增加代码知识库；修复文档处理内容；增加API设置

RAG-TEST-TOOLS/config.py

@@ -0,0 +1,97 @@
+# config.py
+import os
+import sys
+from pathlib import Path
+import json
+
+def get_app_base_path():
+    """
+    获取应用程序基础路径，兼容开发和打包环境
+    """
+    if getattr(sys, 'frozen', False):
+        # 如果是打包后的exe文件
+        if hasattr(sys, '_MEIPASS'):
+            # 临时解压目录（包含打包的资源文件）
+            return Path(sys._MEIPASS)
+        else:
+            # exe文件所在目录
+            return Path(sys.executable).parent
+    else:
+        # 开发环境：当前脚本所在目录
+        return Path(__file__).parent
+
+# 获取基础路径
+BASE_PATH = get_app_base_path()
+
+# API 配置 - 硬编码
+QWEN_API_URL = os.getenv(
+    "QWEN_API_URL",
+    os.getenv("DASH_SCOPE_API_BASE", "https://dashscope.aliyuncs.com/compatible-mode/v1"),
+)
+QWEN_API_KEY = (
+    os.getenv("DASHSCOPE_API_KEY")
+    or os.getenv("DASH_SCOPE_API_KEY")
+    or os.getenv("QWEN_API_KEY")
+    or ""
+)
+QWEN_CHAT_MODEL = os.getenv("QWEN_CHAT_MODEL", "qwen-max")
+QWEN_EMBEDDING_MODEL = os.getenv("QWEN_EMBEDDING_MODEL", "text-embedding-v4")
+
+# 项目路径配置 - 硬编码
+PROJECT_ROOT = BASE_PATH
+Default_PROJECT_ROOT = BASE_PATH / "PrjAttCtrlMng"
+
+# 数据文件路径配置 - 硬编码
+VECTOR_DB_PATH = ""
+Default_VECTOR_DB_PATH = str(BASE_PATH / "satellite_rag.faiss")
+METADATA_PATH = ""
+Default_METADATA_PATH = str(BASE_PATH / "satellite_rag_metadata.json")
+GRAPH_DATA_PATH = ""
+Default_GRAPH_DATA_PATH = str(BASE_PATH / "code_knowledge_graph.json")
+
+# 代码处理配置 - 硬编码
+MAX_CODE_LENGTH = 800
+CPP_EXTENSIONS = {'.c', '.cpp', '.h', '.hpp', '.tcc'}
+IGNORE_DIRS = {'build', 'cmake-build', '.git', 'vendor', 'lib', 'external', 'Debug'}
+
+# 功能检索配置 - 硬编码
+KNOWLEDGE_GRAPH_PATH = ""
+Default_KNOWLEDGE_GRAPH_PATH = str(BASE_PATH / "project_code_knowledge_graph.json")
+TOP_K = 8
+MAX_HOPS = 3
+MIN_SIMILARITY_THRESHOLD = 0.6
+
+# 辅助函数，用于实际使用时的动态路径获取
+def get_vector_db_path():
+    """获取向量数据库实际路径"""
+    return VECTOR_DB_PATH or Default_VECTOR_DB_PATH
+
+def get_metadata_path():
+    """获取元数据文件实际路径"""
+    return METADATA_PATH or Default_METADATA_PATH
+
+def get_graph_data_path():
+    """获取图谱数据实际路径"""
+    return GRAPH_DATA_PATH or Default_GRAPH_DATA_PATH
+
+def get_knowledge_graph_path():
+    """获取知识图谱数据实际路径"""
+    return KNOWLEDGE_GRAPH_PATH or Default_KNOWLEDGE_GRAPH_PATH
+
+# 导出路径信息，便于调试
+def print_path_info():
+    """打印路径信息，用于调试"""
+    print("=== 配置文件路径信息 ===")
+    print(f"BASE_PATH: {BASE_PATH}")
+    print(f"PROJECT_ROOT: {PROJECT_ROOT}")
+    print(f"Default_PROJECT_ROOT: {Default_PROJECT_ROOT}")
+    print(f"Default_VECTOR_DB_PATH: {Default_VECTOR_DB_PATH}")
+    print(f"Default_METADATA_PATH: {Default_METADATA_PATH}")
+    print(f"Default_GRAPH_DATA_PATH: {Default_GRAPH_DATA_PATH}")
+    print(f"Default_KNOWLEDGE_GRAPH_PATH: {Default_KNOWLEDGE_GRAPH_PATH}")
+    print(f"是否打包环境: {getattr(sys, 'frozen', False)}")
+    print("=======================")
+
+# 如果直接运行此文件，打印路径信息
+if __name__ == "__main__":
+    print_path_info()