Skip to content

Generate markov chains for chatbots in node

License

Notifications You must be signed in to change notification settings

postlight/node-markov

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

markov

Generate markov chains for chatbots and freestyle rap contests.

examples

qwantz

qwantz.js:

var util = require('util');
var fs = require('fs');

var markov = require('markov');
var m = markov(1);

var s = fs.createReadStream(__dirname + '/qwantz.txt');
m.seed(s, function () {
    var stdin = process.openStdin();
    util.print('> ');
    
    stdin.on('data', function (line) {
        var res = m.respond(line.toString()).join(' ');
        console.log(res);
        util.print('> ');
    });
});

output:

$ node example/qwantz.js 
> Hello friend.
Oh, that hurts me. How could fall apart, not unlike this tiny house. remains a danger when you? As I see him (quite often, Yes, As Thank I you? take have on! forgotten male, That oppression is is a A friend
> That is troubling news!
I've I had must to guard do against with such the a irony part of of their their fundamental fundamental injustices.
> Justice eh? SOMEBODY LIGHT UP THE BATSIGNAL
crazy I Utahraptor feel slipped alot in better! your about problems the put future! behind full You? of go My down perspective. The

methods

markov(order)

Create a new markov object of order order, which defaults to 2.

.seed(s, cb)

Seed the markov object with a string or stream s.

If s is a string, transition probabilities will be updated for every grouping of the previously specified order with dangling links at the front and end in the appropriate direction.

If ss is a stream, data events will be line-buffered and fed into .seed() again line-by-line.

If cb is specified it will fire once the seed text is fully ingested.

.search(text)

Search for and return some key found in the text body text.

Return undefined if no matches were found.

.pick()

Choose a key at random.

.next(key)

Find a key likely to follow after key.

Returns a hash with keys key, the canonical next key and word, a raw form of key as it appeared in the seed text.

.prev(key)

Find a key likely to come before key.

Returns a hash with keys key, the canonical next key and word, a raw form of key as it appeared in the seed text.

.forward(key, limit)

Generate a markov chain forward starting at key and returning an array of the raw word forms along the way.

Stop when the traversal hits a terminal entry or when limit words have been generated if limit is specified.

.backward(key, limit)

Generate a markov chain backward starting at key and returning an array of the raw word forms along the way.

Stop when the traversal hits a terminal entry or when limit words have been generated if limit is specified.

.fill(key, limit)

Generate a markov chain in both directions starting at key. Return an array of the raw word forms along the way including the raw word form of the supplied key.

Stop when the traversal hits a terminal entry or when limit words have been generated if limit is specified.

.respond(text, limit)

Search for a starting key in text and then call .fill(key, limit) on it.

About

Generate markov chains for chatbots in node

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • JavaScript 100.0%