python dataset_service.py <path_to_template> <path_to_target> <number_of_cities>
This scripts takes three arguments, one for the template csv file, one for the target csv file and one for the number of cities you want to add to the dataset. It will take all the sentences found in the template file, and replace all occurences of "X" and "Y" with real french cities. Each city will be an origin and a destination once for each sentence in the template. It will export the result in a new csv file corresponding to the second parameter.
python dataset_service.py <path_to_dataset_csv>
This will convert the dataset generated by the previous script into a train.spacy file used for training spacy models.
Generated datesets (train & eval) file's name should be corresponded to the naming that is available in config.cfg
Example:
[paths] train = "./data/train_output.spacy" dev = "./d
5D32
ata/eval_output.spacy"