julien-f / js-arff

ARFF parsing and formatting

arff

ARFF parsing and formatting

Install

Download manually or with package-manager.

npm

> npm install --save arff

Usage

var arff = require('arff');

Parsing

var readFile = require('fs').readFile;

readFile('data.arff', 'utf8', function (error, content) {
  if (error) {
    return console.error(error);
  }

  console.log(arff.parse(content));
});

For the folling data.arff:

% My ARFF file.
@RELATION user

@ATTRIBUTE name       STRING
@ATTRIBUTE lastSignIn DATE
@ATTRIBUTE group      {none,read,write,admin}

@DATA
root,?,admin
julien-f,2014-12-16T19:42:01,write

It will give this output:

{
  relation: 'user',
  attributes: [
    {
      name: 'name',
      type: 'string'
    },
    {
      name: 'lastSignIn',
      type: 'date'
    },
    {
      name: 'name',
      type: 'enum',
      values: [
        'none',
        'read',
        'write',
        'admin'
      ]
    }
  ],
  data: [
    {
      name: 'root',
      group: 'admin'
    },
    {
      name: 'julien-f',
      lastSignIn: '2014-12-16T19:42:01',
      group: 'write'
    }
  ]
}

Formatting

console.log(arff.format(require('./relation')));

Which will display:

@RELATION foo

@ATTRIBUTE date date
@ATTRIBUTE dateWithFormat date "MM/DD/YY"
@ATTRIBUTE numeric numeric
@ATTRIBUTE string string
@ATTRIBUTE enumerate {"foo","bar","baz"}

@DATA
"2014-12-16T19:42:01+00:00","06/23/15",3.259,"can have spaces","bar"
"2014-12-16T19:42:01+00:00",?,42,?,?

Contributing

Contributions are very welcome, either on the documentation or on the code.

You may:

• report any issue you've encountered;
• fork and create a pull request.

License

ISC © Julien Fontanet

js-arff's Issues

Support date format

Trim whitespaces around raw strings

The specification suggests this should be the case.

Data should be objects instead of arrays

Properties should be of the correct type

• Dates should be JS dates.
• Numerics should be JS numbers.
• Strings & enumerates should be JS strings.
• Unknown values should be undefined.

Support tab as a field seperator

The specification does not mention the possibilities to use tab as a field separator but it seems that other softwares do.

It would make sense to support it as long as it does not cause issues.

Data should be a map instead of an array

And key should be entries' indices.