@miravia_test/miravia-publish-skill 0.1.2 → 0.1.3

package/dist/SKILL.md

@@ -24,9 +24,9 @@ description: Miravia 商品发布 AI Skill。仅负责商品创建（手动一
 
 | 文件 | 职责 |
 |------|------|
-| `scripts/modules/publish_module.py` | 发布流程主控（类目/品牌/属性/图片/SKU/合规/确认/创建） |
-| `scripts/modules/extraction_engine.py` | 从源文本抽取 `MiraviaListingSchema` |
-| `scripts/modules/file_reader.py` | 读取本地文件、文件夹或网页 URL |
+| `scripts/publish_module.py` | 发布流程主控（类目/品牌/属性/图片/SKU/合规/确认/创建） |
+| `scripts/extraction_engine.py` | 从源文本抽取 `MiraviaListingSchema` |
+| `scripts/file_reader.py` | 读取本地文件、文件夹或网页 URL |
 | `scripts/field_definitions.py` | 商品字段定义与术语说明，供 LLM 抽取与提示使用 |
 | `scripts/product_client.py` | Miravia 商品/图片/类目/品牌/合规 API 封装 |
 | `scripts/miravia_client.py` | 开放平台请求签名、重试、埋点 |
@@ -99,7 +99,8 @@ description: Miravia 商品发布 AI Skill。仅负责商品创建（手动一
 | 状态 | 含义 | 处理方式 |
 |------|------|----------|
 | `need_authorization` | 需要卖家授权三项操作权限 | 卖家确认后加 `--authorized` 重新运行 |
-| `need_category` | 类目未找到或存在多个候选 | 让用户提供正确的类目 ID（`--category-id`）或从候选中选择一个（`--confirmed-category-id`），**禁止重复搜索同一关键词** |
+| `authorization_cancelled` | 卖家明确拒绝授权 | 提示需授权，重新上传文件将再次触发授权请求 |
+| `need_category` | 类目未找到、类目 ID 无效、推荐多候选或预测失败 | 让用户提供正确的类目 ID（`--category-id`）或从候选中选择一个（`--confirmed-category-id`），**禁止重复搜索同一关键词**；所有 ID 均校验类目树；关键词搜索失败后调 `/product/category/recommend/get` 推荐，Top1 自动选中，TopN 展示候选，空结果展示全量类目树 |
 | `need_brand` | 品牌未找到或存在多个候选 | 让用户提供正确的品牌名（`--brand-name`）或从候选中选择一个（`--confirmed-brand-id`），**禁止重复搜索同一关键词** |
 | `need_ctb_confirm` | 需要确认类目/标题/品牌 | 回传 `ctb_confirm_token` |
 | `need_brand_create` | 品牌未找到，需去 Seller Center 创建 | 创建后重新运行命令 |
@@ -107,7 +108,20 @@ description: Miravia 商品发布 AI Skill。仅负责商品创建（手动一
 | `need_category_attributes` | 类目必填属性缺失（AI抽取模式） | 通过 `--attribute-data` 补充或使用 `--draft` |
 | `need_field_summary` | 需要确认字段摘要或选择草稿模式 | 回传 `field_summary_token`，可加 `--draft` |
 | `need_compliance` | 制造商/EU 负责人匹配失败 | 按返回候选选择或补充信息后重跑 |
+| `compliance_missing_suggest_draft` | 合规字段缺失（制造商/欧代/Safety Warning内容/合规文档） | 补充合规字段或加 `--draft` 发布为草稿 |
+| `gpsr_missing_suggest_draft` | GPSR 字段缺失（制造商和/或欧盟负责人为空） | 补充 GPSR 字段或加 `--draft` 发布为草稿 |
+| `post_to_inactive_missing_suggest_draft` | 类目必填字段缺失（EAN） | 补充字段或加 `--draft` 发布为草稿 |
+| `special_category_block` | 特殊类目规则阻断（PPU 值无效 / 图书类目 ISBN 缺失） | 修正错误后重跑，不可转 Draft |
+| `special_category_suggest_draft` | 特殊类目字段缺失（PPU 缺失 / 食品属性缺失） | 补充字段或加 `--draft` 发布为草稿 |
+| `dbm_limit_exceeded_no_dbs` | DBM 超限但无 DBS 权限（重量>40KG / 三边和>300cm / 体积重>40） | 减小包裹尺寸/重量或申请 DBS 权限 |
+| `sku_already_exists` | 发布时 SKU 已被占用（并发冲突或 SKU 重复） | 刷新商品列表检查或修改 Seller SKU 后重跑 |
