fix:优化试题内容和样式排版

backend/app/main.py

@@ -1,5 +1,9 @@
 import os
+import json
 import uuid
+import zipfile
+import base64
+from typing import Any
 from datetime import date, datetime, timedelta
 from io import BytesIO
 from pathlib import Path
@@ -14,7 +18,7 @@ from reportlab.lib.pagesizes import A4
 from reportlab.pdfbase import pdfmetrics
 from reportlab.pdfbase.cidfonts import UnicodeCIDFont
 from reportlab.pdfgen import canvas
-from sqlalchemy import asc, desc, func, or_, select
+from sqlalchemy import asc, desc, func, inspect, or_, select, text
 from sqlalchemy.orm import Session
 
 from .database import Base, engine, get_db
@@ -27,6 +31,8 @@ from .schemas import (
     MistakeCreate,
     MistakeOut,
     MistakeUpdate,
+    OcrParseIn,
+    OcrParseOut,
     ResourceBatchUpdate,
     ResourceCreate,
     ResourceOut,
@@ -39,6 +45,23 @@ from .schemas import (
 
 Base.metadata.create_all(bind=engine)
 
+
+def _migrate_mistake_columns() -> None:
+    inspector = inspect(engine)
+    if "mistakes" not in inspector.get_table_names():
+        return
+    existed = {col["name"] for col in inspector.get_columns("mistakes")}
+    with engine.begin() as conn:
+        if "question_content" not in existed:
+            conn.execute(text("ALTER TABLE mistakes ADD COLUMN question_content TEXT"))
+        if "answer" not in existed:
+            conn.execute(text("ALTER TABLE mistakes ADD COLUMN answer TEXT"))
+        if "explanation" not in existed:
+            conn.execute(text("ALTER TABLE mistakes ADD COLUMN explanation TEXT"))
+
+
+_migrate_mistake_columns()
+
 app = FastAPI(title="公考助手 API", version="1.0.0")
 UPLOAD_DIR = Path(os.getenv("UPLOAD_DIR", "/app/uploads"))
 UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
@@ -69,6 +92,7 @@ def _query_mistakes_for_export(
     category: str | None,
     start_date: date | None,
     end_date: date | None,
+    ids: list[int] | None = None,
 ):
     stmt = select(Mistake)
     if category:
@@ -77,12 +101,175 @@ def _query_mistakes_for_export(
         stmt = stmt.where(Mistake.created_at >= datetime.combine(start_date, datetime.min.time()))
     if end_date:
         stmt = stmt.where(Mistake.created_at <= datetime.combine(end_date, datetime.max.time()))
+    if ids:
+        stmt = stmt.where(Mistake.id.in_(ids))
     items = db.scalars(stmt.order_by(desc(Mistake.created_at))).all()
     if len(items) > 200:
         raise HTTPException(status_code=400, detail="单次最多导出 200 题")
     return items
 
 
+def _validate_mistake_payload(payload: MistakeCreate | MistakeUpdate) -> None:
+    has_image = bool((payload.image_url or "").strip())
+    has_question = bool((payload.question_content or "").strip())
+    has_answer = bool((payload.answer or "").strip())
+    if not has_image and not has_question and not has_answer:
+        raise HTTPException(status_code=400, detail="请上传题目图片或填写试题/答案后再保存")
+
+
+def _normalize_multiline_text(value: str | None) -> str:
+    if not value:
+        return ""
+    text_value = value.replace("\r\n", "\n").replace("\r", "\n")
+    lines = [line.strip() for line in text_value.split("\n")]
+    compact = [line for line in lines if line]
+    return "\n".join(compact).strip()
+
+
+def _wrap_pdf_text(text: str, max_width: float, font_name: str = "STSong-Light", font_size: int = 12) -> list[str]:
+    normalized = _normalize_multiline_text(text)
+    if not normalized:
+        return []
+    wrapped: list[str] = []
+    for raw_line in normalized.split("\n"):
+        current = ""
+        for ch in raw_line:
+            candidate = f"{current}{ch}"
+            if pdfmetrics.stringWidth(candidate, font_name, font_size) <= max_width:
+                current = candidate
+            else:
+                if current:
+                    wrapped.append(current)
+                current = ch
+        if current:
+            wrapped.append(current)
+    return wrapped
