MCPcopy Create free account
hub / github.com/Topdu/OpenOCR / get_ext_data

Method get_ext_data

tools/data/simple_dataset.py:104–137  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

102 return file_name
103
104 def get_ext_data(self):
105 ext_data_num = 0
106 for op in self.ops:
107 if hasattr(op, 'ext_data_num'):
108 ext_data_num = getattr(op, 'ext_data_num')
109 break
110 load_data_ops = self.ops[:self.ext_op_transform_idx]
111 ext_data = []
112
113 while len(ext_data) < ext_data_num:
114 file_idx = self.data_idx_order_list[np.random.randint(
115 self.__len__())]
116 data_line = self.data_lines[file_idx]
117 data_line = data_line.decode('utf-8')
118 substr = data_line.strip('\n').split(self.delimiter)
119 file_name = substr[0]
120 file_name = self._try_parse_filename_list(file_name)
121 label = substr[1]
122 img_path = os.path.join(self.data_dir, file_name)
123 data = {'img_path': img_path, 'label': label}
124 if not os.path.exists(img_path):
125 continue
126 with open(data['img_path'], 'rb') as f:
127 img = f.read()
128 data['image'] = img
129 data = transform(data, load_data_ops)
130
131 if data is None:
132 continue
133 if 'polys' in data.keys():
134 if data['polys'].shape[1] != 4:
135 continue
136 ext_data.append(data)
137 return ext_data
138
139 def __getitem__(self, idx):
140 file_idx = self.data_idx_order_list[idx]

Callers 2

__getitem__Method · 0.95
__getitem__Method · 0.45

Calls 4

__len__Method · 0.95
transformFunction · 0.90
decodeMethod · 0.45

Tested by

no test coverage detected