Refactor new Flow trigger conditions to use a new Flow trigger property preconditions with flows, where and timeWindow subproperties

[anna-geller]

Feature description

Context:

1. We got feedback that some people considered the timeSLA property confusing → timeWindow
2. We identified a path forward to add a dedicated flow-level sla property: Add flow-level SLA property #5857
3. The addition of Condition to all conditions is not strictly necessary (validated with Ludo)

We'll refactor as show in this example:

id: myflow42
namespace: company.team
description: |
  Next steps:
    1. Remove the LABEL magic syntax with label:value for now
    2. Add labels to the preconditions.flows array map
    3. Refactor to include preconditions with flows and where
    
    EQUAL_TO,
    NOT_EQUAL_TO,
    IN,
    NOT_IN,
    IS_TRUE,
    IS_FALSE,
    IS_NULL,
    IS_NOT_NULL,
    STARTS_WITH,
    ENDS_WITH,
    REGEX,
    CONTAINS,
    EXPRESSION is a field, not a zype, the type is then IS_TRUE

triggers:
  - id: wait_for_upstream
    type: io.kestra.plugin.core.trigger.Flow
    conditions:
      - type: io.kestra.plugin.core.condition.HasRetryAttemptCondition
    preconditions:
      timeWindow:
        type: DURATION_WINDOW
        window: PT1D
      flows: # AND by default for all list elements
        - namespace: company.team
          flowId: myflow1
          states: [SUCCESS, WARNING]
        - namespace: company.team
          flowId: myflow2
          states: [SUCCESS]
          labels:
            mykey: myvalue
            another: labelvalue
          inputs: # TODO
            myuser: Loic
# executionFilters, query filters
      where:
        - id: flow1_states # operand: AND # default, needs to be specified only if OR
          filters:
            - field: NAMESPACE
              type: STARTS_WITH
              value: company
            - field: FLOW_ID
              type: IN
              values: 
                - flow1
                - flow2
            - field: STATE
              type: IN
              values: [SUCCESS, WARNING, CANCELLED]

        - id: flow2
          operand: AND
          filters:          
            - field: NAMESPACE
              type: EQUAL_TO
              value: company.team
            - field: FLOW_ID
              type: EQUAL_TO
              value: flow2
            - field: STATE
              type: EQUAL_TO
              value: SUCCESS
            - field: EXPRESSION
              type: IS_TRUE
              value: "{{outputs.output.values.variable == 'value'}}"
            # - field: LABEL
            #   type: EQUAL_TO
            #   values:
            #     mylabelkey: value
            #     another: value
            - field: EXPRESSION
              type: IS_TRUE
              value: "{{ labels.mylabelkey == 'mykey' and labels.mylabelvalue == 'value' }}"
tasks:
  - id: hello
    type: io.kestra.plugin.core.log.Log
    message: I'm triggered after two flows!

Optionally, we can rename all other conditions to not use the Condition word at the end + add an alias with the old name.

Big thanks @loicmathieu for all the iterations and work put into this!

[loicmathieu]

@anna-geller

Optionally, we can rename all other conditions to not use the Condition word at the end + add an alias with the old name.

You call ;)
Renaming is easy

[anna-geller]

Alright, I'd say let's do it!

Any time I see:

• io.kestra.plugin.core.condition.DateTimeBetweenCondition
• io.kestra.plugin.core.condition.DayWeekInMonthCondition
• io.kestra.plugin.core.condition.HasRetryAttemptCondition

leads me to believe that this below is way easier to read/grasp:

• io.kestra.plugin.core.condition.DateTimeBetween
• io.kestra.plugin.core.condition.DayWeekInMonth
• io.kestra.plugin.core.condition.HasRetryAttempt