normalize.py

import change_manager
import wordnet
import wordnet_yaml


def normalize(text):
    return text.strip().replace("  ", " ").replace("( ", "(").replace(" )", ")")


def main():
    wn = change_manager.parse_wordnet("wn.xml")

    for synset in wn.synsets:
        synset.definitions = [
            wordnet.Definition(normalize(d.text))
            for d in synset.definitions
            if d.text.strip()
        ]

        if synset.ili_definition:
            synset.ili_definition = wordnet.Definition(
                normalize(synset.ili_definition.text)
            )

        for example in synset.examples:
            example.text = normalize(example.text)

        synset.examples = [
            example for example in synset.examples if example.text.strip()
        ]

    change_manager.save_all_xml(wn)
    wordnet_yaml.save(wn)


if __name__ == "__main__":
    main()