[Epic] "data init" improvements

[Sébastien Lavoie]

Tasks

• rewrite init module in an OOP way extending EventEmitter class - datopian/data-cli#335
• implement default vs interactive modes - datopian/data-cli#336
• detect licenses - datopian/data-cli#239
• don't interrupt init due to validation errors - datopian/data-cli#241
• blog post

Analysis

data init command works only from cwd - you don't need to provide any arguments

Options:

• -i or --interactive - Run init in interactive mode
• -h or --help - Output usage information

Name/title of a dataset

By default it uses the name of the cwd as dataset's name and uses the unslugified version of the name as a title, eg, cwd's name is my-dataset:

dp.json:

{
  "name": "my-dataset",
  "title": "My dataset"
}

In the interactive mode, it will ask for confirmation.

Adding resources and scanning dirs

Without interactive mode, the command will scan all files and nested directories in the cwd and generate datapackage.json from every file it has found. In the interactive mode it would ask before adding a file into resources and before scanning any nested directory.

dp.json already exists

This will work only in interactive mode as we don't want to break anything, e.g., overwrite existing properties.

other work to be done

We want to rewrite old functions in an OOP way so we can emit events and print out statuses for better UX + for reusability from other libraries.

Created by: @anuveyatsu