Can I create a corpus from a collection of strings in NLTK?

Question

Is there a way to create a corpus without having to have items in files. For instance, I want to manipulate Tweets or paragraphs that I am grabbing from the web. Can I do something like

myCorpus = MyCorpus([
    ('id', 'item', 'category'), 
    ('id', 'item', 'category'),
    ('id', 'item', 'category'), 
    ... ])

Or

myCorpus.add('id', 'item', 'category')

The purpose is to manipulate the corpus with existing NLTK capabilities. I checked TextCollection but it seems that it doesn't handle categories.

score -1 · Answer 1 · answered Dec 29 '10 at 18:33

-1

Why not just write the strings out to a file or files and then process them as a corpus?

answered Dec 29 '10 at 18:33

Jamie Forrest

10,895
6
51
68

It may be faster to keep things in memory. For example, if you're scraping content from the web. – Tim McNamara Jan 10 '11 at 07:22

Can I create a corpus from a collection of strings in NLTK?

1 Answers1