+| `multiple_products_detected` | Excel 中包含多个商品（>1 行） | 展示商品列表，用户选择序号/all/cancel |
 | `need_images` | 部分图片迁移失败 | 替换图片后重跑 |
+| `unsupported_format` | 上传了不支持的文件格式（zip/txt/psd 等） | 提示仅支持 PDF/Word/Excel/Image/URL，允许重新上传 |
+| `file_empty_or_corrupted` | 文件内容为空或损坏（0KB PDF / 乱码 Word） | 提示文件内容为空或损坏，允许重新上传 |
+| `invalid_url` | 输入的 URL 无法抓取 | 提示检查 URL 后重试 |
+| `title_invalid` | 商品标题不符合规范（过短/过长/词数不足） | 提示具体错误 + 标题格式建议（Brand+Category+Model+Features+Scenario），阻断后续流程，禁止转 Draft |
+| `need_category_permission` | 卖家无该类目发布权限 | 询问是否以 Draft 模式发布（待审批通过后自动激活），提供类目权限申请链接 https://sellercenter.miravia.es/apps/setting/category；或更换类目后重跑 |
 | `preview` | 最终发布预览 | 回传 `publish_preview_token` |
 | `success` | 创建成功 | 返回 `item_id`、`pdp_url`、审核提示 |
 | `brand_permission_failed` | 品牌权限不足/待审批 (BRAND_PENDING / BRAND_PERMISSION_DENIED) | 加 `--draft` 发布为草稿，或在 Seller Center 申请品牌权限后重跑 |
@@ -178,6 +192,7 @@ python scripts/product_skill.py --agent-type qoder --scenario create_product \
 | `--sku-data` | 否 | — | SKU JSON 或文件路径 |
 | `--draft` | 否 | false | 关键字段缺失时发布为 draft |
 | `--authorized` | 否 | false | 卖家已确认授权，跳过授权步骤 |
+| `--decline-auth` | 否 | false | 卖家明确拒绝授权，停止流程并提示需授权 |
 | `--output-schema` | 否 | false | 仅输出抽取结构，不发布 |
 | `--ctb-confirm-token` | 否 | — | CTB 阶段返回的 HMAC token |
 | `--field-summary-token` | 否 | — | 字段摘要阶段返回的 HMAC token |
@@ -191,7 +206,7 @@ python scripts/product_skill.py --agent-type qoder --scenario create_product \
 ## 7. 首次激活问候语
 
 ```text
-只需在此处上传您产品的所有资料，例如上传源文件，或选择产品信息所在的文件夹位置，或输入产品信息的网页 URL，本技能将相应地提取信息并将产品发布到 Miravia。
+只需在此处上传您产品的资料，例如上传源文件，或选择产品信息所在的文件夹位置，或输入产品信息的网页 URL。支持格式：PDF、Word、Excel、图片、URL。本技能将相应地提取信息并将产品发布到 Miravia。注意：当前仅支持发布单个商品。
 ```
 
 问候语支持 `zh` / `en` / `es` / `it`，见 `scripts/i18n.py` 中的 `pub_greeting`。
@@ -236,8 +251,7 @@ scripts/skill_keypair.py
 scripts/skill_logger.py
 scripts/telemetry.py
 scripts/trace_logger.py
-scripts/modules/__init__.py
-scripts/modules/publish_module.py
-scripts/modules/extraction_engine.py
-scripts/modules/file_reader.py
+scripts/publish_module.py
+scripts/extraction_engine.py
+scripts/file_reader.py
 ```

package/dist/TECHNICAL.md

@@ -120,7 +120,7 @@ zip -r miravia_product_listing.zip . \
 A: 尚未完成授权，请先执行 `python scripts/auth_cli.py login`。
 
 **Q: AI 抽取效果差？**
-A: 配置 `OPENAI_API_KEY` 或 `DASHSCOPE_API_KEY` 环境变量，使 `modules/extraction_engine.py` 走 LLM 抽取；否则 fallback 到 heuristic 规则，准确度有限。
+A: 配置 `OPENAI_API_KEY` 或 `DASHSCOPE_API_KEY` 环境变量，使 `extraction_engine.py` 走 LLM 抽取；否则 fallback 到 heuristic 规则，准确度有限。
 
 **Q: 品牌未找到怎么办？**
 A: Skill 会返回 `need_brand_create` 状态与 Seller Center 创建链接。卖家创建品牌后，重新运行相同命令即可继续。

