structure_assignment/scripts/process.py

import argparse

from structure_assignment.pipeline import Runner

resource_directory = '../resources'

if (__name__ == '__main__'):

    arg_parser = argparse.ArgumentParser(description='Run part or all of structure pipeline.')
    arg_parser.add_argument('-part', type=str, help='Part name')
    arg_parser.add_argument('-infile', type=str, help='Input file')
    arg_parser.add_argument('-outfile', type=str, help='Output file')
    arg_parser.add_argument('-structures', type=str, help='Updated structure file')
    arguments = arg_parser.parse_args()

    part_name = arguments.part
    input_file_name = arguments.infile
    output_file_name = arguments.outfile
    structure_file_name = arguments.structures

    nlp_needed = part_name in {'strings_to_parse', 'strings_to_dictionary', 'all'}
    runner = Runner(resource_directory, nlp_needed)
    if (part_name == 'strings_to_parse'):
        runner.strings_to_parse(input_file_name, output_file_name)
    elif (part_name == 'strings_to_dictionary'):
        runner.strings_to_dictionary(input_file_name, output_file_name, structure_file_name)
    elif (part_name == 'parse_to_dictionary'):
        runner.parse_to_dictionary(input_file_name, output_file_name, structure_file_name)
    elif (part_name == 'validate_structures'):
        runner.validate_structures(input_file_name)
    elif (part_name == 'validate_dictionary'):
        runner.validate_dictionary(input_file_name)
    elif (part_name == 'all'):
        runner.run_all(input_file_name, output_file_name, nlp, structure_file_name)