Function: when()

when<TInput, TOutput>(config): ConditionalStep<TInput, TOutput>

Defined in: packages/core/src/pipeline/helpers.ts:156

Create a conditional branching step for pipelines. Evaluates a condition and executes onPass or onFail branch accordingly.

The condition can be:

• String: Natural language condition (auto-creates BaleybotConditional with GPT-4)
• Function: Returns boolean or Promise
• Processable<TInput, boolean>: Returns boolean result

Type Parameters

TInput

TInput

Input type for condition and branches

TOutput

TOutput

Output type from both branches

Parameters

config

WhenConfig<TInput, TOutput>

Configuration object with condition, onPass, and optional onFail

Returns

ConditionalStep<TInput, TOutput>

ConditionalStep object for use in pipelines

Example

// With function condition
const myPipeline = pipeline(
  validatePost,
  when({
    condition: (post) => post.wordCount > 1000,
    onPass: publishImmediately,
    onFail: scheduleForReview
  })
);

// With string condition (uses AI)
const myPipeline = pipeline(
  analyzeContent,
  when({
    condition: 'this content is high quality and ready to publish',
    onPass: publishPost,
    onFail: flagForReview
  })
);

// With BaleybotConditional
const isHighQuality = BaleybotConditional.create({
  goal: 'Determine if content meets quality standards'
});

const myPipeline = pipeline(
  processContent,
  when({
    condition: isHighQuality,
    onPass: approve,
    onFail: reject
  })
);

// onFail is optional - input passes through if condition fails
const myPipeline = pipeline(
  loadData,
  when({
    condition: needsEnrichment,
    onPass: enrichData
    // No onFail - data passes through unchanged
  }),
  saveData
);