package/dist/VERSION

@@ -1 +1 @@
-0.1.2
+0.1.3

package/dist/reference_publish.md

@@ -243,7 +243,7 @@ PRD 中标记为新增 API，但当前 Miravia Open Platform 文档站点未上
 
 ## 附录 A. AI 发品字段映射表
 
-> 本附录说明 `modules/extraction_engine.MiraviaListingSchema` 与 Miravia `/product/create` 请求体之间的映射关系。抽取层使用 snake_case，API 层使用 camelCase / Miravia 原生字段名。
+> 本附录说明 `extraction_engine.MiraviaListingSchema` 与 Miravia `/product/create` 请求体之间的映射关系。抽取层使用 snake_case，API 层使用 camelCase / Miravia 原生字段名。
 
 ### A.1 顶层字段
 

package/dist/scripts/cli_common.py

@@ -43,8 +43,8 @@ def add_common_args(parser: argparse.ArgumentParser) -> None:
     )
     parser.add_argument(
         "--lang",
-        default="auto",
-        help="language: en | es | it | zh",
+        default="es",
+        help="language: en | es | it | zh (default: es for Miravia Spain)",
     )
     parser.add_argument(
         "--format",

package/dist/scripts/{modules/extraction_engine.py → extraction_engine.py}

@@ -223,12 +223,21 @@ class ExtractionEngine:
             r"https?://[^\s\"]+\.(?:jpg|jpeg|png|webp)", text, re.IGNORECASE
         )
 
-        # Prices
-        prices = re.findall(r"(?:€|EUR|Euro|price)\s*[:=]?\s*(\d+(?:\.\d{1,2})?)", text, re.I)
-        if prices:
-            schema.general_price = prices[0]
-        if len(prices) > 1:
-            schema.discounted_price = prices[1]
+        # Prices - handle currency symbols before or after the number
+        # Matches: €10.00, $10, 10 EUR, 10 USD, price: 10.00, price=10
+        prices = re.findall(
+            r"(?:€|\$|EUR|Euro|USD|price)\s*[:=]?\s*(\d+(?:\.\d{1,2})?)"
+            r"|"
+            r"(\d+(?:\.\d{1,2})?)\s*(?:EUR|Euro|USD)",
+            text,
+            re.I,
+        )
+        # Flatten tuples and filter empty strings
+        cleaned_prices = [p for match in prices for p in match if p]
+        if cleaned_prices:
+            schema.general_price = cleaned_prices[0]
+        if len(cleaned_prices) > 1:
+            schema.discounted_price = cleaned_prices[1]
 
         # Stock
         stock_match = re.search(r"(?:stock|quantity)\s*[:=]?\s*(\d+)", text, re.I)

package/dist/scripts/{modules/file_reader.py → file_reader.py}

@@ -23,7 +23,13 @@ logger = logging.getLogger("miravia.product.file_reader")
 
 
 SUPPORTED_IMAGE_SUFFIXES = (".jpg", ".jpeg", ".png", ".webp")
-SUPPORTED_DOC_SUFFIXES = (".pdf", ".doc", ".docx", ".txt", ".md", ".csv", ".xls", ".xlsx")
+SUPPORTED_DOC_SUFFIXES = (".pdf", ".doc", ".docx", ".xls", ".xlsx")
+
+# Formats that are explicitly blocked with a clear error message.