+
+
+def _mistake_export_blocks(item: Mistake, content_mode: str) -> list[str]:
+    question = _normalize_multiline_text(item.question_content)
+    answer = _normalize_multiline_text(item.answer)
+    explanation = _normalize_multiline_text(item.explanation)
+    if not question:
+        question = "无题干与选项内容"
+
+    blocks: list[str] = [question]
+    if content_mode == "full":
+        # 「答案:」「解析:」与正文同一行开头，避免标签单独成行（与题号+题干规则一致）
+        blocks.append(f"答案: {answer or '无'}")
+        blocks.append(f"解析: {explanation or '无'}")
+    return blocks
+
+
+def _extract_upload_filename(url: str | None) -> str | None:
+    if not url or not url.startswith("/uploads/"):
+        return None
+    return Path(url).name
+
+
+def _safe_datetime(value: str | None) -> datetime:
+    if not value:
+        return datetime.utcnow()
+    try:
+        return datetime.fromisoformat(value.replace("Z", "+00:00")).replace(tzinfo=None)
+    except ValueError:
+        return datetime.utcnow()
+
+
+def _safe_date(value: str | None) -> date:
+    if not value:
+        return date.today()
+    try:
+        return date.fromisoformat(value)
+    except ValueError:
+        return date.today()
+
+
+def _extract_json_text(raw_text: str) -> str:
+    content = raw_text.strip()
+    if content.startswith("```"):
+        lines = content.splitlines()
+        if lines:
+            lines = lines[1:]
+        if lines and lines[-1].strip() == "```":
+            lines = lines[:-1]
+        content = "\n".join(lines).strip()
+    return content
+
+
+def _dump_all_data(db: Session) -> dict:
+    resources = db.scalars(select(Resource).order_by(asc(Resource.id))).all()
+    mistakes = db.scalars(select(Mistake).order_by(asc(Mistake.id))).all()
+    scores = db.scalars(select(ScoreRecord).order_by(asc(ScoreRecord.id))).all()
+    return {
+        "meta": {
+            "exported_at": datetime.utcnow().isoformat(),
+            "version": "1.1.0",
+        },
+        "resources": [
+            {
+                "id": item.id,
+                "title": item.title,
+                "resource_type": item.resource_type,
+                "url": item.url,
+                "file_name": item.file_name,
+                "category": item.category,
+                "tags": item.tags,
+                "created_at": item.created_at.isoformat() if item.created_at else None,
+            }
+            for item in resources
+        ],
+        "mistakes": [
+            {
+                "id": item.id,
+                "title": item.title,
+                "image_url": item.image_url,
+                "category": item.category,
+                "difficulty": item.difficulty,
+                "question_content": item.question_content,
+                "answer": item.answer,
+                "explanation": item.explanation,
+                "note": item.note,
+                "wrong_count": item.wrong_count,
+                "created_at": item.created_at.isoformat() if item.created_at else None,
+            }
+            for item in mistakes
+        ],
+        "scores": [
+            {
+                "id": item.id,
+                "exam_name": item.exam_name,
+                "exam_date": item.exam_date.isoformat() if item.exam_date else None,
+                "total_score": item.total_score,
+                "module_scores": item.module_scores,
+                "created_at": item.created_at.isoformat() if item.created_at else None,
+            }
+            for item in scores
+        ],
+    }
+
+
+def _restore_upload_url_from_zip(url: str | None, zip_ref: zipfile.ZipFile) -> str | None:
+    if not url:
+        return None
+    file_name = _extract_upload_filename(url)
+    if not file_name:
+        return url
+
+    zip_path = f"uploads/{file_name}"
+    if zip_path not in zip_ref.namelist():
+        return url
+
+    data = zip_ref.read(zip_path)
+    target_name = file_name
