jina.jaml.parsers.flow.v1 module

class jina.jaml.parsers.flow.v1.V1Parser[source]

Bases: jina.jaml.parsers.base.VersionedYAMLParser

V1Parser introduces new syntax and features:

  • It has a top-level field version

  • deployments is now a List of Dict (rather than a Dict as prev.)

  • name is now optional

  • new field method can be used to specify how to add this Deployment into the Flow, availables are:

    • add: (default) equal to Flow.add(…)

    • needs: (default) equal to Flow.needs(…)

    • inspect: (default) equal to Flow.inspect(…)

  • An example V1 YAML config can be found below:

    1. !Flow
    2. version: '1.0'
    3. deployments:
    4.   - name: executor0  # notice the change here, name is now an attribute
    5.     method: add  # by default method is always add, available: add, needs, inspect
    6.     needs: gateway
    7.   - name: executor1  # notice the change here, name is now an attribute
    8.     method: add  # by default method is always add, available: add, needs, inspect
    9.     needs: gateway
    10.   - method: inspect  # add an inspect node on executor1
    11.   - method: needs  # let's try something new in Flow YAML v1: needs
    12.     needs: [executor1, executor0]

  • version = ‘1’

    the version number this parser designed for

  • parse(cls, data)[source]

    • Parameters

      • cls (type) – the class registered for dumping/loading

      • data (Dict) – flow yaml file loaded as python dict

      Return type

      Flow

      Returns

      the Flow YAML parser given the syntax version number

  • dump(data)[source]

    • Parameters

      data (Flow) – versioned flow object

      Return type

      Dict

      Returns

      the dictionary given a versioned flow object