nagao 0.1.1

Counting word frequency based on Nagao algorithm

Nagao

An implementation of the paper: A new method of n-gram statistics for large number of n and automatic extraction of words and phrases from large text data of Japanese

Install

Two ways to install Nagao:

Install Nagao from Pypi:

pip install nagao

Install Nagao from the Github source:

git clone https://github.com/Chiang97912/nagao.git
cd nagao
python setup.py install

Usage

You can use Nagao in Python file:

from nagao import Nagao

nagao = Nagao(lang='en', min_ngram=2, max_ngram=6, min_freq=5, min_lrc=2, min_lre=0.5, min_pmi=0, min_eta=0, threshold=0,\
              use_disk=True, use_db=True, lower=True, clean=True, verbose=True)
ts = time.time()
nagao.process('path/to/corpus/file')
nagao.save('path/to/output/file')
print('total spend:', time.time() - ts)

From the command line, you can run:

nagao -c "path/to/corpus/file" -o "path/to/output/file" -l zh --clean --verbose

You can use nagao --help to find the usage of nagao cli:

Options:
  -c, --corpus TEXT           Corpus file path.
  -o, --output TEXT           Output file path.
  -l, --lang TEXT             Corpus language.
  -minn, --min_ngram INTEGER  Minimum n-gram size.
  -maxn, --max_ngram INTEGER  Max n-gram size.
  --min_freq INTEGER          Minimum frequency of word.
  --min_lrc INTEGER           Minimum count between left and right neighbor.
  --min_lre FLOAT             Minimum entropy between left and right neighbor.
  --min_pmi FLOAT             Minimum pmi(pointwise mutual information).
  --min_eta FLOAT             Minimum balanced value for left and right
                              neighbor count.

  --threshold FLOAT           Minimum word probability.
  --stopwords TEXT            Stopword file path.
  --punctuations TEXT         Punctuation file path.
  --lower                     If use lower option, keep lowered dictionary.
  --clean                     If use clean option, the cache file will be
                              cleaned at the end of the program.
  --verbose                   If use verbose option, logs will be displayed on
                              the terminal.

  --help                      Show this message and exit.

Dependencies

• Python version 3.6

• nltk version 3.5

References

• @newword-py