+    target_path = UPLOAD_DIR / target_name
+    if target_path.exists():
+        target_name = f"{uuid.uuid4().hex}_{file_name}"
+        target_path = UPLOAD_DIR / target_name
+    target_path.write_bytes(data)
+    return f"/uploads/{target_name}"
+
+
 @app.post("/api/upload")
 async def upload_file(file: UploadFile = File(...)):
     suffix = Path(file.filename or "").suffix.lower()
@@ -197,7 +384,16 @@ def list_mistakes(
     if category:
         stmt = stmt.where(Mistake.category == category)
     if keyword:
-        stmt = stmt.where(or_(Mistake.note.ilike(f"%{keyword}%"), Mistake.title.ilike(f"%{keyword}%")))
+        stmt = stmt.where(
+            or_(
+                Mistake.note.ilike(f"%{keyword}%"),
+                Mistake.title.ilike(f"%{keyword}%"),
+                Mistake.question_content.ilike(f"%{keyword}%"),
+                Mistake.answer.ilike(f"%{keyword}%"),
+                Mistake.explanation.ilike(f"%{keyword}%"),
+                Mistake.image_url.ilike(f"%{keyword}%"),
+            )
+        )
     sort_col = Mistake.created_at if sort_by == "created_at" else Mistake.wrong_count
     stmt = stmt.order_by(desc(sort_col) if order == "desc" else asc(sort_col))
     return db.scalars(stmt).all()
@@ -205,6 +401,7 @@ def list_mistakes(
 
 @app.post("/api/mistakes", response_model=MistakeOut)
 def create_mistake(payload: MistakeCreate, db: Session = Depends(get_db)):
+    _validate_mistake_payload(payload)
     item = Mistake(**payload.model_dump())
     db.add(item)
     db.commit()
@@ -214,6 +411,7 @@ def create_mistake(payload: MistakeCreate, db: Session = Depends(get_db)):
 
 @app.put("/api/mistakes/{item_id}", response_model=MistakeOut)
 def update_mistake(item_id: int, payload: MistakeUpdate, db: Session = Depends(get_db)):
+    _validate_mistake_payload(payload)
     item = db.get(Mistake, item_id)
     if not item:
         raise HTTPException(status_code=404, detail="Mistake not found")
@@ -239,32 +437,44 @@ def export_mistakes_pdf(
     category: str | None = None,
     start_date: date | None = None,
     end_date: date | None = None,
+    ids: str | None = None,
+    content_mode: str = Query("full", pattern="^(full|question_only)$"),
     db: Session = Depends(get_db),
 ):
-    items = _query_mistakes_for_export(db, category, start_date, end_date)
+    id_list = [int(x) for x in ids.split(",") if x.strip().isdigit()] if ids else None
+    items = _query_mistakes_for_export(db, category, start_date, end_date, id_list)
     buf = BytesIO()
     pdf = canvas.Canvas(buf, pagesize=A4)
     pdfmetrics.registerFont(UnicodeCIDFont("STSong-Light"))
     pdf.setFont("STSong-Light", 12)
 
     y = 800
-    pdf.drawString(50, y, "公考助手 - 错题导出")
-    y -= 30
+    left = 48
+    right = 560
+    max_width = right - left
+    pdf.drawString(left, y, "公考助手 - 错题导出")
+    y -= 28
     for idx, item in enumerate(items, start=1):
-        lines = [
-            f"{idx}. {item.title}",
-            f"分类: {item.category}  难度: {item.difficulty or '未设置'}  错误频次: {item.wrong_count}",
-            f"备注: {item.note or '无'}",
-            "答题区: _______________________________",
-        ]
-        for line in lines:
-            if y < 70:
-                pdf.showPage()
-                pdf.setFont("STSong-Light", 12)
-                y = 800
-            pdf.drawString(50, y, line[:90])
-            y -= 22
-        y -= 6
+        if y < 90:
+            pdf.showPage()
+            pdf.setFont("STSong-Light", 12)
+            y = 800
+        blocks = _mistake_export_blocks(item, content_mode)
+        for bi, block in enumerate(blocks):
+            # 题号与题干同一行开头，避免「1.」单独成行
+            text = f"{idx}. {block}" if bi == 0 else block
+            lines = _wrap_pdf_text(text, max_width=max_width)
+            if not lines:
+                continue
+            for line in lines:
+                if y < 70:
