X-Git-Url: https://git.njae.me.uk/?a=blobdiff_plain;f=language_models.py;h=babbea19ceed80bd9d95fee347c64e53256b5626;hb=32a4467e6f7ac8ff2e6738118242ec4e4c255e8a;hp=929746888d036fb54de3f1fbf228e296e0bcd027;hpb=3e8d2bd8cd7c623116fa3d2b77db954f51b191e4;p=cipher-training.git

diff --git a/language_models.py b/language_models.py
index 9297468..babbea1 100644
--- a/language_models.py
+++ b/language_models.py
@@ -1,6 +1,10 @@
+"""Language-specific functions, including models of languages based on data of
+its use.
+"""
+
 import string
-import norms
 import random
+import norms
 import collections
 import unicodedata
 import itertools
@@ -16,7 +20,7 @@ def letters(text):
     return ''.join([c for c in text if c in string.ascii_letters])
 
 def unaccent(text):
-    """Remove all accents from letters. 
+    """Remove all accents from letters.
     It does this by converting the unicode string to decomposed compatability
     form, dropping all the combining accents, then re-encoding the bytes.
 
@@ -37,7 +41,7 @@ def unaccent(text):
 
 def sanitise(text):
     """Remove all non-alphabetic characters and convert the text to lowercase
-    
+
     >>> sanitise('The Quick')
     'thequick'
     >>> sanitise('The Quick BROWN fox jumped! over... the (9lazy) DOG')
@@ -72,20 +76,20 @@ with open('words.txt', 'r') as f:
 
 
 def weighted_choice(d):
-	"""Generate random item from a dictionary of item counts
-	"""
-	target = random.uniform(0, sum(d.values()))
-	cuml = 0.0
-	for (l, p) in d.items():
-		cuml += p
-		if cuml > target:
-			return l
-	return None
+    """Generate random item from a dictionary of item counts
+    """
+    target = random.uniform(0, sum(d.values()))
+    cuml = 0.0
+    for (l, p) in d.items():
+        cuml += p
+        if cuml > target:
+            return l
+    return None
 
 def random_english_letter():
-	"""Generate a random letter based on English letter counts
-	"""
-	return weighted_choice(normalised_english_counts)
+    """Generate a random letter based on English letter counts
+    """
+    return weighted_choice(normalised_english_counts)
 
 
 def ngrams(text, n):
@@ -120,14 +124,22 @@ def log_probability_of_unknown_word(key, N):
     return -log10(N * 10**((len(key) - 2) * 1.4))
 
 Pw = Pdist(datafile('count_1w.txt'), log_probability_of_unknown_word)
+Pw_wrong = Pdist(datafile('count_1w.txt'), lambda _k, N: log10(1/N))
 Pl = Pdist(datafile('count_1l.txt'), lambda _k, _N: 0)
 P2l = Pdist(datafile('count_2l.txt'), lambda _k, _N: 0)
+P3l = Pdist(datafile('count_3l.txt'), lambda _k, _N: 0)
 
 def Pwords(words): 
     """The Naive Bayes log probability of a sequence of words.
     """
     return sum(Pw[w.lower()] for w in words)
 
+def Pwords_wrong(words): 
+    """The Naive Bayes log probability of a sequence of words.
+    """
+    return sum(Pw_wrong[w.lower()] for w in words)
+
+
 def Pletters(letters):
     """The Naive Bayes log probability of a sequence of letters.
     """
@@ -139,15 +151,21 @@ def Pbigrams(letters):
     """
     return sum(P2l[p] for p in ngrams(letters, 2))
 
+def Ptrigrams(letters):
+    """The Naive Bayes log probability of the trigrams formed from a sequence 
+    of letters.
+    """
+    return sum(P3l[p] for p in ngrams(letters, 3))
+
 
-def cosine_distance_score(text):
+def cosine_similarity_score(text):
     """Finds the dissimilarity of a text to English, using the cosine distance
     of the frequency distribution.
 
-    >>> cosine_distance_score('abcabc') # doctest: +ELLIPSIS
-    0.370847405...
+    >>> cosine_similarity_score('abcabc') # doctest: +ELLIPSIS
+    0.26228882...
     """
-    return norms.cosine_distance(english_counts, 
+    return norms.cosine_similarity(english_counts, 
         collections.Counter(sanitise(text)))