+# .txt is intentionally excluded from supported formats per TC_P2:
+# only PDF / Word / Excel / Image / URL are allowed.
+UNSUPPORTED_SUFFIXES = (".zip", ".txt", ".psd", ".rar", ".7z", ".tar", ".gz",
+                        ".bmp", ".gif", ".tiff", ".svg", ".mp4", ".avi", ".mov")
 
 
 def _guess_mime(path: Path) -> str:
@@ -51,6 +57,10 @@ def _read_text(path: Path) -> str:
 
 
 def _read_pdf(path: Path) -> str:
+    # Detect empty / 0KB files before attempting to parse
+    if path.stat().st_size == 0:
+        raise ValueError(f"File is empty (0 bytes): {path}")
+
     text_parts: List[str] = []
 
     try:
@@ -61,7 +71,10 @@ def _read_pdf(path: Path) -> str:
                 page_text = page.extract_text()
                 if page_text:
                     text_parts.append(page_text)
-        return "\n\n".join(text_parts)
+        result = "\n\n".join(text_parts)
+        if not result.strip():
+            raise ValueError(f"File content is empty or corrupted: {path}")
+        return result
     except ImportError:
         pass
 
@@ -72,7 +85,10 @@ def _read_pdf(path: Path) -> str:
             reader = PyPDF2.PdfReader(f)
             for page in reader.pages:
                 text_parts.append(page.extract_text() or "")
-        return "\n\n".join(text_parts)
+        result = "\n\n".join(text_parts)
+        if not result.strip():
+            raise ValueError(f"File content is empty or corrupted: {path}")
+        return result
     except ImportError:
         raise RuntimeError(
             "PDF reading requires pdfplumber or PyPDF2. Install one of them with:\n"
@@ -82,6 +98,10 @@ def _read_pdf(path: Path) -> str:
 
 
 def _read_word(path: Path) -> str:
+    # Detect empty / 0KB files before attempting to parse
+    if path.stat().st_size == 0:
+        raise ValueError(f"File is empty (0 bytes): {path}")
+
     try:
         import docx
     except ImportError:
@@ -90,9 +110,16 @@ def _read_word(path: Path) -> str:
             "  pip install python-docx"
         )
 
-    doc = docx.Document(path)
+    try:
+        doc = docx.Document(path)
+    except Exception as exc:
+        raise ValueError(f"File content is empty or corrupted: {path} ({exc})")
+
     paragraphs = [p.text for p in doc.paragraphs if p.text.strip()]
-    return "\n\n".join(paragraphs)
+    result = "\n\n".join(paragraphs)
+    if not result.strip():
+        raise ValueError(f"File content is empty or corrupted: {path}")
+    return result
 
 
 def _read_excel(path: Path) -> List[Dict[str, Any]]:
@@ -119,6 +146,27 @@ def _read_excel(path: Path) -> List[Dict[str, Any]]:
     return result
 
 
+def get_excel_row_count(content_items: List[Dict[str, Any]]) -> int:
+    """Count total data rows across all Excel sheets in content items.
+
+    Returns 0 if no Excel/table data found.