+                    pdf.showPage()
+                    pdf.setFont("STSong-Light", 12)
+                    y = 800
+                pdf.drawString(left, y, line)
+                y -= 18
+            y -= 6
+        y -= 8
 
     pdf.save()
     buf.seek(0)
@@ -280,16 +490,20 @@ def export_mistakes_docx(
     category: str | None = None,
     start_date: date | None = None,
     end_date: date | None = None,
+    ids: str | None = None,
+    content_mode: str = Query("full", pattern="^(full|question_only)$"),
     db: Session = Depends(get_db),
 ):
-    items = _query_mistakes_for_export(db, category, start_date, end_date)
+    id_list = [int(x) for x in ids.split(",") if x.strip().isdigit()] if ids else None
+    items = _query_mistakes_for_export(db, category, start_date, end_date, id_list)
     doc = Document()
     doc.add_heading("公考助手 - 错题导出", level=1)
     for idx, item in enumerate(items, start=1):
-        doc.add_paragraph(f"{idx}. {item.title}")
-        doc.add_paragraph(f"分类: {item.category} | 难度: {item.difficulty or '未设置'} | 错误频次: {item.wrong_count}")
-        doc.add_paragraph(f"备注: {item.note or '无'}")
-        doc.add_paragraph("答题区: ________________________________________")
+        blocks = _mistake_export_blocks(item, content_mode)
+        for bi, block in enumerate(blocks):
+            # 题号与题干同段，避免单独一行只有「1.」
+            para = f"{idx}. {block}" if bi == 0 else block
+            doc.add_paragraph(para)
 
     buf = BytesIO()
     doc.save(buf)
@@ -362,8 +576,108 @@ def score_stats(db: Session = Depends(get_db)):
     return ScoreStats(highest=highest, lowest=lowest, average=round(float(avg), 2), improvement=improvement)
 
 
+def _qwen_base_url() -> str:
+    return os.getenv("QWEN_BASE_URL", "https://dashscope.aliyuncs.com/compatible-mode/v1").strip().rstrip("/")
+
+
+def _get_qwen_api_key() -> str:
+    """去除首尾空白与常见误加的引号，避免 .env 里写成 'sk-xxx' 导致鉴权失败。"""
+    raw = os.getenv("QWEN_API_KEY", "") or ""
+    return raw.strip().strip('"').strip("'").strip()
+
+
+def _raise_for_qwen_http_error(resp: httpx.Response, prefix: str) -> None:
+    """HTTP 非 2xx 时解析 DashScope 错误体，对 invalid_api_key 返回 401 + 明确说明。"""
+    if resp.status_code < 300:
+        return
+    text = resp.text
+    try:
+        data = resp.json()
+        err = data.get("error")
+        if isinstance(err, dict):
+            code = str(err.get("code") or "")
+            if code == "invalid_api_key":
+                raise HTTPException(
+                    status_code=401,
+                    detail=(
+                        "阿里云 DashScope API Key 无效或未生效。"
+                        "请到阿里云百炼 / Model Studio 控制台创建 API Key（通常以 sk- 开头），"
+                        "写入项目根目录 .env 的 QWEN_API_KEY=，勿加引号；"
+                        "修改后执行: docker compose up -d --build backend"
+                    ),
+                )
+            msg = err.get("message") or text
+            raise HTTPException(status_code=502, detail=f"{prefix}: {msg}")
+    except HTTPException:
+        raise
+    except (ValueError, TypeError, KeyError):
+        pass
+    raise HTTPException(status_code=502, detail=f"{prefix}: {text[:1200]}")
+
+
+def _httpx_trust_env() -> bool:
+    """默认不信任环境变量中的代理，避免 Docker/IDE 注入空代理导致 ConnectError；需走系统代理时设 HTTPX_TRUST_ENV=1。"""
+    return os.getenv("HTTPX_TRUST_ENV", "0").lower() in ("1", "true", "yes")
+
+
+def _qwen_http_client(timeout_sec: float = 60.0) -> httpx.AsyncClient:
+    return httpx.AsyncClient(
+        timeout=httpx.Timeout(timeout_sec, connect=20.0),
+        trust_env=_httpx_trust_env(),
+        limits=httpx.Limits(max_keepalive_connections=5, max_connections=10),
+    )
+
+
+def _message_content_to_str(content: Any) -> str:
+    """OpenAI 兼容接口里 message.content 可能是 str 或多段结构。"""
+    if content is None:
+        return ""
+    if isinstance(content, str):
+        return content
+    if isinstance(content, list):
+        parts: list[str] = []
