regex4dummies 1.4.2

A NLP library that simplifies pattern finding in strings

Simple pattern finding in strings and natural language processing. Checkout regex4dummies’ website at https://darkmattervale.github.io/regex4dummies/

Features

• Automatic pattern detection ( semantic and literal )

• Multiple parsers ( implementations of nltk, pattern, and nlpnet )

• Keyword searching to find specific phrases

• Topic analysis

• Tokenizer and sentence dependency identifier

Roadmap

Some features I plan to implement in the future:

• Machine Learning. This will allow the parsers to learn multiple grammatical “styles” and be able to successfully parse a much wider selection of strings

• Additional parsers

• Continued improvements on the current parsers. This will be a long-term project, and additional details can be found on the main Github page

If you have feature requests, feel free to add an issue to the Github issue tracker. All contributions and requests are appreciated!

Usage

regex4dummies is very easy to use. Simply import the library, get some strings, and compare them!

from regex4dummies import regex4dummies
from regex4dummies import Toolkit

# Creating strings
strings = [ "This is the first test string.", "This is the second test string." ]

regex = regex4dummies()

# Identifying literal patterns in strings
print regex.compare_strings( 'default', True, strings )

# Identifying semantic patterns in strings using the nltk parser
print regex.compare_strings( 'nltk', False, strings )

Above is regex4dummies in its simplest form. It allows for additional features as well, including:

# Display the version of regex4dummies you are using
print regex.__version__

# To use the other parsers, replace the above line of code with either of the following:
# print regex.compare_strings( 'pattern', False, strings )
# print regex.compare_strings( 'nlpnet', False, strings )

# To call all of the parsers, replace the above line of code with the following:
# print regex.compare_strings( '', False, strings )

# To get the topics of the strings, call the get_pattern_topics function
print regex.get_pattern_topics()

# Printing pattern information
pattern_information = regex.get_sentence_information()
  for objects in pattern_information:
      print "[ Pattern ]             : " + objects.pattern
      print "[ Subject ]             : " + objects.subject
      print "[ Verb ]                : " + objects.verb
      print "[ Object ]              : " + objects.object[0]
      print "[ Prep Phrases ]        : " + str( objects.prepositional_phrases )
      print "[ Reliability Score ]   : " + str( objects.reliability_score )
      print "[ Applicability Score ] : " + str( objects.applicability_score )
      print ""

A newly released set of features include a tokenizer and dependency finder function. To use them, simply give a string as the first parameter, and the name of the parser you would like to use as the second for both functions.

# Testing the toolkit functions
tool_tester = Toolkit()

# Testing the tokenizer functions
print tool_tester.tokenize( "This is a test string.", "pattern" )

# Testing the dependency functions
print tool_tester.find_dependencies( "This is a test string.", "pattern" )

Installation

To install this library, use pip.

$ pip install regex4dummies

In addition to the library, wget is a required command-line command to use the nlpnet parser. If you do not have wget or cannot get it, follow the below directions to still get the functionality of the nlpnet parser.

Instructions to install the required dependency for nlpnet:

1. Download the nlpnet_dependency file on the most recent release found in Github ( please not, when uncompressed, this file is over 350 MB large ).

2. Place this directory into the same directory that nltk-data is located ( if you don’t have that installed, just run the library and go through the GUI downloader )

That’s it! The nlpnet parser should now be able to be used.

Patch Notes

( Latest ) v1.4.2: Miscellaneous Update & overall performance improvement

• Accuracy tester is a lot better ( except for semantic comparisons ). A new “corpus” will be added soon to get accurate semantic comparisons

• Literal parser improvements ( which are numerous ). Multiple bug fixes, sub-patterns are not gathered ( patterns within a larger pattern no longer count as individual patterns ), and some general improvements

• Behind-the-scenes work to update code and make it more flexible for future updates

• GUI update. It is now backwards compatible

• Toolkit update! It should now work, and I apologize for it not working in previous updates

• Topic analyzer update. It should find more topics, as well as returning the topics in order from most-likely to be the main topic to least likely. In the future, an update will allow the extraction of the most important information as a whole ( not just the topics )

• Source code is smaller. Removed some code that was not being used, helping reduce the size of the repository

Contributing

Contributors are welcome and much needed! regex4dummies is still under heavy development, and needs all of the help it can get. If you have any feature ideas, feel free to create an issue on the github repository ( https://github.com/darkmattervale/regex4dummies/issues ) or fork the repository and create your addition.

Any help you can give is much appreciated. The more help we get, the better regex4dummies will perform. Thanks for contributing!

License

Please see LICENSE.txt for information about the MIT license

Citations

nlpnet:

• Fonseca, E. R. and Rosa, J.L.G. Mac-Morpho Revisited: Towards Robust Part-of-Speech Tagging. Proceedings of the 9th Brazilian Symposium in Information and Human Language Technology, 2013. p. 98-107 [PDF]