Classification.py

'''Now given a mail, split it in terms of spaces  ,  then ,  add up the log probability of each .  Multiply it with the spam probability  . Do the same thing for non-spam
   Whichever is higher  wins .  Lets start
'''
import sys,os
from math import *
def makeDict(f):
    temp = {}
    data = open(f,'r')
    for line in data:
        prob = line.split(" ")
        temp[prob[0]] = prob[1]
    return temp
def predict(basepath,f):
    toClassify = open(os.path.join(basepath,f),'r')
    for line in toClassify:
        words = line.split(" ")
        #print 'words are',words
        spamP = 0
        nonspamP = 0
        for w in words:
            try:
                spamP = spamP + float(spamProbs[w].strip("\n"))
            except:
                continue

            try:
                nonspamP = nonspamP + float(nonspamProbs[w].strip("\n"))
            except:
                continue

    totalSpamP = spamP + log ( 0.5 )
    totalnonSpamP = nonspamP + log ( 0.5 )
    #print 'TOtal spam and non-spam probs are ',totalSpamP,totalnonSpamP
    if(totalSpamP > totalnonSpamP):
        return True
    else:
        return False

spamProbs = makeDict(sys.argv[1]) #Pass the spam log probs here
nonspamProbs = makeDict(sys.argv[2]) #Pass the non-spam log probs here
#print spamProbs
spamCount = 0
nonspamCount = 0
print 'No of files in spam is',len(os.listdir(sys.argv[3]))
for f in os.listdir(sys.argv[3]):

    if(predict(sys.argv[3],f)):
        spamCount = spamCount + 1
    else:
        nonspamCount = nonspamCount + 1
print 'No. of spam in ',sys.argv[3],' is ',str(spamCount),' no. of non-spam',str(nonspamCount)
print 'No of files in non-spam is',len(os.listdir(sys.argv[4]))
spamCount = 0
nonspamCount =0
for f in os.listdir(sys.argv[4]):

    if(predict(sys.argv[4],f)):
        spamCount = spamCount + 1
    else:
        nonspamCount = nonspamCount + 1
print 'No. of spam in ',sys.argv[4],' is ',spamCount,' no. of non-spam',nonspamCount