Comparison of search algorithms

pass.py

import os
import sys
import re
import difflib

store = os.path.abspath('/home/matthias/dev/password-store/')


def get_password_names_old(name):
    names = []

    for root, dirs, files in os.walk(store):
        dir_path = root[len(store) + 1:]

        if dir_path.startswith('.'):
            continue

        for filename in files:
            path = os.path.join(dir_path, filename)
            if re.match(r'.*{0}.*\.gpg$'.format(name), path, re.IGNORECASE):
                names.append(path[:-4])

    return names


def get_password_names_new(name):
    def is_junk(x):
        return x == ' '

    matcher = difflib.SequenceMatcher(isjunk=is_junk, b=name, autojunk=False)
    matches = []

    for root, dirs, files in os.walk(store):
        dir_path = root[len(store) + 1:]

        if dir_path.startswith('.'):
            continue

        for filename in files:
            path = os.path.join(dir_path, filename)
            matcher.set_seq1(path[:-4])
            score = matcher.ratio()

            if score >= 0.5:
                matches.append((score, path[:-4]))

    return [x[1] for x in sorted(matches, key=lambda x: x[0], reverse=True)]


def get_password_names_other(name):
    matcher = difflib.SequenceMatcher(b=name, autojunk=False)
    matches = {}

    for root, dirs, files in os.walk(store):
        dir_path = root[len(store) + 1:]

        if dir_path.startswith('.'):
            continue

        for filename in files:
            path = os.path.join(dir_path, filename)[:-4]

            for name in path.split('/'):
                matcher.set_seq1(name)
                score = matcher.ratio()

                if score >= 0.5 and (path not in matches or score > matches[path]):
                    matches[path] = score

    return sorted(matches, key=matches.__getitem__, reverse=True)


if __name__ == '__main__':
    print "old =", get_password_names_old(sys.argv[1])
    print "new =", get_password_names_new(sys.argv[1])
    print "oth =", get_password_names_other(sys.argv[1])