+    """
+    total = 0
+    for item in content_items:
+        if item.get("type") != "table":
+            continue
+        content = item.get("content")
+        if not content:
+            continue
+        tables = content if isinstance(content, list) else [content]
+        for table in tables:
+            if isinstance(table, dict) and "rows" in table:
+                total += len(table.get("rows", []))
+            elif isinstance(table, list):
+                total += len(table)
+    return total
+
+
 def _read_csv(path: Path) -> List[Dict[str, str]]:
     encodings = ["utf-8-sig", "utf-8", "gbk"]
     for enc in encodings:
@@ -161,6 +209,13 @@ def _read_local_file(path: Path) -> Dict[str, Any]:
     suffix = path.suffix.lower()
     source = str(path)
 
+    # Block explicitly unsupported formats with a clear error
+    if suffix in UNSUPPORTED_SUFFIXES:
+        raise ValueError(
+            f"Unsupported file format: {source}. "
+            "Only PDF, Word, Excel, Image, and URL are supported."
+        )
+
     if _is_image(path):
         return {"source": source, "type": "image", "content": _read_image(path)}
 
@@ -176,14 +231,12 @@ def _read_local_file(path: Path) -> Dict[str, Any]:
     if suffix == ".csv":
         return {"source": source, "type": "table", "content": _read_csv(path)}
 
-    if suffix in (".txt", ".md", "") or _guess_mime(path).startswith("text/"):
-        return {"source": source, "type": "text", "content": _read_text(path)}
-
-    # Fallback: try as text
-    try:
-        return {"source": source, "type": "text", "content": _read_text(path)}
-    except Exception as exc:
-        raise ValueError(f"Unsupported file type: {path} ({exc})")
+    # .txt, .md and other text-like formats are NOT supported per TC_P2.
+    # Only PDF / Word / Excel / Image / URL are allowed.
+    raise ValueError(
+        f"Unsupported file format: {source}. "
+        "Only PDF, Word, Excel, Image, and URL are supported."
+    )
 
 
 def _read_url(url: str, timeout: int = 30) -> Dict[str, Any]:
@@ -194,8 +247,11 @@ def _read_url(url: str, timeout: int = 30) -> Dict[str, Any]:
             "Chrome/120.0.0.0 Safari/537.36"
         )
     }
-    resp = requests.get(url, headers=headers, timeout=timeout)
-    resp.raise_for_status()
+    try:
+        resp = requests.get(url, headers=headers, timeout=timeout)
+        resp.raise_for_status()
+    except requests.RequestException as exc:
+        raise ValueError(f"Failed to fetch URL: {url} ({exc})")
 
     content_type = resp.headers.get("Content-Type", "")
     if "text/html" in content_type:
@@ -222,6 +278,9 @@ def read_sources(sources: List[str]) -> List[Dict[str, Any]]:
 
     Returns a list of uniform content dicts:
     ``{"source": "...", "type": "text|table|image", "content": ...}``.
+
+    Raises ValueError for user-facing errors (unsupported format, empty file,
+    invalid URL) so the caller can surface a clear message to the user.
     """
     results: List[Dict[str, Any]] = []
 
@@ -231,7 +290,12 @@ def read_sources(sources: List[str]) -> List[Dict[str, Any]]:
             continue
 
         if _is_url(src):
-            results.append(_read_url(src))
+            try:
+                results.append(_read_url(src))
+            except ValueError:
+                raise  # Propagate user-facing URL errors
+            except Exception as exc:
+                raise ValueError(f"Failed to fetch URL: {src} ({exc})")
             continue
 
         path = Path(src).expanduser()
@@ -252,11 +316,15 @@ def read_sources(sources: List[str]) -> List[Dict[str, Any]]:
             for f in files:
                 try:
                     results.append(_read_local_file(f))
+                except ValueError:
+                    raise  # Propagate user-facing file errors
                 except Exception as exc:
                     logger.warning("Failed to read %s: %s", f, exc)
         else:
             try:
                 results.append(_read_local_file(path))
+            except ValueError:
+                raise  # Propagate user-facing file errors
             except Exception as exc:
                 logger.warning("Failed to read %s: %s", src, exc)