def compressionRatio(self,encodingWordUsage, compressionUsage,tree = None):
# the first two arguments can either be speeches or dictionaries with words and frequencies
if(isinstance(encodingWordUsage,speech.Speech)):
encodingWordUsage = encodingWordUsage.wordUsage
if(isinstance(compressionUsage,speech.Speech)):
compressionUsage = compressionUsage.wordUsage
#tree can be provided, to help performance, otherwise make a tree
if(tree == None):
tree = huffmanCoding.huffmanCodingTree(encodingWordUsage)
wordCount = 0
bitsInHuffmanCoding = 0
#note, don't use the additive smoothing for finding ratio
for entry in compressionUsage.iteritems():
#bits needed dictionary has the number of bits that each node of tree would encode. This could
#be found using tree.findNode(), but looking things up in the tree took a long time, relative
#to the number of times that it has to be done.
bitsInHuffmanCoding += (entry[1]-1)*tree.bitsNeededDictionary[entry[0]]
wordCount += entry[1] - 1
bitsInBlockCoding = wordCount*int(math.ceil(math.log(len(compressionUsage),2)))
#for debugging
#print "huffmanCoding = "+str(bitsInHuffmanCoding)+" blockCoding: "+str(bitsInBlockCoding)
if(bitsInBlockCoding > 0):
return float(bitsInHuffmanCoding)/float(bitsInBlockCoding)
else:
return 0
if(len(sys.argv) >= 3):
print "finding speech set"
set = speechSet.SpeechSet(sys.argv[1])
reverse = False
if(len(sys.argv) == 4):
reverse = True
print "finding all filenames"
if(reverse):#note that reversing the list puts it in chronological non-reverse
listRecentFilenames = sorted(glob.glob(set.directoryPath+"/"+set.fileType),reverse = False)
else:
listRecentFilenames = sorted(glob.glob(set.directoryPath+"/"+set.fileType),reverse = True)
print "finding word usage"
wordUsage = determineWordUsage(set.wordUsage,int(sys.argv[2]))
wordsSorted = sorted(wordUsage.iteritems(),key = operator.itemgetter(1))
print "start finding coding tree "+str(time.clock())
tree = huffmanCoding.huffmanCodingTree(wordUsage)
print "bits of the longest codeword:"+str(wordsSorted[0])+", "+str(tree.bitsNeededDictionary[wordsSorted[0][0]])
print "bits of the shortest codeword:"+str(wordsSorted[len(wordsSorted)-1])+", "+str(tree.bitsNeededDictionary[wordsSorted[len(wordsSorted)-1][0]])
#print "bit the "+ str(tree.bitsNeededDictionary["the"])
#print "bit a " + str(tree.bitsNeededDictionary["a"])
#print "bit for " + str(tree.bitsNeededDictionary["for"])
print "completed finding coding tree "+str(time.clock())
file = None
fileForWordCount = open("WordCount.txt","w+")
if(reverse):
file = open(sys.argv[2]+"LeastRecentSpeech.txt","w+")
else:
file = open(sys.argv[2]+"MostRecentSpeech.txt","w+")
print "starting compression ratio, "+str(time.clock())
for filename in listRecentFilenames: