Canonical tests for parsers

Datastructure

Every tests looks like this:

testBasicDirection: ‹ 1
    source: |               ‹ 2
      Add a bit of chilli
    result:                      ‹ 3
      steps:
        -
          - type: text                   ‹ 4
            value: "Add a bit of chilli"
      metadata:                             ‹ 5
        "cooking time": 30 mins

Key components:

1. Test name. It gives an idea what feature this test is testing and can be useful for debugging.

2. Source. Recipe text file as it'd provided by a user.

3. Result. Expected result of parsing. It consists of multiple steps and metadata.

4. Direction item. Every step consists of multiple direction items. There're four types:

   • text has value which contains text used in step.

   • ingredient has name, quantity and units, and optional: true when marked with ? (optional defaults to false and is omitted when false).

   • cookware has name and quantity, and optional like ingredients.

   • timer has name, quantity and units.

5. Metadata is in key/value form.

Example usage in https://github.com/cooklang/CookInSwift.