feat: 百度通用文字识别提取供应商/日期（config api_key 持久化）

2026-07-19 19:23:38 +08:00
parent cc66448327
commit 056e7d8e75
14 changed files with 1198 additions and 17 deletions
@@ -0,0 +1,162 @@
+"""批量处理服务：扫描 data/input/ 下图片，串行跑完整 OCR → 采购单流程。
+
+- 不合并（每张图片单独出一个 result）
+- 默认串行（避免百度 OCR 限流）
+- 返回 dict 汇总：total / success / failed / results[]
+"""
+from __future__ import annotations
+
+import logging
+import os
+from pathlib import Path
+from typing import Any, Callable, Dict, List, Optional
+
+from ..config.settings import ConfigManager
+from ..core.utils.log_utils import get_logger
+from .ocr_service import OCRService
+from .order_service import OrderService
+
+logger = get_logger(__name__)
+
+
+IMAGE_EXTS = {'.png', '.jpg', '.jpeg', '.bmp', '.webp', '.tif', '.tiff'}
+
+
+class BatchService:
+    """批量处理服务。"""
+
+    def __init__(self, config: Optional[ConfigManager] = None):
+        self.config = config or ConfigManager()
+        self.ocr_service = OCRService(self.config)
+        self.order_service = OrderService(self.config)
+        self._input_folder = self.config.get_path(
+            'Paths', 'input_folder', fallback='data/input', create=True
+        )
+
+    @property
+    def input_folder(self) -> str:
+        return self._input_folder
+
+    def list_input_images(self) -> List[str]:
+        """列出 data/input/ 下所有待处理图片（按 mtime 升序，先入先出）。"""
+        if not os.path.isdir(self._input_folder):
+            return []
+        files = []
+        for name in os.listdir(self._input_folder):
+            p = os.path.join(self._input_folder, name)
+            if not os.path.isfile(p):
+                continue
+            ext = os.path.splitext(name)[1].lower()
+            if ext in IMAGE_EXTS:
+                files.append(p)
+        files.sort(key=lambda p: os.path.getmtime(p))
+        return files
+
+    def process_all_inputs(
+        self,
+        progress_cb: Optional[Callable[[int, int, dict], None]] = None,
+    ) -> Dict[str, Any]:
+        """批量处理 data/input/ 下全部图片。
+
+        Args:
+            progress_cb: 可选回调 fn(done_count, total_count, latest_result_dict)
+
+        Returns:
+            {
+                'total': int,
+                'success': int,
+                'failed': int,
+                'results': [
+                    {
+                        'image': str,
+                        'hash': str,
+                        'status': 'success' | 'failed',
+                        'supplier': str,
+                        'bill_date': str,
+                        'total_amount': float,
+                        'result_file': str | None,
+                        'error': str | None,
+                    },
+                    ...
+                ]
+            }
+        """
+        images = self.list_input_images()
+        total = len(images)
+        results: List[Dict[str, Any]] = []
+
+        if total == 0:
+            logger.info("data/input/ 下无待处理图片")
+            return {'total': 0, 'success': 0, 'failed': 0, 'results': []}
+
+        logger.info(f"批量处理开始: 共 {total} 张图片")
+        success = 0
+        failed = 0
+
+        for idx, image_path in enumerate(images, start=1):
+            entry: Dict[str, Any] = {
+                'image': image_path,
+                'hash': '',
+                'status': 'failed',
+                'supplier': '',
+                'bill_date': '',
+                'total_amount': 0.0,
+                'result_file': None,
+                'error': None,
+            }
+            try:
+                file_hash = Path(image_path).stem
+                entry['hash'] = file_hash
+
+                # 1) OCR（写 data/output/{hash}.xlsx + .meta.json）
+                excel_path = self.ocr_service.process_image(image_path)
+                if not excel_path:
+                    entry['error'] = 'OCR 失败'
+                    failed += 1
+                    results.append(entry)
+                    if progress_cb:
+                        progress_cb(idx, total, entry)
+                    continue
+
+                # 2) process_excel（识别元信息 + 重命名 + 落库）
+                result_file = self.order_service.process_excel(excel_path)
+                if not result_file:
+                    entry['error'] = '处理失败'
+                    failed += 1
+                    results.append(entry)
+                    if progress_cb:
+                        progress_cb(idx, total, entry)
+                    continue
+
+                # 3) 回查元信息（process_excel 内部已落库）
+                meta_row = self.order_service.metadata_db.get(file_hash) or {}
+                entry['result_file'] = result_file
+                entry['supplier'] = meta_row.get('supplier', '') or ''
+                entry['bill_date'] = meta_row.get('bill_date', '') or ''
+                entry['total_amount'] = float(meta_row.get('total_amount') or 0.0)
+                entry['status'] = 'success'
+                success += 1
+                logger.info(
+                    f"[{idx}/{total}] 处理完成: {Path(image_path).name} -> "
+                    f"{Path(result_file).name} | 供应商={entry['supplier']!r} "
+                    f"日期={entry['bill_date']!r} 金额={entry['total_amount']:.2f}"
+                )
+            except Exception as e:
+                logger.error(f"[{idx}/{total}] 处理异常: {image_path}: {e}", exc_info=True)
+                entry['error'] = str(e)
+                failed += 1
+
+            results.append(entry)
+            if progress_cb:
+                progress_cb(idx, total, entry)
+
+        summary = {
+            'total': total,
+            'success': success,
+            'failed': failed,
+            'results': results,
+        }
+        logger.info(
+            f"批量处理完成: 总 {total} 张, 成功 {success}, 失败 {failed}"
+        )
+        return summary