fix eval

Author: QiuYuFong

mindocr/data/transforms/det_east_transforms.py

@@ -4,6 +4,8 @@
 import numpy as np
 from PIL import Image
 from shapely.geometry import Polygon
+import json
+import ast
 
 __all__ = ["EASTProcessTrain"]
 
@@ -414,8 +416,17 @@ def _extract_vertices(self, data_labels):
         """
         vertices_list = []
         labels_list = []
-        data_labels = eval(data_labels)
-        for data_label in data_labels:
+        try:
+            parsed_data = json.loads(data_labels)
+        except json.JSONDecodeError:
+            try:
+                parsed_data = ast.literal_eval(data_labels)
+            except (ValueError, SyntaxError) as e:
+                raise ValueError(f"Invalid data format: {str(e)}") from e
+
+        if not isinstance(parsed_data, list):
+            raise ValueError("Data labels should be a list")
+        for data_label in parsed_data:
             vertices = data_label["points"]
             vertices = [item for point in vertices for item in point]
             vertices_list.append(vertices)

mindocr/data/transforms/svtr_transform.py

@@ -546,10 +546,18 @@ def __init__(self, max_text_length, character_dict_path=None, use_space_char=Fal
 
         self.ctc_encode = CTCLabelEncodeForSVTR(max_text_length, character_dict_path, use_space_char, **kwargs)
         self.gtc_encode_type = gtc_encode
+        # Pls explicitly specify the supported gtc_encode classes and obtain the class objects through dictionaries.
+        supported_gtc_encode = {}
         if gtc_encode is None:
             self.gtc_encode = SARLabelEncodeForSVTR(max_text_length, character_dict_path, use_space_char, **kwargs)
         else:
-            self.gtc_encode = eval(gtc_encode)(max_text_length, character_dict_path, use_space_char, **kwargs)
+            # Mindocr currently does not have a module that requires a custom `gtc_encode` input parameter, and will not
+            # enter this branch at present. If it is supported later, please directly obtain the class reference through
+            # a specific dict, and do not use the `eval` function.
+            if gtc_encode not in supported_gtc_encode:
+                raise ValueError(f"Get unsupported gtc_encode {gtc_encode}")
+            self.gtc_encode = supported_gtc_encode[gtc_encode](max_text_length, character_dict_path,
+                                                               use_space_char, **kwargs)
 
     def __call__(self, data):
         data_ctc = copy.deepcopy(data)

mindocr/postprocess/builder.py

@@ -23,18 +23,24 @@
 
 __all__ = ["build_postprocess"]
 
-supported_postprocess = (
-    det_db_postprocess.__all__
-    + det_pse_postprocess.__all__
-    + det_east_postprocess.__all__
-    + rec_postprocess.__all__
-    + cls_postprocess.__all__
-    + rec_abinet_postprocess.__all__
-    + kie_ser_postprocess.__all__
-    + kie_re_postprocess.__all__
-    + layout_postprocess.__all__
-    + table_postprocess.__all__
-)
+SUPPORTED_POSTPROCESS = {
+    "DBPostprocess": DBPostprocess,
+    "PSEPostprocess": PSEPostprocess,
+    "EASTPostprocess": EASTPostprocess,
+    "CTCLabelDecode": CTCLabelDecode,
+    "RecCTCLabelDecode": RecCTCLabelDecode,
+    "RecAttnLabelDecode": RecAttnLabelDecode,
+    "RecMasterLabelDecode": RecMasterLabelDecode,
+    "VisionLANPostProcess": VisionLANPostProcess,
+    "SARLabelDecode": SARLabelDecode,
+    "ClsPostprocess": ClsPostprocess,
+    "ABINetLabelDecode": ABINetLabelDecode,
+    "VQASerTokenLayoutLMPostProcess": VQASerTokenLayoutLMPostProcess,
+    "VQAReTokenLayoutLMPostProcess": VQAReTokenLayoutLMPostProcess,
+    "YOLOv8Postprocess": YOLOv8Postprocess,
+    "Layoutlmv3Postprocess": Layoutlmv3Postprocess,
+    "TableMasterLabelDecode": TableMasterLabelDecode,
+}
 
 
 def build_postprocess(config: dict):
@@ -57,11 +63,11 @@ def build_postprocess(config: dict):
         >>> postprocess
     """
     proc = config.pop("name")
-    if proc in supported_postprocess:
-        postprocessor = eval(proc)(**config)
-    elif proc is None:
+    if proc is None:
         return None
+    if proc in SUPPORTED_POSTPROCESS:
+        postprocessor = SUPPORTED_POSTPROCESS[proc](**config)
     else:
-        raise ValueError(f"Invalid postprocess name {proc}, support postprocess are {supported_postprocess}")
+        raise ValueError(f"Invalid postprocess name {proc}, support postprocess are {SUPPORTED_POSTPROCESS.keys()}")
 
     return postprocessor

tools/arg_parser.py

@@ -62,7 +62,10 @@ def _parse_options(opts: list):
             "=" in opt_str
         ), "Invalid option {}. A valid option must be in the format of {{key_name}}={{value}}".format(opt_str)
         k, v = opt_str.strip().split("=")
-        options[k] = yaml.load(v, Loader=yaml.Loader)
+        try:
+            options[k] = yaml.load(v, Loader=yaml.SafeLoader)
+        except yaml.YAMLError as e:
+            raise ValueError(f"Failed to parse value for key '{k}': {str(e)}") from e
     # print('Parsed options: ', options)
 
     return options