diff --git a/configs/rec/multi_language/generate_multi_language_configs.py b/configs/rec/multi_language/generate_multi_language_configs.py index 027a65c0c73e4b39ebb3b407271450d23281aa80..d2302a41d14085395de0f82f4970bce7cdd7eb95 100644 --- a/configs/rec/multi_language/generate_multi_language_configs.py +++ b/configs/rec/multi_language/generate_multi_language_configs.py @@ -19,21 +19,38 @@ import logging logging.basicConfig(level=logging.INFO) support_list = { - 'it':'italian', 'xi':'spanish', 'pu':'portuguese', 'ru':'russian', 'ar':'arabic', - 'ta':'tamil', 'ug':'uyghur', 'fa':'persian', 'ur':'urdu', 'rs':'serbian latin', - 'oc':'occitan', 'rsc':'serbian cyrillic', 'bg':'bulgarian', 'uk':'ukranian', 'be':'belarusian', - 'te':'telugu', 'ka':'kannada', 'chinese_cht':'chinese tradition','hi':'hindi','mr':'marathi', - 'ne':'nepali', + 'it': 'italian', + 'es': 'spanish', + 'pt': 'portuguese', + 'ru': 'russian', + 'ar': 'arabic', + 'ta': 'tamil', + 'ug': 'uyghur', + 'fa': 'persian', + 'ur': 'urdu', + 'rs': 'serbian latin', + 'oc': 'occitan', + 'rsc': 'serbian cyrillic', + 'bg': 'bulgarian', + 'uk': 'ukranian', + 'be': 'belarusian', + 'te': 'telugu', + 'ka': 'kannada', + 'chinese_cht': 'chinese tradition', + 'hi': 'hindi', + 'mr': 'marathi', + 'ne': 'nepali', } -assert( - os.path.isfile("./rec_multi_language_lite_train.yml") - ),"Loss basic configuration file rec_multi_language_lite_train.yml.\ +assert (os.path.isfile("./rec_multi_language_lite_train.yml") + ), "Loss basic configuration file rec_multi_language_lite_train.yml.\ You can download it from \ https://github.com/PaddlePaddle/PaddleOCR/tree/dygraph/configs/rec/multi_language/" - -global_config = yaml.load(open("./rec_multi_language_lite_train.yml", 'rb'), Loader=yaml.Loader) + +global_config = yaml.load( + open("./rec_multi_language_lite_train.yml", 'rb'), Loader=yaml.Loader) project_path = os.path.abspath(os.path.join(os.getcwd(), "../../../")) + class ArgsParser(ArgumentParser): def __init__(self): super(ArgsParser, self).__init__( @@ -41,15 +58,30 @@ class ArgsParser(ArgumentParser): self.add_argument( "-o", "--opt", nargs='+', help="set configuration options") self.add_argument( - "-l", "--language", nargs='+', help="set language type, support {}".format(support_list)) + "-l", + "--language", + nargs='+', + help="set language type, support {}".format(support_list)) self.add_argument( - "--train",type=str,help="you can use this command to change the train dataset default path") + "--train", + type=str, + help="you can use this command to change the train dataset default path" + ) self.add_argument( - "--val",type=str,help="you can use this command to change the eval dataset default path") + "--val", + type=str, + help="you can use this command to change the eval dataset default path" + ) self.add_argument( - "--dict",type=str,help="you can use this command to change the dictionary default path") + "--dict", + type=str, + help="you can use this command to change the dictionary default path" + ) self.add_argument( - "--data_dir",type=str,help="you can use this command to change the dataset default root path") + "--data_dir", + type=str, + help="you can use this command to change the dataset default root path" + ) def parse_args(self, argv=None): args = super(ArgsParser, self).parse_args(argv) @@ -68,20 +100,28 @@ class ArgsParser(ArgumentParser): return config def _set_language(self, type): - assert(type),"please use -l or --language to choose language type" + assert (type), "please use -l or --language to choose language type" assert( type[0] in support_list.keys() ),"the sub_keys(-l or --language) can only be one of support list: \n{},\nbut get: {}, " \ "please check your running command".format(support_list, type) - global_config['Global']['character_dict_path'] = 'ppocr/utils/dict/{}_dict.txt'.format(type[0]) - global_config['Global']['save_model_dir'] = './output/rec_{}_lite'.format(type[0]) - global_config['Train']['dataset']['label_file_list'] = ["train_data/{}_train.txt".format(type[0])] - global_config['Eval']['dataset']['label_file_list'] = ["train_data/{}_val.txt".format(type[0])] + global_config['Global'][ + 'character_dict_path'] = 'ppocr/utils/dict/{}_dict.txt'.format(type[ + 0]) + global_config['Global'][ + 'save_model_dir'] = './output/rec_{}_lite'.format(type[0]) + global_config['Train']['dataset'][ + 'label_file_list'] = ["train_data/{}_train.txt".format(type[0])] + global_config['Eval']['dataset'][ + 'label_file_list'] = ["train_data/{}_val.txt".format(type[0])] global_config['Global']['character_type'] = type[0] - assert( - os.path.isfile(os.path.join(project_path,global_config['Global']['character_dict_path'])) - ),"Loss default dictionary file {}_dict.txt.You can download it from \ -https://github.com/PaddlePaddle/PaddleOCR/tree/dygraph/ppocr/utils/dict/".format(type[0]) + assert ( + os.path.isfile( + os.path.join(project_path, global_config['Global'][ + 'character_dict_path'])) + ), "Loss default dictionary file {}_dict.txt.You can download it from \ +https://github.com/PaddlePaddle/PaddleOCR/tree/dygraph/ppocr/utils/dict/".format( + type[0]) return type[0] @@ -110,43 +150,51 @@ def merge_config(config): cur[sub_key] = value else: cur = cur[sub_key] - + + def loss_file(path): - assert( - os.path.exists(path) - ),"There is no such file:{},Please do not forget to put in the specified file".format(path) + assert ( + os.path.exists(path) + ), "There is no such file:{},Please do not forget to put in the specified file".format( + path) + - if __name__ == '__main__': FLAGS = ArgsParser().parse_args() merge_config(FLAGS.opt) save_file_path = 'rec_{}_lite_train.yml'.format(FLAGS.language) if os.path.isfile(save_file_path): os.remove(save_file_path) - + if FLAGS.train: global_config['Train']['dataset']['label_file_list'] = [FLAGS.train] - train_label_path = os.path.join(project_path,FLAGS.train) + train_label_path = os.path.join(project_path, FLAGS.train) loss_file(train_label_path) if FLAGS.val: global_config['Eval']['dataset']['label_file_list'] = [FLAGS.val] - eval_label_path = os.path.join(project_path,FLAGS.val) + eval_label_path = os.path.join(project_path, FLAGS.val) loss_file(eval_label_path) if FLAGS.dict: global_config['Global']['character_dict_path'] = FLAGS.dict - dict_path = os.path.join(project_path,FLAGS.dict) + dict_path = os.path.join(project_path, FLAGS.dict) loss_file(dict_path) if FLAGS.data_dir: global_config['Eval']['dataset']['data_dir'] = FLAGS.data_dir global_config['Train']['dataset']['data_dir'] = FLAGS.data_dir - data_dir = os.path.join(project_path,FLAGS.data_dir) + data_dir = os.path.join(project_path, FLAGS.data_dir) loss_file(data_dir) - + with open(save_file_path, 'w') as f: - yaml.dump(dict(global_config), f, default_flow_style=False, sort_keys=False) + yaml.dump( + dict(global_config), f, default_flow_style=False, sort_keys=False) logging.info("Project path is :{}".format(project_path)) - logging.info("Train list path set to :{}".format(global_config['Train']['dataset']['label_file_list'][0])) - logging.info("Eval list path set to :{}".format(global_config['Eval']['dataset']['label_file_list'][0])) - logging.info("Dataset root path set to :{}".format(global_config['Eval']['dataset']['data_dir'])) - logging.info("Dict path set to :{}".format(global_config['Global']['character_dict_path'])) - logging.info("Config file set to :configs/rec/multi_language/{}".format(save_file_path)) + logging.info("Train list path set to :{}".format(global_config['Train'][ + 'dataset']['label_file_list'][0])) + logging.info("Eval list path set to :{}".format(global_config['Eval'][ + 'dataset']['label_file_list'][0])) + logging.info("Dataset root path set to :{}".format(global_config['Eval'][ + 'dataset']['data_dir'])) + logging.info("Dict path set to :{}".format(global_config['Global'][ + 'character_dict_path'])) + logging.info("Config file set to :configs/rec/multi_language/{}". + format(save_file_path)) diff --git a/ppocr/data/imaug/label_ops.py b/ppocr/data/imaug/label_ops.py index 7a32d870bfc7f532896ce6b11aac5508a6369993..75c6530d22c5b0d5ad9c0c4414132ec5fa4e3dd1 100644 --- a/ppocr/data/imaug/label_ops.py +++ b/ppocr/data/imaug/label_ops.py @@ -94,7 +94,7 @@ class BaseRecLabelEncode(object): use_space_char=False): support_character_type = [ 'ch', 'en', 'EN_symbol', 'french', 'german', 'japan', 'korean', - 'EN', 'it', 'xi', 'pu', 'ru', 'ar', 'ta', 'ug', 'fa', 'ur', 'rs', + 'EN', 'it', 'es', 'pt', 'ru', 'ar', 'ta', 'ug', 'fa', 'ur', 'rs', 'oc', 'rsc', 'bg', 'uk', 'be', 'te', 'ka', 'chinese_cht', 'hi', 'mr', 'ne' ] diff --git a/ppocr/utils/dict/xi_dict.txt b/ppocr/utils/dict/es_dict.txt similarity index 100% rename from ppocr/utils/dict/xi_dict.txt rename to ppocr/utils/dict/es_dict.txt diff --git a/ppocr/utils/dict/pu_dict.txt b/ppocr/utils/dict/pt_dict.txt similarity index 100% rename from ppocr/utils/dict/pu_dict.txt rename to ppocr/utils/dict/pt_dict.txt