Python version of Ruby script to preprocess tweets for use in GloVe featurization http://nlp.stanford.edu/projects/glove/

preprocess-twitter.py

"""
preprocess-twitter.py

python preprocess-twitter.py "Some random text with #hashtags, @mentions and http://t.co/kdjfkdjf (links). :)"

Script for preprocessing tweets by Romain Paulus
with small modifications by Jeffrey Pennington
with translation to Python by Motoki Wu

Translation of Ruby script to create features for GloVe vectors for Twitter data.
http://nlp.stanford.edu/projects/glove/preprocess-twitter.rb
"""

import sys
import re

FLAGS = re.MULTILINE | re.DOTALL

def hashtag(text):
    text = text.group()
    hashtag_body = text[1:]
    if hashtag_body.isupper():
        result = "<hashtag> {} <allcaps>".format(hashtag_body)
    else:
        result = " ".join(["<hashtag>"] + re.split(r"(?=[A-Z])", hashtag_body, flags=FLAGS))
    return result

def allcaps(text):
    text = text.group()
    return text.lower() + " <allcaps>"


def tokenize(text):
    # Different regex parts for smiley faces
    eyes = r"[8:=;]"
    nose = r"['`\-]?"

    # function so code less repetitive
    def re_sub(pattern, repl):
        return re.sub(pattern, repl, text, flags=FLAGS)

    text = re_sub(r"https?:\/\/\S+\b|www\.(\w+\.)+\S*", "<url>")
    text = re_sub(r"/"," / ")
    text = re_sub(r"@\w+", "<user>")
    text = re_sub(r"{}{}[)dD]+|[)dD]+{}{}".format(eyes, nose, nose, eyes), "<smile>")
    text = re_sub(r"{}{}p+".format(eyes, nose), "<lolface>")
    text = re_sub(r"{}{}\(+|\)+{}{}".format(eyes, nose, nose, eyes), "<sadface>")
    text = re_sub(r"{}{}[\/|l*]".format(eyes, nose), "<neutralface>")
    text = re_sub(r"<3","<heart>")
    text = re_sub(r"[-+]?[.\d]*[\d]+[:,.\d]*", "<number>")
    text = re_sub(r"#\S+", hashtag)
    text = re_sub(r"([!?.]){2,}", r"\1 <repeat>")
    text = re_sub(r"\b(\S*?)(.)\2{2,}\b", r"\1\2 <elong>")

    ## -- I just don't understand why the Ruby script adds <allcaps> to everything so I limited the selection.
    # text = re_sub(r"([^a-z0-9()<>'`\-]){2,}", allcaps)
    text = re_sub(r"([A-Z]){2,}", allcaps)

    return text.lower()


if __name__ == '__main__':
    _, text = sys.argv
    if text == "test":
        text = "I TEST alllll kinds of #hashtags and #HASHTAGS, @mentions and 3000 (http://t.co/dkfjkdf). w/ <3 :) haha!!!!!"
    tokens = tokenize(text)
    print tokens

Forked version with edits suggested by above comments: https://gist.github.com/ppope/0ff9fa359fb850ecf74d061f3072633a

What's the license on this? Same as the original?
Public Domain and Dedication license http://opendatacommons.org/licenses/pddl/

Found some bugs:

• It doesn't add the tag to hashtags in caps.
• It doesn't split hashtags as: #WelcomeRefugees = welcome refugees as the original does.

I paste here my version:

`"""
preprocess-twitter.py
python preprocess-twitter.py "Some random text with #hashtags, @mentions and http://t.co/kdjfkdjf (links). :)"
Script for preprocessing tweets by Romain Paulus
with small modifications by Jeffrey Pennington
with translation to Python by Motoki Wu
Translation of Ruby script to create features for GloVe vectors for Twitter data.
http://nlp.stanford.edu/projects/glove/preprocess-twitter.rb
"""
import sys
import regex as re

FLAGS = re.MULTILINE | re.DOTALL

def hashtag(text):
text = text.group()
hashtag_body = text[1:]
if hashtag_body.isupper():
result = " {} ".format(hashtag_body.lower())
else:
result = " ".join([""] + [re.sub(r"([A-Z])",r" \1", hashtag_body, flags=FLAGS)])
return result

def allcaps(text):
text = text.group()
return text.lower() + " "

def tweet_preprocessing(text):
# Different regex parts for smiley faces
eyes = r"[8:=;]"
nose = r"['`-]?"

# function so code less repetitive
def re_sub(pattern, repl):
    return re.sub(pattern, repl, text, flags=FLAGS)

text = re_sub(r"https?:\/\/\S+\b|www\.(\w+\.)+\S*", "<url>")
text = re_sub(r"@\w+", "<user>")
text = re_sub(r"{}{}[)dD]+|[)dD]+{}{}".format(eyes, nose, nose, eyes), "<smile>")
text = re_sub(r"{}{}p+".format(eyes, nose), "<lolface>")
text = re_sub(r"{}{}\(+|\)+{}{}".format(eyes, nose, nose, eyes), "<sadface>")
text = re_sub(r"{}{}[\/|l*]".format(eyes, nose), "<neutralface>")
text = re_sub(r"/"," / ")
text = re_sub(r"<3","<heart>")
text = re_sub(r"[-+]?[.\d]*[\d]+[:,.\d]*", "<number>")
text = re_sub(r"#\S+", hashtag)
text = re_sub(r"([!?.]){2,}", r"\1 <repeat>")
text = re_sub(r"\b(\S*?)(.)\2{2,}\b", r"\1\2 <elong>")

## -- I just don't understand why the Ruby script adds <allcaps> to everything so I limited the selection.
# text = re_sub(r"([^a-z0-9()<>'`\-]){2,}", allcaps)
text = re_sub(r"([A-Z]){2,}", allcaps)

return text.lower()

if name == 'main':
_, text = sys.argv
if text == "test":
text = "I TEST alllll kinds of #hashtags and #HASHTAGS and #HashTags, @mentions and 3000 (http://t.co/dkfjkdf). w/ <3 :) haha!!!!!"
tokens = tweet_preprocessing(text)
print tokens`