fix: 物流表格截图识别改用LLM视觉能力，移除阿里云OCR SDK依赖

服务器未安装alibabacloud_ocr_api20210707 SDK导致识别失败。
改为直接将图片base64发给qwen-plus LLM视觉接口识别表格，
跳过OCR步骤，减少依赖链路。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/app/services/order_service.py

@@ -2400,27 +2400,63 @@ class OrderService:
 
 
     def parse_logistics_table_images(self, image_urls, session):
-        """识别物流表格截图，用 LLM 从 OCR 文本中提取快递单号和收件人。"""
+        """识别物流表格截图，用 LLM 视觉能力直接从图片中提取快递单号和收件人。"""
+        import base64 as _b64
+        import json as _json
         import logging as _log
-        _logger = _log.getLogger(__name__)
+        from urllib import request as urllib_request
-        from backend.app.services.ai_service import ai_service
+        from backend.app.services.ai_service import read_image_bytes
         from backend.app.core.config import get_settings
+
+        _logger = _log.getLogger(__name__)
         settings = get_settings()
+        api_key = settings.llm_parse_api_key or settings.aliyun_ai_access_key_id
+
+        system_prompt = (
+            "你是物流信息提取助手。从物流表格截图中提取每一行的快递单号和收件人姓名。\n"
+            "表格通常包含：运单号、运单状态、件数、体积、计费重量、付款方式、运费、"
+            "保价费、包装服务费、信息费、代收货款、签收费、声明价值、产品类型、"
+            "增值服务、服务方式、业务属性、托寄物、寄件人、收件人、目的网点 等列。\n\n"
+            "严格按以下 JSON 数组格式输出，不要添加任何其他内容：\n"
+            '[{"tracking_number":"运单号","express_company":"快递公司名或null","recipient_name":"收件人姓名"}]'
+        )
+
         results = []
         seen_tracking = set()
         for url in image_urls:
             try:
                 _logger.info("[parse_logistics_table_images] 处理图片: %s", url)
-                ctx = ai_service._ocr_and_parse_image(url)
+                image_bytes = read_image_bytes(url)
-                raw_text = ctx.get("raw_text", "")
+                image_b64 = _b64.b64encode(image_bytes).decode("utf-8")
-                _logger.info("[parse_logistics_table_images] OCR原始文本长度: %d, 前500字: %s",
+                suffix = url.rsplit(".", 1)[-1].lower() if "." in url else "jpeg"
-                             len(raw_text), raw_text[:500])
+                mime = {"jpg": "image/jpeg", "jpeg": "image/jpeg", "png": "image/png", "bmp": "image/bmp"}.get(suffix, "image/jpeg")
-                if not raw_text.strip():
+
-                    _logger.warning("[parse_logistics_table_images] OCR未识别到文字")
+                payload = _json.dumps({
-                    continue
+                    "model": settings.llm_parse_model,
-                # 用 LLM 从表格文本中提取快递单号和收件人
+                    "messages": [
-                llm_items = self._llm_extract_logistics(raw_text, settings)
+                        {"role": "system", "content": system_prompt},
-                _logger.info("[parse_logistics_table_images] LLM提取结果: %d条", len(llm_items))
+                        {"role": "user", "content": [
+                            {"type": "image_url", "image_url": {"url": f"data:{mime};base64,{image_b64}"}},
+                            {"type": "text", "text": "请从这个物流表格截图中提取每一行的快递单号和收件人姓名："},
+                        ]},
+                    ],
+                    "temperature": 0.1,
+                    "max_tokens": 4096,
+                }).encode("utf-8")
+
+                req = urllib_request.Request(
+                    url=settings.llm_parse_api_url, data=payload,
+                    headers={"Content-Type": "application/json", "Authorization": f"Bearer {api_key}"},
+                    method="POST",
+                )
+                with urllib_request.urlopen(req, timeout=60) as resp:
+                    result = _json.loads(resp.read().decode("utf-8"))
+
+                content = result["choices"][0]["message"]["content"]
+                _logger.info("[parse_logistics_table_images] LLM响应: %s", content[:500])
+                llm_items = self._extract_json_array(content)
+                _logger.info("[parse_logistics_table_images] 提取结果: %d条", len(llm_items))
+
                 for item in llm_items:
                     tn = (item.get("tracking_number") or "").strip()
                     if not tn or len(tn) < 10 or tn in seen_tracking: