A tool to solve cryptograms, specifically the ones in Clever Cryptograms by Louise B. Moll.
An example of installation and usage might look like:
$ git clone https://github.com/isaacazuelos/cryptogram
$ cd cryptogram
$ stack install
$ cryptogram "ZYHQ HQ XS LDXNURL UCOORL" -d ./sample-dict.txt | head -n 1
key: X***L**YH**RNS*U**QZC**D*O message: THIS IS AN EXAMPLE PUZZLE
It solves cryptograms. By cryptogram, we mean a specific type of substitution cypher.
A substitution cypher is when each glyph is swapped with another, and the mapping used is the same for all glyphs across the entire encoded message.
In our case, we only care about uppercase English alphabet letters. Any other glyphs are passed over unchanged. Unlike Louise B. Moll, we don't assume a no identity rule.
Most of this is handled by stack
.
This includes development stuff like dependency fetching and test running.
Usage: cryptogram [-e|--encrypt] [-k|--key ARG] MESSAGE [-d|--dictionary ARG]
To encrypt some text, pass the message in with the --encrypt
flag set, otherwise decryption is assumed.
A key can be specified. When encrypting without a key, one is randomly generated. Decryption without a key will print out all solutions which decode the valid-seeming words in the text.
Specifying a dictionary file assumes a dictionary with whitespace-delimited words. You can get a decent dictionary for simple messages from here. By default it uses the dictionary at /usr/share/dict/words
, which is awful for this.
The dictionary you use matters a lot.
The default dictionary for me in OS X 10.11 has too many low probability words, and is sorted lexicographically, so it is truly awful for this purpose. Feeding it a word or two from a puzzle is a decent way to get a hint though.
This program preserves the order of the words in the dictionary when searching. That means it will try "cat" before "the" using the default dictionary, if both are valid options.
If you use a version of the suggested dictionary from first20hours, it will put more likely solutions first. If you're going to use it, you'll want remove a bunch of the nonsense short words (it only takes a few minutes in a REPL.)
I'd like to thank Edwin Olson's work at http://www.quipqiup.com/howwork.php and his paper on the topic for helping me think through how to make this work.
It's MIT Licensed. See the included LICENSE
file.