3 次代码提交 5bb909781f ... 3bb58ac275

作者 SHA1 备注 提交日期
  zhch158_admin 3bb58ac275 feat(添加严农业银行文档配置): 新增严农业银行的OCR文档配置及相关工具设置 1 周之前
  zhch158_admin 87d30e6dc5 feat(移除VL识别模型参数): 删除不再需要的模型参数配置以简化设置 1 周之前
  zhch158_admin 0d52ee2e18 feat(更新处理器配置): 添加PaddleOCR处理器配置,更新日志目录,新增YUSYS本地处理器支持 1 周之前

+ 2 - 1
ocr_tools/ocr_batch/pdf_list.txt

@@ -15,5 +15,6 @@ B用户_扫描流水.pdf,bank_statement
 湛_平安银行图.pdf,bank_statement
 张_微信图.pdf,bank_statement
 朱_中信银行图.pdf,bank_statement
-韩_中国银行图.pdf,bank_statement
+韩_中国银行图.pdf,bank_statemen
+严_农业银行.pdf,bank_statement
 

+ 1 - 0
ocr_tools/ocr_batch/pdf_list_local.txt

@@ -5,3 +5,4 @@ B用户_扫描流水.pdf,bank_statement
 施博深.pdf,bank_statement
 山西云集科技有限公司.pdf,bank_statement
 许_民生银行图.pdf,bank_statement
+严_农业银行.pdf,bank_statement

+ 17 - 1
ocr_tools/ocr_batch/processor_configs.yaml

@@ -74,10 +74,26 @@ processors:
       - "--debug"
       - "--log_level=DEBUG"
     output_subdir: "bank_statement_yusys_local"
-    log_subdir: "logs/bank_statement_yusys_v4_local"
+    log_subdir: "logs/bank_statement_yusys_local"
     venv: "conda activate mineru2"
     description: "YUSYS(local) Wired UNET OCR GLM-OCR"
 
+  yusys_paddleocr_local:
+    script: "/Users/zhch158/workspace/repository.git/ocr_platform/ocr_tools/universal_doc_parser/main_v2.py"
+    input_arg: "--input"
+    output_arg: "--output_dir"
+    scene_arg: "--scene"
+    extra_args:
+      - "--config=/Users/zhch158/workspace/repository.git/ocr_platform/ocr_tools/universal_doc_parser/config/bank_statement_paddle_vl_local.yaml"
+      - "--pages=1-35"
+      - "--streaming"
+      - "--debug"
+      - "--log_level=DEBUG"
+    output_subdir: "bank_statement_yusys_paddleocr_local"
+    log_subdir: "logs/bank_statement_yusys_paddleocr_local"
+    venv: "conda activate mineru2"
+    description: "YUSYS(local) Wired UNET OCR PaddleOCR-VL"
+
   # -------------------------------------------------------------------------
   # PaddleOCR-VL 处理器
   # -------------------------------------------------------------------------

+ 0 - 16
ocr_tools/universal_doc_parser/config/bank_statement_paddle_vl_local.yaml

@@ -176,22 +176,6 @@ vl_recognition:
     formula: "Formula Recognition:"
     seal: "Seal Recognition:"  # 印章识别的专用提示词
   
-  # 模型参数
-  model_params:
-    connection_pool_size: 128  # HTTP 连接池大小(应 >= max_workers)
-    http_timeout: 300          # HTTP 请求超时时间(秒)
-    connect_timeout: 30        # 连接超时时间(秒)
-    retry_max_attempts: 2      # 最大重试次数
-    retry_backoff_base_seconds: 0.5
-    retry_backoff_max_seconds: 8.0
-    retry_jitter_ratio: 0.2
-    retry_status_codes: [429, 500, 502, 503, 504]
-    max_tokens: 16384
-    temperature: 0.1
-    top_p: 0.0001
-    top_k: 1
-    repetition_penalty: 1.1
-  
   # 场景特定配置
   table_recognition:
 

+ 1 - 0
ocr_validator/config/global.yaml

@@ -162,3 +162,4 @@ data_sources:
   - 许_民生银行图.yaml
   - 韩_中国银行图.yaml
   - 杨万益_福建农信.yaml
+  - 严_农业银行.yaml

+ 20 - 0
ocr_validator/config/严_农业银行.yaml

@@ -0,0 +1,20 @@
+# 文档: 严_农业银行
+document:
+  name: "严_农业银行"
+  base_dir: "/Users/zhch158/workspace/data/流水分析/严_农业银行"
+  
+  # 🎯 关键改进:定义该文档使用的 OCR 工具及其结果目录
+  ocr_results:
+    # bank_statement_yusys_local
+    - tool: "mineru"
+      result_dir: "bank_statement_yusys_local"
+      image_dir: "bank_statement_yusys_local/{{name}}"
+      description: "YUSYS-OCR框架(local) GLM-OCR"
+      enabled: true
+
+    # bank_statement_yusys_paddleocr_local
+    - tool: "mineru"
+      result_dir: "bank_statement_yusys_paddleocr_local"
+      image_dir: "bank_statement_yusys_paddleocr_local/{{name}}"
+      description: "YUSYS-OCR框架(local) PaddleOCR-VL"
+      enabled: true