Skip to main content

Automatic Keyword Extraction from Document Collections

Project description


Distiller provides convenient auto-extraction of document key words
based on term-frequency/inverse-document-frequency (TF-IDF) and word

Distiller handles all of the pre-processing details and produces final
statistic reports in JSON format.


Distiller uses the [Natural Language Toolkit](

You will need to download a couple of NLTK packages:

>>> import nltk
Downloader> d
Download which package (l=list; x=cancel)?
Identifier> maxent_treebank_pos_tagger
Downloader> d
Download which package (l=list; x=cancel)?
Identifier> stopwords


Installation using pip:

$ pip install Distiller


Typical usage from within the Python interpreter:

>>> from Distiller.distiller import Distiller
>>> distiller = Distiller(data, target, options)


### data

Path to file containing the document collection in JSON format.

'metadata': {
'base_url': 'The document's source URL (if any)'
'documents': [
'id': 'The document's unique identifier (if any)',
'body': 'The entire body of the document in a single text blob.',
}, ...


Path where Distiller will output the following reports:

keywords: A list of words and the frequency with which they were detected as being keywords of documents.

bigrams: A list of word pairs and the frequency with which they were detected as being key pairs in documents.

trigrams: A list of word triples and the frequency with which they were detected as being key pairs in documents.

docmap: A mapping of document IDs to their respective keywords, n-grams, and other statistics.

keymap: A mapping of keywords to the documents they appear in.


An optional dictionary containing document processing arguments in this format:

'normalize': True, # normalize tokens during pre processing
'stem': True, # stems tokens during pre processing
'lemmatize': False, # lemmatize during pre processing
'tfidf_cutoff': 0.001, # cutoff value to use for term-freq/doc-freq score
'pos_list': ['NN','NNP'], # POS white list used to filter for candidates
'black_list': [] # token list used to filter out from candidates

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for Distiller, version 0.1.2
Filename, size File type Python version Upload date Hashes
Filename, size Distiller-0.1.2.tar.gz (9.1 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page