Source code for tamil.tweetparser

# This Python file uses the following encoding: utf-8
# (C) 2013-2016 Muthiah Annamalai
# 
# This file is now part of open-tamil project
# 
# N.B.: Originally written to demo the Python Twitter API for a lightning talk
# at the Boston Python group on 30th July, 2013.

import unicodedata
import re

class TweetParser:
    def __init__(self,timeline_owner,tweet):
        self.Owner = timeline_owner;
        self.tweet = tweet;
        self.RT = False;         
        self.tweet = tweet;
        self.UserHandles = TweetParser.getUserHandles(tweet);
        self.Hashtags = TweetParser.getHashtags(tweet);
        self.URLs = TweetParser.getURLs(tweet);
        self.RT = TweetParser.getAttributeRT(tweet);
        self.MT = TweetParser.getAttributeMT(tweet);
        
        # additional intelligence
        if ( self.RT and len(self.UserHandles) > 0 ): #change the owner of tweet?
            self.Owner = self.UserHandles[0];
        
        return

    def __str__(self):
        return "owner %s, urls: %d, hashtags %d, user_handles %d, len_tweet %d, RT = %s, MT = %s"%(self.Owner,len(self.URLs),len(self.Hashtags),len(self.UserHandles), len(self.tweet), self.RT,self.MT)

    @staticmethod
    def getAttributeRT( tweet ):
        """ see if tweet is a RT """
        return re.search(r'^RT',tweet.strip()) != None

    @staticmethod
    def getAttributeMT( tweet ):
        """ see if tweet is a MT """
        return re.search(r'^MT',tweet.strip()) != None
    
    @staticmethod
    def getUserHandles( tweet ):
        """ given a tweet we try and extract all user handles in order of occurrence"""
        return re.findall(r'(@[a-zA-Z0-9_]+)',tweet);

    @staticmethod
    def getHashtags( tweet ):
        """ return all hashtags"""
        return re.findall(r'(#[\w\d]+)',tweet);

    @staticmethod
    def getURLs( tweet ):
        """ URL : [http://]?[\w\.?/]+"""
        return re.findall(r'([http://]?[a-zA-Z\d\/]+[\.]+[a-zA-Z\d\/\.]+)',tweet);
    


class TamilTweetParser( TweetParser ):
    def __init__(self,timeline_owner,tweet):
        TweetParser.__init__(self,timeline_owner,tweet)
        self.TAWords = TamilTweetParser.getTamilWords(tweet);
        return
    
    def __str__(self):
        """ tack on the parent """
        return TweetParser.__str__(self) + ", TA words = %d "%len(self.TAWords)
    
    @staticmethod
    def isTamilPredicate(word):
        """ is Tamil word : boolean True/False"""
        for c in word:
            if unicodedata.name(c).split()[0] != u'TAMIL' :
                return False
        return True

    @staticmethod
    def cleanupPunct( tweet ):
        """ NonEnglishOrTamilOr """
        tweet = ''.join( map( lambda c: (unicodedata.name(c).split()[0] in [u'TAMIL',u'LATIN']) and c or u' ', tweet) )
        return tweet

    @staticmethod
    def getTamilWords( tweet ):
        """" word needs to all be in the same tamil language """
        tweet = TamilTweetParser.cleanupPunct( tweet );
        nonETwords = filter( lambda x: len(x) > 0 , re.split(r'\s+',tweet) );#|"+|\'+|#+
        tamilWords = filter( TamilTweetParser.isTamilPredicate, nonETwords );
        return tamilWords