aepiotti / preprocessor

Elegant and Easy Tweet Preprocessing in Python

Preprocessor

Preprocessor is a preprocessing library for tweet data written in Python. When building Machine Learning systems based on tweet and text data, a preprocessing is required. This is required because of quality of the data as well as dimensionality reduction purposes.

This library makes it easy to clean, parse or tokenize the tweets so you don't have to write the same helper functions over and over again ever time.

Features

Currently supports cleaning, tokenizing and parsing:

• URLs
• Hashtags
• Mentions
• Reserved words (RT, FAV)
• Emojis
• Smileys
• Numbers
• JSON and .txt file support

Preprocessor v0.6.0 supports Python 3.4+ on Linux, macOS and Windows. Tests run on following setups:

Linux Xenial with Python 3.4.8, 3.5.6, 3.6.7, 3.7.1, 3.8.0, 3.8.3+
macOS with Python 3.7.5, 3.8.0
Windows with Python 3.5.4, 3.6.8

Usage

Basic cleaning:

>>> import preprocessor as p
>>> p.clean('Preprocessor is #awesome 👍 https://github.com/s/preprocessor')
'Preprocessor is'

Tokenizing:

>>> p.tokenize('Preprocessor is #awesome 👍 https://github.com/s/preprocessor')
'Preprocessor is $HASHTAG$ $EMOJI$ $URL$'

Parsing:

>>> parsed_tweet = p.parse('Preprocessor is #awesome https://github.com/s/preprocessor')
<preprocessor.parse.ParseResult instance at 0x10f430758>
>>> parsed_tweet.urls
[(25:58) => https://github.com/s/preprocessor]
>>> parsed_tweet.urls[0].start_index
25
>>> parsed_tweet.urls[0].match
'https://github.com/s/preprocessor'
>>> parsed_tweet.urls[0].end_index
58

Fully customizable:

>>> p.set_options(p.OPT.URL, p.OPT.EMOJI)
>>> p.clean('Preprocessor is #awesome 👍 https://github.com/s/preprocessor')
'Preprocessor is #awesome'

Preprocessor will go through all of the options by default unless you specify some options.

Processing files:

Preprocessor currently supports processing .json and .txt formats. Please see below examples for the correct input format.

Example JSON file

[
    "Preprocessor now supports files. https://github.com/s/preprocessor",
    "#preprocessing is a cruical part of @ML projects.",
    "@RT @Twitter raw text data usually has lots of #residue. http://t.co/g00gl"
]

Example Text file

Preprocessor now supports files. https://github.com/s/preprocessor
#preprocessing is a cruical part of @ML projects.
@RT @Twitter raw text data usually has lots of #residue. http://t.co/g00gl

Preprocessing JSON file:

# JSON example
>>> input_file_name = "sample_json.json"
>>> p.clean_file(input_file_name, options=[p.OPT.URL, p.OPT.MENTION])
Saved the cleaned tweets to:/tests/artifacts/24052020_013451892752_vkeCMTwBEMmX_clean_file_sample.json

Preprocessing text file:

# Text file example
>>> input_file_name = "sample_txt.txt"
>>> p.clean_file(input_file_name, options=[p.OPT.URL, p.OPT.MENTION])
Saved the cleaned tweets to:/tests/artifacts/24052020_013451908865_TE9DWX1BjFws_clean_file_sample.txt

Available Options:

Option Name	Option Short Code
URL	p.OPT.URL
Mention	p.OPT.MENTION
Hashtag	p.OPT.HASHTAG
Reserved Words	p.OPT.RESERVED
Emoji	p.OPT.EMOJI
Smiley	p.OPT.SMILEY
Number	p.OPT.NUMBER

Installation

Using pip:

$ pip install tweet-preprocessor

Using Anaconda:

$ conda install -c saidozcan tweet-preprocessor

Using manual installation:

$ python setup.py build
$ python setup.py install

Contributing

Are you willing to contribute to preprocessor? That's great! Please follow below steps to contribute to this project:

1. Create a bug report or a feature idea using the templates on Issues page.
2. Fork the repository and make your changes.
3. Open a PR and make sure your PR has tests and all the checks pass.
4. And that's all!