+        for part in content:
+            if isinstance(part, dict):
+                if part.get("type") == "text" and "text" in part:
+                    parts.append(str(part["text"]))
+                elif "text" in part:
+                    parts.append(str(part["text"]))
+            elif isinstance(part, str):
+                parts.append(part)
+        return "".join(parts)
+    return str(content)
+
+
+def _openai_completion_assistant_text(data: dict) -> str:
+    """从 chat/completions JSON 中取出助手文本；若含 error 或无 choices 则抛错。"""
+    err = data.get("error")
+    if err is not None:
+        if isinstance(err, dict):
+            code = str(err.get("code") or "")
+            if code == "invalid_api_key":
+                raise HTTPException(
+                    status_code=401,
+                    detail=(
+                        "阿里云 DashScope API Key 无效。"
+                        "请在 .env 中填写正确的 QWEN_API_KEY 并重启 backend。"
+                    ),
+                )
+            msg = err.get("message") or err.get("code") or json.dumps(err, ensure_ascii=False)
+        else:
+            msg = str(err)
+        raise HTTPException(status_code=502, detail=f"千问接口错误: {msg}")
+    choices = data.get("choices")
+    if not choices:
+        raise HTTPException(
+            status_code=502,
+            detail=f"千问返回异常（无 choices），请检查模型名与权限。原始片段: {json.dumps(data, ensure_ascii=False)[:800]}",
+        )
+    msg = choices[0].get("message") or {}
+    return _message_content_to_str(msg.get("content"))
+
+
 async def _call_qwen(system_prompt: str, user_prompt: str) -> str:
-    api_key = os.getenv("QWEN_API_KEY", "")
+    api_key = _get_qwen_api_key()
     if not api_key:
         return (
             "当前未配置千问 API Key，已返回本地降级提示。\n"
@@ -373,7 +687,7 @@ async def _call_qwen(system_prompt: str, user_prompt: str) -> str:
             "QWEN_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1\n"
             "QWEN_MODEL=qwen-plus"
         )
-    base_url = os.getenv("QWEN_BASE_URL", "https://dashscope.aliyuncs.com/compatible-mode/v1")
+    base_url = _qwen_base_url()
     model = os.getenv("QWEN_MODEL", "qwen-plus")
     headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"}
     payload = {
@@ -384,12 +698,333 @@ async def _call_qwen(system_prompt: str, user_prompt: str) -> str:
         ],
         "temperature": 0.4,
     }
-    async with httpx.AsyncClient(timeout=40) as client:
-        resp = await client.post(f"{base_url}/chat/completions", headers=headers, json=payload)
-    if resp.status_code >= 300:
-        raise HTTPException(status_code=502, detail=f"千问请求失败: {resp.text}")
-    data = resp.json()
-    return data["choices"][0]["message"]["content"]
+    url = f"{base_url}/chat/completions"
+    try:
+        async with _qwen_http_client(40.0) as client:
+            resp = await client.post(url, headers=headers, json=payload)
+    except httpx.ConnectError as e:
+        raise HTTPException(
+            status_code=502,
+            detail=(
+                f"无法连接千问接口（{url}）。请检查本机/容器能否访问外网、DNS 是否正常；"
+                "若在 Docker 中可尝试为 backend 配置 dns 或关闭错误代理。"
+                "默认已忽略 HTTP(S)_PROXY，若需代理请设置 HTTPX_TRUST_ENV=1。"
+                f" 原始错误: {e!s}"
+            ),
+        ) from e
+    except httpx.TimeoutException as e:
+        raise HTTPException(status_code=504, detail=f"千问请求超时: {e!s}") from e
+    _raise_for_qwen_http_error(resp, "千问请求失败")
+    try:
+        data = resp.json()
+    except ValueError:
+        raise HTTPException(status_code=502, detail=f"千问返回非 JSON: {resp.text[:600]}")
+    return _openai_completion_assistant_text(data)
+
+
+async def _call_qwen_vision(system_prompt: str, user_prompt: str, image_data_url: str) -> str:
+    api_key = _get_qwen_api_key()
+    if not api_key:
+        return (
+            "当前未配置千问 API Key，无法执行 OCR。\n"
+            "请在 .env 中配置 QWEN_API_KEY 后重试。"
