Markov Bot

A Markov Chain bot that wants to get to know you.

Quickstart™

Prerequisites:

Node.js - version 4+
npm - version 3+
Redis - tested against version 2.8, but probably works with older versions
Elasticsearch - tested against version 2.2 and 2.3

Clone this repo:

git clone git@github.com:mmacaulay/markov-bot.git

Install npm packages

npm install

Run the tests

npm test

API

Run in dev mode (runs the babel preprocessor and monitors filesystem changes with nodemon

npm run dev

Seed data

There is a lovely collection of seed data available to play with in the seeds directory.

bin/seed --orders 1,2 --namespaces alice seeds/alice.txt

For a full list of options, run the bin/seed utility without arguments.

Learn

POST /learn

Learn from a corpus of text.

Options:

text (String) [required] - Text to learn from.
namespaces (Array of String) [optional] - List of namespaces to learn, create an additional index on each namespace. The index 'all' is always learned.
orders (Array of Integer) [optional] - List of word orders to learn. The order 1 is always learned.

Request:

POST /learn HTTP/1.1
Content-Type: application/json
...
{
  "text": "The quick brown fox jumps over the lazy dog.",
  "namespaces": ["mac"]
  "orders": [2, 3]
}

Response:

HTTP/1.1 200 OK
Content-Type: application/json; charset=utf-8
...

ok

Speak

GET /speak

Generate some occasionally witty banter. For best results, make at least one prior call to /learn.

Options:

namespace (String) [optional] - Generate text based on past things learned as namespace.
about (String) [optional] - Generate text containing the about word.
order (String) [optional] - Generate text with the specified word order. Default 1.

Request:

GET /speak?namespace=alice&about=trees&order=2 HTTP/1.1

Response:

HTTP/1.1 200 OK
Content-Type: application/json; charset=utf-8
...

{"data":"There was nothing but the trees behind him."}