cleveragents-core/tests/features/rxpy_route_validation.feature

Feature: RxPY Route Validation for Run Command
  As a CleverAgents user
  I want RxPY stream routes to be restricted to interactive mode only
  So that I can avoid incomplete processing due to quiescence detection issues

  Background:
    Given I have a working directory for CLI tests


  Scenario: Detect RxPY routes in configuration
    Given I have a configuration with RxPY stream routes:
      """
      agents:
        test_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true

      routes:
        main:
          type: stream
          stream_type: cold
          operators:
            - type: map
              params:
                agent: test_agent
          publications:
            - __output__

      merges:
        - sources: [__input__]
          target: main
      """
    When I create a ReactiveCleverAgentsApp with this configuration
    Then the app should detect RxPY stream routes


  Scenario: Allow LangGraph routes in configuration
    Given I have a configuration with LangGraph routes:
      """
      agents:
        test_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true

      routes:
        main:
          type: graph
          entry_point: start
          nodes:
            start:
              type: agent
              agent: test_agent
          edges:
            - source: start
              target: end

      merges:
        - sources: [__input__]
          target: main
      """
    When I create a ReactiveCleverAgentsApp with this configuration
    Then the app should not detect RxPY stream routes


  Scenario: Run command rejects RxPY routes
    Given I have a configuration file with RxPY stream routes "rxpy_config.yaml":
      """
      agents:
        echo_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true

      routes:
        echo_stream:
          type: stream
          stream_type: cold
          operators:
            - type: map
              params:
                agent: echo_agent
          publications:
            - __output__

      merges:
        - sources: [__input__]
          target: echo_stream
      """
    When I run "cleveragents run -c rxpy_config.yaml -p 'test' --unsafe"
    Then the command should fail
    And the error should contain "RxPY stream routes which are only supported in 'interactive' mode"
    And the error should contain "Use 'interactive' command instead"
    And the error should contain "Migrate routes to LangGraph"


  Scenario: Run command with RxPY routes and context does not create context
    Given I have a configuration file with RxPY stream routes "rxpy_config.yaml"
    And the context "rxpy_test_context" does not exist
    When I run "cleveragents run -c rxpy_config.yaml -p 'test' --unsafe --context rxpy_test_context"
    Then the command should fail
    And the error should contain "RxPY stream routes which are only supported in 'interactive' mode"
    And the context "rxpy_test_context" should not exist


  Scenario: Interactive command accepts RxPY routes
    Given I have a configuration file with RxPY stream routes "rxpy_config.yaml"
    When I start interactive session "cleveragents interactive -c rxpy_config.yaml --unsafe"
    Then the interactive session should start successfully
    And no error about RxPY routes should be shown


  Scenario: Run command with LangGraph routes succeeds
    Given I have a configuration file with LangGraph routes "langgraph_config.yaml":
      """
      agents:
        echo_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true

      routes:
        main:
          type: graph
          entry_point: start
          nodes:
            start:
              type: agent
              agent: echo_agent
          edges:
            - source: start
              target: end

      merges:
        - sources: [__input__]
          target: main
      """
    When I run "cleveragents run -c langgraph_config.yaml -p 'echo test' --unsafe"
    Then the command should succeed
    And no error about route types should be shown


  Scenario: App method _has_rxpy_stream_routes with no routes
    Given I have a configuration with no routes:
      """
      agents:
        test_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true
      """
    When I create a ReactiveCleverAgentsApp with this configuration
    Then the app should not detect RxPY stream routes


  Scenario: App method _has_rxpy_stream_routes with mixed routes
    Given I have a configuration with both RxPY and LangGraph routes:
      """
      agents:
        test_agent:
          type: tool
          config:
            tools: ["echo"]
            safe_mode: true

      routes:
        stream_route:
          type: stream
          stream_type: cold
          operators:
            - type: map
              params:
                agent: test_agent
          publications:
            - graph_route

        graph_route:
          type: graph
          entry_point: start
          nodes:
            start:
              type: agent
              agent: test_agent
          edges:
            - source: start
              target: end

      merges:
        - sources: [__input__]
          target: stream_route
      """
    When I create a ReactiveCleverAgentsApp with this configuration
    Then the app should detect RxPY stream routes


  Scenario: Validation error provides helpful guidance
    Given I have a configuration file with RxPY stream routes "rxpy_config.yaml"
    When I run "cleveragents run -c rxpy_config.yaml -p 'test' --unsafe"
    Then the command should fail
    And the error message should explain the reason for restriction
    And the error message should provide solution alternatives
    And the error message should mention quiescence detection
    And the error message should mention nested operators