+        )
+    base_url = _qwen_base_url()
+    model = os.getenv("QWEN_VL_MODEL", "qwen-vl-plus")
+    headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"}
+    # 与 DashScope 文档一致：先图后文，利于多模态路由
+    payload = {
+        "model": model,
+        "messages": [
+            {"role": "system", "content": system_prompt},
+            {
+                "role": "user",
+                "content": [
+                    {"type": "image_url", "image_url": {"url": image_data_url}},
+                    {"type": "text", "text": user_prompt},
+                ],
+            },
+        ],
+        "temperature": 0.2,
+    }
+    url = f"{base_url}/chat/completions"
+    try:
+        async with _qwen_http_client(60.0) as client:
+            resp = await client.post(url, headers=headers, json=payload)
+    except httpx.ConnectError as e:
+        raise HTTPException(
+            status_code=502,
+            detail=(
+                f"无法连接千问接口（OCR，{url}）。请检查网络与 DNS；"
+                "默认已忽略 HTTP(S)_PROXY，若需代理请设置 HTTPX_TRUST_ENV=1。"
+                f" 原始错误: {e!s}"
+            ),
+        ) from e
+    except httpx.TimeoutException as e:
+        raise HTTPException(status_code=504, detail=f"OCR 请求超时: {e!s}") from e
+    except httpx.RequestError as e:
+        raise HTTPException(status_code=502, detail=f"OCR 网络请求失败: {e!s}") from e
+    _raise_for_qwen_http_error(resp, "OCR 请求失败")
+    try:
+        data = resp.json()
+    except ValueError:
+        raise HTTPException(status_code=502, detail=f"OCR 返回非 JSON: {resp.text[:600]}")
+    return _openai_completion_assistant_text(data)
+
+
+@app.get("/api/data/export")
+def export_user_data(
+    format: str = Query("zip", pattern="^(zip|json)$"),
+    include_files: bool = True,
+    db: Session = Depends(get_db),
+):
+    payload = _dump_all_data(db)
+    timestamp = datetime.utcnow().strftime("%Y%m%d_%H%M%S")
+
+    if format == "json":
+        buf = BytesIO(json.dumps(payload, ensure_ascii=False, indent=2).encode("utf-8"))
+        return StreamingResponse(
+            buf,
+            media_type="application/json",
+            headers={"Content-Disposition": f'attachment; filename="exam_helper_backup_{timestamp}.json"'},
+        )
+
+    used_upload_files: set[str] = set()
+    for item in payload["resources"]:
+        name = _extract_upload_filename(item.get("url"))
+        if name:
+            used_upload_files.add(name)
+    for item in payload["mistakes"]:
+        name = _extract_upload_filename(item.get("image_url"))
+        if name:
+            used_upload_files.add(name)
+
+    buf = BytesIO()
+    with zipfile.ZipFile(buf, "w", zipfile.ZIP_DEFLATED) as zip_ref:
+        zip_ref.writestr("data.json", json.dumps(payload, ensure_ascii=False, indent=2))
+        if include_files:
+            for file_name in sorted(used_upload_files):
+                path = UPLOAD_DIR / file_name
+                if path.exists() and path.is_file():
+                    zip_ref.write(path, arcname=f"uploads/{file_name}")
+    buf.seek(0)
+    return StreamingResponse(
+        buf,
+        media_type="application/zip",
+        headers={"Content-Disposition": f'attachment; filename="exam_helper_backup_{timestamp}.zip"'},
+    )
+
+
+@app.post("/api/data/import")
+async def import_user_data(
+    file: UploadFile = File(...),
+    mode: str = Query("merge", pattern="^(merge|replace)$"),
+    db: Session = Depends(get_db),
+):
+    content = await file.read()
+    if not content:
+        raise HTTPException(status_code=400, detail="导入文件为空")
+    if len(content) > 100 * 1024 * 1024:
+        raise HTTPException(status_code=400, detail="导入文件不能超过 100MB")
+
+    suffix = Path(file.filename or "").suffix.lower()
+    payload: dict
+    zip_ref: zipfile.ZipFile | None = None
+
+    if suffix == ".json":
+        try:
+            payload = json.loads(content.decode("utf-8"))
