26 lines
5.5 KiB
Python
26 lines
5.5 KiB
Python
|
import numpy as np
|
||
|
|
transposes = ['GADSJFEDVAWFYJSGLSGFVELTSWZGKNYGWGKWFZWZJFJZFDLFJKFLLJLESJUVFHMJWKJNWAVSKSHWSWHGWEZWGATVDJLMWWWWODAUUJFTSDFFLSWAAEUFKJQSKJFGYWAAVFVLSSVLJWTVSWJLJFFFZJWLQSAGQFAZEDFSZGDFVSKEJXDLFOZGVKHNGETLLKGKWHJLZLJEXGWDHWAGLJGGOKSLHMSXWAVWGZWEWAAJVSJGWYXUAKXYZVXJGYFDGXMKFASFAAHGDAHYJJZWXWWGWXJZWLHVLDIFJWJWLGALFEWWMAFKZLAWLAZGVHXQSKGZJLLDVWFWAYWFZFVWYFSDDDXMWWAWQQLXZGFYFAVBSALVJDAJGELLGDW', 'BUSYARSCSBFUSOAAHHFCWSVSFFCCVOVAOFHFHSBSSGIDHZOSSHUCWSFPHSSCRFZVBOISPKGACAWFFOSCFCOFRBHFSSVBGBCGVSGVYSROAZAOOHTRGMSKOMVAASRDHBFBOCOVBACKRRSIZADOSMUHSBQKQBGQUHBHSMWBSFZZWWOSSCDBPOGJCHHSCSIHVDBGKCHSSSQIFKASSTFFVCOIVWBTDBGORGSRDCCCBBBOOHRCJWHMBCMFOCMCFOUSHWFCRBRWCBZSJBZCSSHBFBFCRZWZGVWBVSISUBODVDFVHDSGBGQVOFGPSASFWFPHHPHWFVOQAOCORSEBOCCYUHBISIHZOGHAPVVASFBGUBPIAGVSSOACBMVZBZF', 'CHIAACORDGEANPBEEMOTSMERAWSREITMSRLAMAORAASUHESVSIOOSNAEEAAONEGENILNUAOEFENDCRNRERTITGIEVAEBTONPINPOTMNCBMEGPLOISWLHNEEOYTNEHWTTSFYECIRIOARPLYENTBMOGDOORDARRRGLWNGSREEYSNNALRAOESEEOUUHRATIEOEOOUHRRRETINOASODEEKPBATDRYMTSFSNIELFRSGTICIAREBOUKMOISMOAEGBEHENMMIANNGIREGIESTTSOSATOOEEESTEYNAPEOVRIHDEOEHODDHOUUTALPRETOIHBYHSOIILANRNELUETFREEODTAMHHVSTYEEYYREETSGUSBEEMAMSWTSAITBM', 'YNJHHOZYHMWMNJYPLSHBNVVUNLNYXHNSYIYLUHLYHXNLLXNYIFZXMNHLPNNLIMLMIXSYNMALUNAIYXYXUYXHBZHUYGMLCNFIMIYGBYIEYSVUJILMICYUXRUGMCIHYCUBNNMFYHMNLVXIUWLXBCSMLYOLYMCUCUZSBUBBUGXZWFMHYQAFCVYBLJLCQMNHJEQOLLYYNYFNYVLPBLMMMYNNNMMIGYYNUIVLNSHYNUIAOHHNFYFHCCLGNCLENYORYLYMIHNAMNAVNQALMBBYGYJZHIXHUYYJGIZYNLYIHYILLMYFUYUGHFBFFFUJBJLUYNUMQHXUCAYXHMIPQJZXNNNIPYUUYOBXUUZXUUPCNIMNYSMIGJNMBINYBYI', 'MVGVQJXOWMSGDAJAWUDAAWDFWGMSLDGTGOFWAYWXVMDHAEAJLDEJGJUVAAELLWWADAQKLFWSASSGOQVSFVSYWWYEJTAGDGQCHJJMWJLAJKMAAMWMEFLLLHJWLKLZKLFWSZGWEMLZDGGFKZUFWJKESUFWKZVNENJKSELGNMLGGQOANWJAFDAATGWKAFZYDWGDVFJVZVQWFWJWSWSLLFDDAGLESJJSKFGYZTWTADKZKXVZNLALFFWMDFWAZVLHXQVUJYWGNZZMNAZKSWWJSJZSLJLLFFJWWJXFLWFHYLJLOLJSMKFWLQWEEGNJAZVLFZVGDLWKVWUJOFLWGSAMLZZFWSLLEFWGCJGGNFWDAFLSJWWFADJZWMKKWDJ', 'CEAJNVBUNQZSLAQFQULAAERQQHCTULOBSSBNQRABGAVYYRPSUGLRZRRBFALUUAJGBBBFUNACACGENBJEQREYENQFQRYXYXJRRRREJRUAGBEAAQFERQZGUYGARGURUUQEGRSNNGNZNIBNWNUBAQNVIBGGGNNRNRBUGRFERPUJHGRANPRITRAFVAQPGRRYNBEVUBGARGZEQSBZIGVNVOLRGABFFPSAGTERNHIHYYZGUEORRBABTBJALBZAVVAEBRVBRJNARNGGVGGUUANCAFVYUJUGTGRAZRGGUDAUBFQRUGRGARGORVEVRERBARBUQNBHNUACRYYNUNUEEERCURRVAFYULOOBSGEBRQEYYGBOQFRFAVRVSYUSSVR', 