feat/ eval

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

@@ -49,3 +49,47 @@ def extract_dialogue_turns_from_export_md(text: str) -> list[tuple[str, str]]:
         raw_ai = ((ai_m.group(1) if ai_m else "") or "").strip()
         out.append((u, _normalize_export_ai_block(raw_ai)))
     return out
+
+
+_MEMOIR_SECTION_HEADER = re.compile(
+    r"^##\s*回忆录章节（生成正文）\s*$",
+    re.MULTILINE | re.IGNORECASE,
+)
+
+_IMAGE_REF = re.compile(r"\{\{IMAGE:[^}]*\}\}\s*", re.DOTALL)
+
+
+def extract_source_user_id_from_export_md(text: str) -> str | None:
+    """匹配导出头 ``**User ID:** `uuid` ``。"""
+    m = re.search(r"\*\*User ID:\*\*\s*`([0-9a-fA-F-]{36})`", text)
+    if not m:
+        return None
+    return m.group(1).strip()
+
+
+def extract_memoir_chapter_sections_from_export_md(text: str) -> list[tuple[str, str]]:
+    """从 ``## 回忆录章节（生成正文）`` 起按 ``##`` / ``###`` 标题切分基线正文（去掉 IMAGE 占位）。"""
+    m = _MEMOIR_SECTION_HEADER.search(text)
+    if not m:
+        return []
+    tail = (text[m.end() :] or "").strip()
+    if not tail:
+        return []
+    pieces = re.split(r"\n(?=(?:###\s|##\s+))", tail)
+    out: list[tuple[str, str]] = []
+    for piece in pieces:
+        piece = piece.strip()
+        if not piece.startswith("#"):
+            continue
+        first_nl = piece.find("\n")
+        if first_nl == -1:
+            title = piece.lstrip("#").strip()
+            body = ""
+        else:
+            title = piece[:first_nl].lstrip("#").strip()
+            body = (piece[first_nl + 1 :] or "").strip()
+        body = _IMAGE_REF.sub("", body)
+        body = re.sub(r"\n{3,}", "\n\n", body).strip()
+        if title and body:
+            out.append((title, body))
+    return out