+        except (UnicodeDecodeError, json.JSONDecodeError) as exc:
+            raise HTTPException(status_code=400, detail=f"JSON 解析失败: {exc}") from exc
+    elif suffix == ".zip":
+        try:
+            zip_ref = zipfile.ZipFile(BytesIO(content))
+        except zipfile.BadZipFile as exc:
+            raise HTTPException(status_code=400, detail="ZIP 文件损坏或格式错误") from exc
+        if "data.json" not in zip_ref.namelist():
+            raise HTTPException(status_code=400, detail="ZIP 中缺少 data.json")
+        try:
+            payload = json.loads(zip_ref.read("data.json").decode("utf-8"))
+        except (UnicodeDecodeError, json.JSONDecodeError) as exc:
+            raise HTTPException(status_code=400, detail=f"data.json 解析失败: {exc}") from exc
+    else:
+        raise HTTPException(status_code=400, detail="仅支持 .json 或 .zip 导入")
+
+    resources = payload.get("resources", [])
+    mistakes = payload.get("mistakes", [])
+    scores = payload.get("scores", [])
+    if not isinstance(resources, list) or not isinstance(mistakes, list) or not isinstance(scores, list):
+        raise HTTPException(status_code=400, detail="导入文件结构错误")
+
+    if mode == "replace":
+        for item in db.scalars(select(Resource)).all():
+            db.delete(item)
+        for item in db.scalars(select(Mistake)).all():
+            db.delete(item)
+        for item in db.scalars(select(ScoreRecord)).all():
+            db.delete(item)
+        db.commit()
+
+    imported = {"resources": 0, "mistakes": 0, "scores": 0}
+
+    for item in resources:
+        url = item.get("url")
+        if zip_ref is not None:
+            url = _restore_upload_url_from_zip(url, zip_ref)
+        obj = Resource(
+            title=item.get("title") or "未命名资源",
+            resource_type=item.get("resource_type") if item.get("resource_type") in {"link", "file"} else "link",
+            url=url,
+            file_name=item.get("file_name"),
+            category=item.get("category") or "未分类",
+            tags=item.get("tags"),
+            created_at=_safe_datetime(item.get("created_at")),
+        )
+        db.add(obj)
+        imported["resources"] += 1
+
+    for item in mistakes:
+        image_url = item.get("image_url")
+        if zip_ref is not None:
+            image_url = _restore_upload_url_from_zip(image_url, zip_ref)
+        difficulty = item.get("difficulty")
+        obj = Mistake(
+            title=item.get("title") or "未命名错题",
+            image_url=image_url,
+            category=item.get("category") or "其他",
+            difficulty=difficulty if difficulty in {"easy", "medium", "hard"} else None,
+            question_content=item.get("question_content"),
+            answer=item.get("answer"),
+            explanation=item.get("explanation"),
+            note=item.get("note"),
+            wrong_count=max(int(item.get("wrong_count") or 1), 1),
+            created_at=_safe_datetime(item.get("created_at")),
+        )
+        db.add(obj)
+        imported["mistakes"] += 1
+
+    for item in scores:
+        score = float(item.get("total_score") or 0)
+        obj = ScoreRecord(
+            exam_name=item.get("exam_name") or "未命名考试",
+            exam_date=_safe_date(item.get("exam_date")),
+            total_score=max(min(score, 200), 0),
+            module_scores=item.get("module_scores"),
+            created_at=_safe_datetime(item.get("created_at")),
+        )
+        db.add(obj)
+        imported["scores"] += 1
+
+    db.commit()
+    return {"success": True, "mode": mode, "imported": imported}
+
+
+@app.post("/api/ocr/parse", response_model=OcrParseOut)
+async def parse_ocr(payload: OcrParseIn):
+    file_name = _extract_upload_filename(payload.image_url)
+    if not file_name:
+        raise HTTPException(status_code=400, detail="仅支持 /uploads 下的图片做 OCR")
+    target = UPLOAD_DIR / file_name
+    if not target.exists() or not target.is_file():
