feat/ 导出开发容器内的数据用于评估

api/app/features/evaluation/importers/script_json.py

@@ -0,0 +1,23 @@
+"""JSON 脚本导入评测用例。"""
+
+from __future__ import annotations
+
+import json
+from typing import Any
+
+
+def parse_script_json(raw: str | bytes) -> tuple[list[str], dict[str, Any]]:
+    data = json.loads(raw if isinstance(raw, str) else raw.decode("utf-8"))
+    if isinstance(data, list):
+        utterances = [str(x).strip() for x in data if str(x).strip()]
+        return utterances, {}
+    if isinstance(data, dict):
+        u = data.get("utterances") or data.get("user_utterances") or []
+        if not isinstance(u, list):
+            raise ValueError("utterances 必须是数组")
+        utterances = [str(x).strip() for x in u if str(x).strip()]
+        meta = {
+            k: v for k, v in data.items() if k not in ("utterances", "user_utterances")
+        }
+        return utterances, meta
+    raise ValueError("根须为数组或对象")

api/app/features/evaluation/importers/user_export_markdown.py

@@ -0,0 +1,19 @@
+"""从 extract_sql_to_user_md 产出的 Markdown 中提取用户轮次。"""
+
+from __future__ import annotations
+
+import re
+
+
+def extract_user_utterances_from_export_md(text: str) -> list[str]:
+    """匹配 ``**用户:**`` 块之间的正文。"""
+    out: list[str] = []
+    for m in re.finditer(
+        r"\*\*用户:\*\*\s*\n+(.+?)(?=\n+\*\*AI:\*\*|\n+####|\Z)",
+        text,
+        flags=re.DOTALL | re.IGNORECASE,
+    ):
+        chunk = (m.group(1) or "").strip()
+        if chunk and chunk != "（空）":
+            out.append(chunk)
+    return out