dictdiff

#!/usr/bin/env python3

import argparse
import sys

from pathlib import Path

def wordlist_to_dictionary(path):
    with path.open() as f:
        lines = [l.strip() for l in f.readlines()] 

    return {l.split(';')[0]:int(l.split(';')[1]) if len(l.split(';')) == 2
                else None
                for l in lines}

def main():
    parser = argparse.ArgumentParser(description='Compare two .dict files and produce files showing what has been added, removed, and rescored.')

    parser.add_argument('olddict', type=Path)
    parser.add_argument('newdict', type=Path)

    args = parser.parse_args()

    oldlist = wordlist_to_dictionary(args.olddict)
    newlist = wordlist_to_dictionary(args.newdict)

    added = {w:newlist[w] for w in newlist.keys() if w not in oldlist.keys()}
    removed = {w:oldlist[w] for w in oldlist.keys() if w not in newlist.keys()}
    
    rescored = {w:[oldlist[w], newlist[w]] for w in oldlist.keys() if
                   w in newlist.keys() and oldlist[w] != newlist[w]}

    print('Added: {}'.format(len(added)))
    print('Removed: {}'.format(len(removed)))
    print('Rescored: {}'.format(len(rescored)))

    oldlist_name = args.olddict.name[:-len(''.join(args.olddict.suffixes))]
    newlist_name = args.newdict.name[:-len(''.join(args.newdict.suffixes))]

    with open('-'.join([oldlist_name, newlist_name, 'ADDED.dict']), 'w') as f:
        f.write('\n'.join([';'.join([w, str(added[w])]) for w in added.keys()]))
    with open('-'.join([oldlist_name, newlist_name, 'REMOVED.dict']), 'w') as f:
        f.write('\n'.join([';'.join([w, str(removed[w])]) for w in removed.keys()]))

    with open('-'.join([oldlist_name, newlist_name, 'RESCORED.txt']), 'w') as f:
        f.write('\n'.join(['{}: {} => {}'.format(w, 
                            rescored[w][0], rescored[w][1]) 
                            for w in rescored.keys()]))
    with open('-'.join([oldlist_name, newlist_name, 'RESCORED.dict']), 'w') as f:
        f.write('\n'.join([';'.join([w, str(newlist[w])]) for w in rescored.keys()]))


if __name__ == '__main__':
    main()