Where to start for a cryptogram?

[Thekid]

Can I get some pointers on writing a program to decipher cryptograms? I'm not sure where to start. This is an example:

bthhy gwtqt! wyo pl cyxq zkv eypre?

[ilikepython]

Can I get some pointers on writing a program to decipher cryptograms? I'm not sure where to start. This is an example:

bthhy gwtqt! wyo pl cyxq zkv eypre?

I think you should start off with a huge word list in a file so python can know what to look for but i'm not sure. Just an idea.

[ilikepython]

Wait do you mean like a word scrambler/unscrambler or a kind of like a decoder that tries to decode the message??

[Thekid]

A decoder. I have written a word unscrambler but I like to do cryptograms from a crossword puzzle book and thought it would be fun to try to write a program that could decipher them.

[bartonc]

A decoder. I have written a word unscrambler but I like to do cryptograms from a crossword puzzle book and thought it would be fun to try to write a program that could decipher them.

After unscrambling a word you could do something like:

import win32com.client
 
msword = win32com.client.Dispatch("Word.Application")
 
string = "Here is some texxt with a coupple of problems"
 
for word in string.split():
    if msword.CheckSpelling(word):
       print "!%s! OK!" % word
    else:
       print "?%s? ->"%word,
       suggestions = msword.GetSpellingSuggestions(word)
       for suggest in suggestions:
           print suggest.Name,
       print
 

on Windows or check with ispell using popen() on *nix.

[Thekid]

Ya, I don't need to unscramble the words first but I have written a simple word unscrambler before. I was wondering where to start to write a program to solve a cryptogram without a key. So if:
a = "h"
b = "i"
c = "j"
d = "x"
e = "t"
f = "r"
g = "e"

and the puzzle was: ab eagrg!
the solution would be: hi there!

BUT.....how can it be done without having a key to the solution? Cryptograms are fun and I can usually solve them with paper and pencil but I figured I'd give myself a challenge and see if I could come up with a program that could do it.

[bartonc]

Ya, I don't need to unscramble the words first but I have written a simple word unscrambler before. I was wondering where to start to write a program to solve a cryptogram without a key. So if:
a = "h"
b = "i"
c = "j"
d = "x"
e = "t"
f = "r"
g = "e"

and the puzzle was: ab eagrg!
the solution would be: hi there!

BUT.....how can it be done without having a key to the solution? Cryptograms are fun and I can usually solve them with paper and pencil but I figured I'd give myself a challenge and see if I could come up with a program that could do it.

That's a good one. For a given problem that you know how to solve with paper and pencil, start by writing down the steps you take with the given info, what new info you have after step 1, and so on, until the entire algorithm unfolds. Once you have a possible algorithm, then it's time to ask "how can I make the computer perform this algorithm?".

[Thekid]

I found the following at:
http://mail.python.org/pipermail/python-list/2002-September/165558.html

"""
 Ha'f ahoy ad ayfa wdqr srxhekdpyr! H phmm zhby
 fhv odeajf dc cryy pys jdfahez ad ajy chrfa kyrfde
 ad fdmby ajhf irwkadzrxo. Ixe wdq chzqry ha dqa?
 
 Fheiy wdq jxby axgye ow ijxmmyezy xet tyirwkayt
 ajy fyirya oyffxzy, H phmm ybye zhby wdq ajy
 irwkadzrxo-oxghez fdqriy idty ad xoxuy wdqr
 crhyetf!
 """
 hokdra fwf, farhez, rxetdo
 ihkjyr = mhfa(farhez.myaayrf[:26])
 rxetdo.fjqccmy(ihkjyr)
 ihkjyr = "".ldhe(ihkjyr)
 ihkjyr += ihkjyr.qkkyr()
 oxk = {}
 cdr kmxhe, irwka he uhk(farhez.myaayrf, ihkjyr):
     oxk[kmxhe] = irwka
 cdr mhey he dkye(fwf.xrzb[1]):
     krhea "".ldhe([oxk.zya(ij,ij) cdr ij he mhey])[:-1]
 
 

Now if I was to start thinking this one out, I'd start with single letters which are usually either 'A' or 'I' and also take a look at any words with an appostrophe, then I'd look for possible 'and' or 'the' type structures, double letters,and any questions, which usually start with: is, am, are, who, what, where, when, how, why, can, can't, do, don't, may, will, won't, etc. but again the appostrophe comes in to play as well as the amount of letters in the word.
I have the solution to this one but wouldn't find it hard to start myself. Here's someones solution to it:

def decrypt(yourstring):
    dedict = {'F': 'S', 'I': 'C', 'H': 'I', 'a': 't', 'c': 'f', 'b': 'v', 'e': 'n',
    'd': 'o', 'f': 's', 'i': 'c', 'h': 'i', 'k': 'p', 'j': 'h', 'm': 'l',
    'l': 'j', 'o': 'm', 'q': 'u', 'p': 'w', 's': 'b', 'u': 'z', 't': 'd',
    'w': 'y', 'v': 'x', 'y': 'e', 'x': 'a', 'z': 'g', 'g': 'k'}
    return "".join([dedict.get(i,i) for i in yourstring])
 
returns 
 
    """
    It's time to test your brainpower! I will give
    six months of free web hosting to the first person
    to solve this cryptogram. Can you figure it out?
 
    Since you have taken my challenge and decrypted
    the secret message, I will even give you the
    cryptogram-making source code to amaze your
    friends!
    """
    import sys, string, random
    cipher = list(string.letters[:26])
    random.shuffle(cipher)
    cipher = "".join(cipher)
    cipher += cipher.upper()
    map = {}
    for plain, crypt in zip(string.letters, cipher):
        map[plain] = crypt
    for line in open(sys.argv[1]):
        print "".join([map.get(ch,ch) for ch in line])[:-1]
 

This was posted in 2002 but I didn't know if some of this code could be used without the given key.