+        raise HTTPException(status_code=404, detail="图片不存在或已删除")
+
+    suffix = target.suffix.lower()
+    mime = {
+        ".jpg": "image/jpeg",
+        ".jpeg": "image/jpeg",
+        ".png": "image/png",
+        ".webp": "image/webp",
+    }.get(suffix)
+    if not mime:
+        raise HTTPException(status_code=400, detail="仅支持 JPG/PNG/WebP OCR")
+
+    b64 = base64.b64encode(target.read_bytes()).decode("utf-8")
+    image_data_url = f"data:{mime};base64,{b64}"
+    ocr_prompt = (
+        "请识别图片中的题目，返回严格 JSON。"
+        "字段说明：text 为整题完整纯文本（含材料、提问句、全部选项）；"
+        "question_content 必须与 text 一致地表示「完整题干」，须包含阅读材料、填空/提问句、所有选项（A B C D 等），"
+        "禁止只填写「依次填入…」等短提示句而省略材料和选项。"
+        "另含 title_suggestion、category_suggestion、difficulty_suggestion、answer、explanation。"
+        "无法确认的字段可填空字符串。"
+    )
+    if payload.prompt:
+        ocr_prompt = f"{ocr_prompt}\n补充要求：{payload.prompt}"
+
+    raw_text = await _call_qwen_vision(
+        "你是公考题目OCR与结构化助手。输出必须是 JSON，不要额外解释。",
+        ocr_prompt,
+        image_data_url,
+    )
+
+    try:
+        parsed = json.loads(_extract_json_text(raw_text))
+        data = (
+            parsed
+            if isinstance(parsed, dict)
+            else {
+                "text": raw_text.strip(),
+                "title_suggestion": None,
+                "category_suggestion": None,
+                "difficulty_suggestion": None,
+                "question_content": raw_text.strip(),
+                "answer": "",
+                "explanation": "",
+            }
+        )
+    except json.JSONDecodeError:
+        data = {
+            "text": raw_text.strip(),
+            "title_suggestion": None,
+            "category_suggestion": None,
+            "difficulty_suggestion": None,
+            "question_content": raw_text.strip(),
+            "answer": "",
+            "explanation": "",
+        }
+
+    def _opt_str(val: Any) -> str | None:
+        if val is None:
+            return None
+        if isinstance(val, (dict, list)):
+            return None
+        s = str(val).strip()
+        return s if s else None
+
+    def _merge_question_body(text_raw: str, qc_raw: str | None) -> str | None:
+        """模型常把全文放在 text，却只把短问句放在 question_content；合并时以更长、更完整的文本为准。"""
+        t = (text_raw or "").strip()
+        q = (qc_raw or "").strip()
+        if not t and not q:
+            return None
+        if not q:
+            return t or None
+        if not t:
+            return q or None
+        if len(t) > len(q):
+            return t
+        if len(q) > len(t):
+            return q
+        # 长度接近或相等：若一方包含另一方，取更长；否则保留 text（整页 OCR 通常更全）
+        if t in q:
+            return q
+        if q in t:
+            return t
+        return t
+
+    text_out = str(data.get("text", "") or "").strip()
+    qc_model = _opt_str(data.get("question_content"))
+    question_merged = _merge_question_body(text_out, qc_model)
+
+    return OcrParseOut(
+        text=text_out,
+        title_suggestion=_opt_str(data.get("title_suggestion")),
+        category_suggestion=_opt_str(data.get("category_suggestion")),
+        difficulty_suggestion=_opt_str(data.get("difficulty_suggestion")),
+        question_content=_opt_str(question_merged),
+        answer=_opt_str(data.get("answer")),
+        explanation=_opt_str(data.get("explanation")),
+    )
 
 
 @app.post("/api/ai/mistakes/{item_id}/analyze", response_model=AiMistakeAnalysisOut)
@@ -403,6 +1038,9 @@ async def ai_analyze_mistake(item_id: int, db: Session = Depends(get_db)):
             f"错题标题: {item.title}\n"
             f"分类: {item.category}\n"
             f"难度: {item.difficulty or '未设置'}\n"
+            f"题目内容: {item.question_content or '无'}\n"
+            f"答案: {item.answer or '无'}\n"
+            f"解析: {item.explanation or '无'}\n"
             f"错误频次: {item.wrong_count}\n"
             f"备注: {item.note or '无'}\n\n"
             "请按以下结构输出:\n"