fix: 修复 NameError/状态污染/类型标注/统计; 补全练习与 main; 新增 config/.gitignore/requirements; 文档统一

config.py

@@ -0,0 +1,87 @@
+"""
+集中读取 .env / 环境变量。
+
+使用方式：
+    from config import settings
+    print(settings.llm_model)
+    print(settings.has_openai_key)
+"""
+
+import os
+from dataclasses import dataclass
+from pathlib import Path
+
+
+def _load_dotenv(env_path: Path) -> None:
+    """极简 .env 解析，避免引入 python-dotenv 依赖。"""
+    if not env_path.is_file():
+        return
+    for raw_line in env_path.read_text(encoding="utf-8").splitlines():
+        line = raw_line.strip()
+        if not line or line.startswith("#"):
+            continue
+        if "=" not in line:
+            continue
+        key, _, value = line.partition("=")
+        key = key.strip()
+        value = value.strip().strip('"').strip("'")
+        # 已存在则不覆盖（让真实环境变量优先）
+        os.environ.setdefault(key, value)
+
+
+_ROOT = Path(__file__).resolve().parent
+_load_dotenv(_ROOT / ".env")
+
+
+@dataclass(frozen=True)
+class Settings:
+    openai_api_key: str
+    anthropic_api_key: str
+    llm_provider: str
+    llm_model: str
+    llm_max_tokens: int
+    rag_chroma_path: str
+    rag_collection_name: str
+    rag_embed_model: str
+    validation_service_url: str
+    log_level: str
+
+    @property
+    def has_openai_key(self) -> bool:
+        return bool(self.openai_api_key) and self.openai_api_key != "your_openai_api_key_here"
+
+    @property
+    def has_anthropic_key(self) -> bool:
+        return bool(self.anthropic_api_key) and self.anthropic_api_key != "your_anthropic_api_key_here"
+
+
+def _int(name: str, default: int) -> int:
+    raw = os.environ.get(name)
+    try:
+        return int(raw) if raw else default
+    except ValueError:
+        return default
+
+
+settings = Settings(
+    openai_api_key=os.environ.get("OPENAI_API_KEY", ""),
+    anthropic_api_key=os.environ.get("ANTHROPIC_API_KEY", ""),
+    llm_provider=os.environ.get("LLM_PROVIDER", "openai"),
+    llm_model=os.environ.get("LLM_MODEL", "gpt-4o-mini"),
+    llm_max_tokens=_int("LLM_MAX_TOKENS", 4096),
+    rag_chroma_path=os.environ.get("RAG_CHROMA_PATH", "./db/chroma"),
+    rag_collection_name=os.environ.get("RAG_COLLECTION_NAME", "jrxml_chunks"),
+    rag_embed_model=os.environ.get("RAG_EMBED_MODEL", "sentence-transformers/all-MiniLM-L6-v2"),
+    validation_service_url=os.environ.get("VALIDATION_SERVICE_URL", "http://localhost:8001"),
+    log_level=os.environ.get("LOG_LEVEL", "INFO"),
+)
+
+
+if __name__ == "__main__":
+    s = settings
+    print(f"provider  = {s.llm_provider}")
+    print(f"model     = {s.llm_model}")
+    print(f"max_tok   = {s.llm_max_tokens}")
+    print(f"openai?   = {s.has_openai_key}")
+    print(f"anthro?   = {s.has_anthropic_key}")
+    print(f"rag_path  = {s.rag_chroma_path}")