

@@ -3,6 +3,7 @@
"""
轻量烟雾测试：初始化 MemorySystem 并运行一次检索，验证 MemoryMetadata.source 访问不再报错
"""
import asyncio
import sys
import os
@@ -11,6 +12,7 @@ sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from src.chat.memory_system.memory_system import MemorySystem
async def main():
ms = MemorySystem()
await ms.initialize()
@@ -19,5 +21,6 @@ async def main():
for i, m in enumerate(results, 1):
print(f"{i}. id={m.metadata.memory_id} source={getattr(m.metadata, 'source', None)}")
if __name__ == '__main__':
if __name__ == "__main__":
asyncio.run(main())