1) Add Emoticons parsing to the Tweet API.

2) Add tests for emoticons in parse_tweet feed 3) Update Authors to reflect this contribution and original parse_tweet API
bear · 8279bade1f42da60c492751fd6dcbeaaf71bc212 · Muthiah Annamalai committed Mar 7, 2015 · Mar 7, 2015 · 8279bad
1 parent b21bc03
commit 8279bade1f42da60c492751fd6dcbeaaf71bc212
Unified Split

Showing with 56 additions and 3 deletions.

+2 −1 AUTHORS.rst

+40 −2 twitter/parse_tweet.py

+14 −0 twitter_test.py
@@ -25,6 +25,7 @@ Now it's a full-on open source project with many contributors over time:
 * Lars Weiler,
 * Sebastian Wiesinger,
 * Jake Robinson,
+* Muthu Annamalai,
 * abloch,
 * cahlan,
 * dpslwk,
@@ -33,4 +34,4 @@ Now it's a full-on open source project with many contributors over time:
 * git-matrix,
 * sbywater,
 * thefinn93,
-* themylogin,
+* themylogin,
@@ -4,11 +4,36 @@
 
 from twitter import TwitterError  # import not used?
 
+class Emoticons:
+    POSITIVE = ["*O","*-*","*O*","*o*","* *",
+                ":P",":D",":d",":p",
+                ";P",";D",";d",";p",
+                ":-)",";-)",":=)",";=)",
+                ":<)",":>)",";>)",";=)",
+                "=}",":)","(:;)",
+                "(;",":}","{:",";}",
+                "{;:]",
+                "[;",":')",";')",":-3",
+                "{;",":]",
+                ";-3",":-x",";-x",":-X",
+                ";-X",":-}",";-=}",":-]",
+                ";-]",":-.)",
+                "^_^","^-^"]
+
+    NEGATIVE = [":(",";(",":'(",
+                "=(","={","):",");",
+                ")':",")';",")=","}=",
+                ";-{{",";-{",":-{{",":-{",
+                ":-(",";-(",
+                ":,)",":'{",
+                "[:",";]"
+                ]
 
 class ParseTweet:
     # compile once on import
     regexp = {"RT": "^RT", "MT": r"^MT", "ALNUM": r"(@[a-zA-Z0-9_]+)",
-              "HASHTAG": r"(#[\w\d]+)", "URL": r"([http://]?[a-zA-Z\d\/]+[\.]+[a-zA-Z\d\/\.]+)"}
+              "HASHTAG": r"(#[\w\d]+)", "URL": r"([https://|http://]?[a-zA-Z\d\/]+[\.]+[a-zA-Z\d\/\.]+)",
+              "SPACES":r"\s+"}
     regexp = dict((key, re.compile(value)) for key, value in regexp.items())
 
     def __init__(self, timeline_owner, tweet):
@@ -25,7 +50,8 @@ def __init__(self, timeline_owner, tweet):
         self.URLs = ParseTweet.getURLs(tweet)
         self.RT = ParseTweet.getAttributeRT(tweet)
         self.MT = ParseTweet.getAttributeMT(tweet)
-
+        self.Emoticon = ParseTweet.getAttributeEmoticon(tweet)
+        
         # additional intelligence
         if ( self.RT and len(self.UserHandles) > 0 ):  # change the owner of tweet?
             self.Owner = self.UserHandles[0]
@@ -37,6 +63,18 @@ def __str__(self):
         self.Owner, len(self.URLs), len(self.Hashtags), len(self.UserHandles), len(self.tweet), self.RT, self.MT)
 
     @staticmethod
+    def getAttributeEmoticon(tweet):
+        """ see if tweet is contains any emoticons, +ve, -ve or neutral """
+        emoji = list()
+        for tok in re.split(ParseTweet.regexp["SPACES"],tweet.strip()):
+            if tok in Emoticons.POSITIVE:
+                emoji.append( tok )
+                continue
+            if tok in Emoticons.NEGATIVE:
+                emoji.append( tok )
+        return emoji
+    
+    @staticmethod
     def getAttributeRT(tweet):
         """ see if tweet is a RT """
         return re.search(ParseTweet.regexp["RT"], tweet.strip()) != None
@@ -651,15 +651,29 @@ def testParseTweets(self):
 
         data = twitter.ParseTweet("@twitter", hashtag_n_URL)
         self.assertEqual([len(data.Hashtags), len(data.URLs)], [2, 1])
+        self.assertEqual(len(data.Emoticon),0)
 
         url_only = u"""The #Rainbow #Nebula, 544,667 #lightyears away. pic.twitter.com/2A4wSUK25A""";
         data = twitter.ParseTweet("@twitter", url_only)
         self.assertEqual([data.MT, len(data.Hashtags), len(data.URLs)], [False, 3, 1])
+        self.assertEqual(len(data.Emoticon),0)
 
         url_handle = u"""RT ‏@BarackObama POTUS recommends Python-Twitter #unrelated picture pic.twitter.com/w8lFIfuUmI""";
         data = twitter.ParseTweet("@twitter", url_handle)
         self.assertEqual([data.RT, len(data.Hashtags), len(data.URLs), len(data.UserHandles)], [True, 1, 1, 1])
+        self.assertEqual(len(data.Emoticon),0)
 
+    def testEmoticon(self):
+        url_handle = u"""RT ‏@BarackObama POTUS recommends :-) Python-Twitter #unrelated picture pic.twitter.com/w8lFIfuUmI""";
+        data = twitter.ParseTweet("@twitter", url_handle)
+        self.assertEqual([data.RT, len(data.Hashtags), len(data.URLs), len(data.UserHandles)], [True, 1, 1, 1])
+        self.assertEqual(len(data.Emoticon),1)
+
+        url_handle = u"""RT @cats ^-^ cute! But kitty litter :-( #unrelated picture""";
+        data = twitter.ParseTweet("@cats", url_handle)
+        self.assertEqual([data.RT, len(data.Hashtags), len(data.URLs), len(data.UserHandles)], [True, 1, 0, 1])
+        self.assertEqual(len(data.Emoticon),2)
+        self.assertEqual(data.Emoticon,['^-^',':-('])
 
 class MockHTTPBasicAuthHandler(object):
     '''A mock replacement for HTTPBasicAuthHandler'''