nltk-api-server

API server for NLTK. Aimed to provide convenient interface to use NLTK over any programming language

Features

• Stemming
• Lemmatization
• Part of Speech Tagging (Unigram/Bigram/Regex)
• Named Entity Recognition
• Sentiment Analysis

1. Stemming

NLTK Stemmers used: Porter, Snowball and Lancaster

• Accepts:

  • /api/stem?words=<words>/

  • /api/stem?words=<words>&stemmer=<porter/snowball/lancaster/default>/

  • /api/stem?words=<words>&stemmer=snowball&language=<language>&ignore_stopwords=<true/false>/

• Query Parameters:

  • Mandatory:

    • words:

      value: string comma separated

  • Optional:

    • stemmer:

      value: porter/snowball/lancaster/default

      default: snowball

    • ignore_stopwords: Only for Snowball Stemmer

      value: true/false

      default: false

    • language: Only for Snowball Stemmer

      value: see SnowballStemmer.languages

      default: english

1.1 Examples

• localhost:9000/api/stem?words=dangerous,monitoring,testing

    {
        "status": true,
        "result": [
            "danger",
            "monitor",
            "test"
        ]
    }
  

• localhost:9000/api/stem/?words=dangerous,monitoring,testing&stemmer=snowball

    {
        "status": true,
        "result": [
            "dog",
            "cat",
            "vertex"
        ]
    }
  

• The above examples do not cover all cases. See the section above examples for more features

2. Lemmatization

NLTK Lemmatizer used: WordNetLemmatizer

• Accepts:

  • /api/lemma?words=<words>/

• Query Parameters:

  • Mandatory:

    • words:

      value: string comma separated

2.1 Examples

• localhost/api/lemma/?words=dogs,cats,vertices

    {
        "status":true,
        "result": [
            "dog",
            "cat",
            "vertex"
        ]
    }
  

3. Part of Speech Tagging

NLTK POS tagger used: pos_tag, UnigramTagger, BigramTagger & RegexpTagger

• Accepts:

  • /api/tag?sentence=<sentence>/

  • /api/tag?sentence=<sentence>&tagger=<pos/unigram/bigram/default>/

  • /api/tag?sentence=<sentence>&tagger=<pos/unigram/bigram/default>&train=<categories>/

  including any query parameter accepted by /api/tag/

• Query Parameters:

  • Mandatory:

    • sentence:

      value: string

  • Optional:

    • tagger:

      value: pos/unigram/bigram/regex

      default: pos_tag

    • train (iff unigram/bigram):

      value: 'news', 'editorial', 'reviews', 'religion', 'learned', 'science_fiction', 'romance', 'humor'

      default: 'news'

    • any query parameter acceptable by /api/tag/

3.1 Examples

• localhost/api/sentence=this is a test

    {
        "status": true,
        "result": [
            [
                "this",
                "DT"
            ],
            [
                "is",
                "VBZ"
            ],
            [
                "a",
                "DT"
            ],
            [
                "test",
                "NN"
            ]
        ]
    }
  

• localhost/api/sentence=this is a test&tagger=unigram

    {
        "status": true,
        "result": [
            [
                "this",
                "DT"
            ],
            [
                "is",
                "BEZ"
            ],
            [
                "a",
                "AT"
            ],
            [
                "test",
                "NN"
            ]
        ]
    }
  

• The above examples do not cover all cases. See the section above examples for more features

• Remember, we can also use trained data along with the unigram/bigram tagger: 'news', 'editorial', 'reviews', 'religion', 'learned', 'science_fiction', 'romance', 'humor'

4. Named Entity Recognition

NLTK NER used: ne_chunk

• Accepts:

  • /api/ner?sentence=<sentence>/

  including any query parameter accepted by /api/tag/

• Query Parameters:

  • Mandatory:

    • sentence:

      value: string

    • Optional:

      • any query parameter acceptable by /api/tag/

4.1 Examples

• localhost/api/ner?sentence=At the Olympics in August, Phelps picked up five gold medal

    {
        "status": true,
        "result": [
            "Phelps"
        ]
    }
  

5. Sentiment Analysis

NLTK Sentiment Analyzer used: vader

• Accepts:

  • /api/sentiment?sentence=<sentence>/

• Query Parameters:

  • Mandatory:

    • sentence:

      value: string

  • Optional:

    • any query parameter acceptable by /api/tag/

5.1 Examples

• localhost/api/sentiment?sentence=At the Olympics in August, Phelps picked up five gold medal

    {
        "status": true,
        "result": {
            "neg": 0,
            "neu": 0.256,
            "pos": 0.744,
            "compound": 0.4404
        }
    }
  

6. Run on local

• git clone [email protected]:vipul-sharma20/nltk-api-server.git
• cd nltk-api-server
• sudo pip install virtualenv
  • Python 2.7.9 and later (on the python2 series), and Python 3.4 and later include pip by default, so you may have pip already.
  • If you don't have pip installed, visit here to see steps to install virtualenv: https://virtualenv.readthedocs.org/en/latest/installation.html
• virtualenv nltk-api
• source nltk-api/bin/activate
• pip install -r requirements.txt (wait till the requirements are installed)
• python manage.py runserver This will run the application on http://127.0.0.1:8000/

IMPORTANT: You will require some corpora and trained models for the code to run. You can refer to: http://www.nltk.org/data.html

• Interactive Method:

    In [1]: import nltk
  
    In [2]: nltk.download()