feature: 增加水印去除/高级匿名化算子 (#151)

* feature: 增加水印去除算子

* feature: clean code

* feature: clean code

* feature: 增加高级匿名化算子

Author: hhhhsc701

README.md

@@ -73,15 +73,15 @@ make install-mineru
 ```
 
 ### Deploy the DeerFlow service
-1. Modify `runtime/deer-flow/.env.example` and add configurations for SEARCH_API_KEY and the EMBEDDING model.
-2. Modify `runtime/deer-flow/.conf.yaml.example` and add basic model service configurations.
-3. Execute `make install-deer-flow`
+```bash
+make install-deer-flow
+```
 
 ### Local Development and Deployment
 After modifying the local code, please execute the following commands to build the image and deploy using the local image.
 ```bash
 make build
-make install REGISTRY=""
+make install dev=true
 ```
 
 ## 🤝 Contribution Guidelines

runtime/ops/mapper/__init__.py

@@ -23,7 +23,6 @@ def _import_operators():
     from . import garble_characters_cleaner
     from . import html_tag_cleaner
     from . import id_number_cleaner
-    from . import img_watermark_remove
     from . import invisible_characters_cleaner
     from . import ip_address_cleaner
     from . import legend_cleaner
@@ -47,6 +46,7 @@ def _import_operators():
     from . import img_resize
     from . import remove_duplicate_sentences
     from . import knowledge_relation_slice
+    from . import pii_ner_detection
 
 
 _import_operators()

runtime/ops/mapper/img_direction_correct/base_model.py

@@ -11,7 +11,6 @@ class BaseModel:
 
     def __init__(self, model_type='vertical'):
         models_path = os.getenv("MODELS_PATH", "/home/models")
-        self.resources_path = str(Path(models_path, 'img_direction_correct', 'resources'))
         args = Namespace()
         args.cls_image_shape = '3, 224, 224'
         args.cls_batch_num = 6
@@ -20,13 +19,14 @@ def __init__(self, model_type='vertical'):
         args.use_gpu = False
         args.use_npu = False
         args.use_xpu = False
+        args.use_mlu = False
         args.enable_mkldnn = False
         if model_type == 'vertical':
-            args.cls_model_dir = str(Path(self.resources_path, 'vertical_model'))
+            args.cls_model_dir = str(Path(models_path, 'ch_ppocr_mobile_v2.0_cls_infer'))
             self.model_name = 'standard model to detect image 0 or 90 rotated'
             args.label_list = ['0', '90']
         else:
-            args.cls_model_dir = str(Path(self.resources_path, 'standard_model'))
+            args.cls_model_dir = str(Path(models_path, 'ch_ppocr_mobile_v2.0_cls_infer'))
             self.model_name = 'standard model to detect image 0 or 180 rotated'
             args.label_list = ['0', '180']
 

runtime/ops/mapper/img_watermark_remove/__init__.py

@@ -0,0 +1,4 @@
+from datamate.core.base_op import OPERATORS
+
+OPERATORS.register_module(module_name='PiiDetector',
+                          module_path='ops.mapper.pii_ner_detection.process')

runtime/ops/mapper/img_watermark_remove/metadata.yml

@@ -0,0 +1,62 @@
+import presidio_analyzer as analyzer
+
+# 中国身份证号识别器
+id_recognizer = analyzer.PatternRecognizer(
+    supported_entity="ID_CHINA",
+    supported_language="zh",
+    patterns=[
+        analyzer.Pattern(
+            name="china_id_pattern",
+            regex=r"\b[1-9]\d{5}(19|20)\d{2}(0[1-9]|1[0-2])(0[1-9]|[12]\d|3[01])\d{3}[\dXx]\b|\b[1-9]\d{7}(0[1-9]|1[0-2])(0[1-9]|[12]\d|3[01])\d{3}\b",
+            score=0.9
+        )
+    ],
+    context=["身份证", "身份证明", "身份证号", "证件号码"]
+)
+
+# 中国电话号码识别器
+phone_recognizer = analyzer.PatternRecognizer(
+    supported_entity="Phone_CHINA",
+    supported_language="zh",
+    patterns=[
+        analyzer.Pattern(
+            name="china_mobile_pattern",
+            regex=r"\b(1[3-9]\d{9})\b",
+            score=0.85
+        ),
+        analyzer.Pattern(
+            name="china_landline_pattern",
+            regex=r"\b(0\d{2,3}-?\d{7,8})\b",
+            score=0.8
+        )
+    ],
+    context=["电话", "手机", "联系方式", "联系电话"]
+)
+
+# 中国邮编识别器
+zipcode_recognizer = analyzer.PatternRecognizer(
+    supported_entity="ZIPCODE_CHINA",
+    supported_language="zh",
+    patterns=[
+        analyzer.Pattern(
+            name="china_zipcode_pattern",
+            regex=r"\b[1-9]\d{5}\b",
+            score=0.7
+        )
+    ],
+    context=["邮编", "邮政编码", "邮编号码"]
+)
+
+# 兼容中文域名的URL识别器
+url_recognizer = analyzer.PatternRecognizer(
+    supported_entity="URL",
+    supported_language="zh",
+    patterns=[
+        analyzer.Pattern(
+            name="url_pattern",
+            regex=r"\b((?:https?://|www\.)[\w-]+\.[\w-]+\S*|(?:https?://|www\.)[\u4e00-\u9fa5]+\.[\u4e00-\u9fa5]+\S*)\b",
+            score=0.9
+        )
+    ],
+    context=["网址", "链接", "网站", "网页"]
+)

runtime/ops/mapper/img_watermark_remove/process.py

@@ -0,0 +1,9 @@
+name: '高级匿名化'
+language: 'Python'
+vendor: 'others'
+raw_id: 'PiiDetector'
+version: '1.0.0'
+description: '高级匿名化算子，检测命名实体并匿名化。'
+modal: 'text'
+inputs: 'text'
+outputs: 'text'