A script to help with the Split Decisions word puzzles in the New York Times

splits.py

#!/usr/bin/env python3
# coding: utf-8
""" splits.py

    Usage:
        ./splits .in,cr..

    This will print out all known words that match both patterns with fixed
    letters per dot. In the above example, one answer pair would be:

        sinew, screw
"""


# ______________________________________________________________________
# Imports

import re
import sys


# ______________________________________________________________________
# Functions

# This expects a dot/letter pattern in `pat`, as in '..a.'.
# This replaces the given (non-dot) letters in `s` with the
# letters provided in `pat`. For example:
#     replace_pattern('hello', 'j...y')  # returns 'jelly'
def replace_pattern(s, pat):
    lets = list(s)
    for i, let in enumerate(pat):
        if let != '.':
            lets[i] = let
    return ''.join(lets)

# This expects a string like '.in,cr..', and extracts two
# strings like '.in..' and '.cr..'.
def parse_patterns(s):
    # Find the start of the AB,CD substring.
    start = [i for i, let in enumerate(s) if let != '.'][0]
    end   = start + 5
    pat1  = s[:start + 2] + s[end:]
    pat2  = s[:start] + s[end - 2:]
    return pat1, pat2

def find_matches(pat1, pat2):

    # Load our dictionary.
    with open('/usr/share/dict/words') as f:
        words = [line.strip().lower() for line in f]
    wordset = set(words)

    # Find candidates that match pat1.
    pat1 = re.compile(pat1)
    candidates = [w for w in wordset if pat1.fullmatch(w)]

    # Find complete matches.
    for c in candidates:
        replaced = replace_pattern(c, pat2)
        if replaced in wordset:
            print(f'{c}, {replaced}')


# ______________________________________________________________________
# Main

if __name__ == '__main__':
    
    if len(sys.argv) < 2:
        print(__doc__)
        sys.exit(0)

    pat1, pat2 = parse_patterns(sys.argv[1])
    find_matches(pat1, pat2)