'OEDEQORIRDESRGOITATGCIAEYGOEEIOOSORNELARHCNELWTEAHHPEAAOIGCIITSANRUIEPTPTPMTSUIKNEKOERRNAFEENEONRTDMOLITUUNIGEUETOEHIOBTRHIRUMFEESYSDEYIDERBUMETTBDLEREHBVRNNNMOTIPENHILRHRGNAEIESGCRTBHHVROCNDNETHONIOESOWEEHDRLYSSULROTIODESESTRETLMIIIOUNTOKFWUHGGUEGSNOEWYNRIISTLTGWOHGENMIEUWMLERYHETSTOQHHIUEEFTEMEOAETREYDIENTENPGTRESTRLDENANSADOMTMDTNSETNAOAISLRURRAMRNTFILHFOOHMTGGASLFALLF', 'JWAWQQCHHAWKWWKPWIDIPPGWEDJNIDNGKNAZJOIAAACYAEAHPAAARPPNPADOJHPPCINEBLHEHEUDODZJKLJJSEAKNKJJAJNSADWQNUJKNHEDONNHDSOAONAWUAJAPWHOHWKPAOAAUILQOXZDDABEWQJKAAPCZSPNDOHMJEOPOKHHYJEJROXDZDQWOAWJEHWPZDAPDHNZDNDWBAJPHNLOPUAIABHBNKPKIZNPXUHSKJOQOGEWDOWWWOWJECOOHAPAOPADRPHDHPHOAANNJQSOPACAHDLDNQEAOKRPAERLPOOUAPZDPIECAMJDKOZWWCASAZEOSJOEIADKKEZPAAEJXPALKKOPKJBMJDHOEALRKWEDWDIOKNZKKP', 'JAAAJBXNHWCVYCAXYKRXUFMLWCCUXJABACOAFWNEBAXDMCACWKJJRRVBXWJRPHARNJOVJRHWHWLJDNNNCRNPXWJVNALJBJMJMRWAMCPCWFWNXCNHRUNAVNBWNFPRCWDCHRAXQCMWYHNBCNJRRPJWWVJDBWBQJJQNHXDDNVDXNDRRHWWPNNRJXNBVDAENMHBQRRWJNUNXJNNBUWNUWNXBCBLNADUXCWRONNWQNORQWCCYRVWWJKCRDKWNBDHBFBXCJQANNQXXNQXQECPOBWQCDCXNBNRNRJBJUCNBEURCQBQNMUXXNYBRUDNNOCNEKXCRWRCJQJYJCUNADWRJKVPBUXCXWTCJVMADNNRBBYJNAEWJWCRQXXXJXN', 'VHNHRECRWUJNYVCAYNBAHJNQPRQHAMXBAQXJQJUNRCOAVQAKXNACBWHXACVCVVXWAMXYLWHPHPQCAACBQWBRWPVXMNNWBWCBFBMNUQVQRRPJVQUCWJNNHUCMGRVONHCNAWNKNXQXNLACJAWWBDWPMXWPQCDJWWNCUWCXEJWQBPCCKXPQAMAVABCKLVNUKCRJMWQOORCCEXFVXCENNYTJNCJDFUXURNUQUWNNPJWNNXJXWHPLCRCWWRCEJNURQWVQCVNLCNJBCNJJNQADNPXROLMKQNCONOTWXQAJRUUNNNXCXJWAUUCUUXECNRENXMNCRBBRXVJWQNANAPBAJYQQJTQTNNJTXCXXEACRBJUVJNPCMXWJAVFCAM', 'IWDONOTNAYPOIYIIIRAEISCSETEIRSWUOERDOMERLAETEFOEWATIIGCMERBIOYNGSARLTGOAOTAIEEHSITSSDDIRTBWDGDHTOIADEIOENTSREAYHGTTASEITPNOLRATPATNENPEFRHCOBDDGEICBSFCHOHRSCDNEOTOTERGEOHTTOTTUYWDBBCAEHONYUHFTOGEELSHHENIYWHEDSLEIROUNHDWLLBLIABVRUNGEDFNNKSFYTRHLTRIETSLNOOYIDYCULLTELLTLROEMEBSNTHHYERERSFIDSTMILI
|
||
|
|
|
||
|
|
def compute_frequencies(block):
|
||
|
|
letter_count = {}
|
||
|
|
letters = [chr(i) for i in range(ord('A'), ord('Z')+1)]
|
||
|
|
for l in letters:
|
||
|
|
letter_count[l] = block.count(l)
|
||
|
|
return letter_count
|
||
|
|
|
||
|
|
def coincidence(block):
|
||
|
|
compute_frequencies(block)
|
||
|
|
freq = compute_frequencies(block)
|
||
|
|
sum = 0
|
||
|
|
for f in freq:
|
||
|
|
sum += freq[f] * (freq[f]-1)
|
||
|
|
kappa = float(sum)/(len(block) * (len(block)-1))
|
||
|
|
return kappa
|
||
|
|
for t in transposes:
|
||
|
|
print("kappa", coincidence(t))
|
||
|
|
|
||
|
|
for t in transposes:
|
||
|
|
d = compute_frequencies(t)
|
||
|
|
|
||
|
|
print(chr((ord('E') - ord(d.keys()[np.argmax(d.values())]))%26+65))
|