@falai/agent 1.2.8 → 2.0.0

package/docs/core/conversation-flows/step-transitions.md

@@ -1,731 +0,0 @@
-# Step Transitions & Conditional Logic
-
-The step transition system provides intelligent conversation flow control, enabling dynamic, context-aware progression through conversation routes. Unlike linear chatbots, this system can skip irrelevant steps, enforce dependencies, and adapt to user context.
-
-## Overview
-
-Step transitions handle:
-
-- **Conditional Skipping**: Skip steps based on collected data or context
-- **Dependency Enforcement**: Ensure prerequisite data exists before proceeding
-- **Branch Resolution**: Handle multiple possible next steps
-- **Route Completion**: Detect when routes reach their end
-- **Loop Prevention**: Avoid infinite traversal in complex flows
-
-## Stopping Conditions in Batch Execution
-
-When executing multiple steps in a batch, the engine stops for specific reasons indicated by the `stoppedReason` field in the response.
-
-### StoppedReason Values
-
-| Reason | Description | Behavior |
-|--------|-------------|----------|
-| `needs_input` | Step requires data not yet available | Batch stops, LLM generates response to collect data |
-| `end_route` | Reached END_ROUTE marker | Route is complete, no more steps to execute |
-| `route_complete` | All steps in route processed | Route finished successfully |
-| `max_steps_reached` | Batch hit the `maxStepsPerBatch` limit | Batch stops, remaining steps execute in subsequent calls |
-| `prepare_error` | Error in prepare hook | Batch stops, error returned with last successful state |
-| `llm_error` | Error during LLM call | Batch stops, session state preserved |
-| `validation_error` | Data validation failed | Batch continues, errors included in response |
-| `finalize_error` | Error in finalize hook | Non-fatal, logged and execution continues |
-
-### Needs-Input Stopping
-
-The most common stopping condition. A step needs input when:
-
-```typescript
-// Step needs input if requires fields are missing
-const step1 = {
-  prompt: "Confirm your booking",
-  requires: ["hotel", "date"], // Both must be present
-};
-
-// Step needs input if collecting and no collect fields have data
-const step2 = {
-  prompt: "What's your preference?",
-  collect: ["preference", "notes"], // Needs input if BOTH are missing
-};
-```
-
-### End-Route Stopping
-
-Batch stops when reaching the END_ROUTE marker:
-
-```typescript
-const finalStep = confirmStep.nextStep({
-  prompt: "Booking confirmed! Anything else?",
-}).endRoute(); // Creates END_ROUTE transition
-
-// Response will have:
-// stoppedReason: "end_route"
-// isRouteComplete: true
-```
-
-### Route-Complete Stopping
-
-When all steps have been processed without hitting END_ROUTE:
-
-```typescript
-// If the last step has no transitions and doesn't need input
-const response = await agent.respond("Complete my booking");
-
-// Response will have:
-// stoppedReason: "route_complete"
-// isRouteComplete: true
-```
-
-### Transitions Within Batched Execution
-
-During batch execution, transitions work as follows:
-
-1. **Linear transitions** - Steps connected via `nextStep()` are evaluated sequentially
-2. **SkipIf evaluation** - Each step's `skipIf` is checked before inclusion
-3. **Needs-input check** - If a step needs input, batch stops there
-4. **END_ROUTE detection** - Batch stops when reaching route end
-
-```typescript
-// Example: Steps A → B → C → END_ROUTE
-// If user provides data for A and B but not C:
-
-const response = await agent.respond("Data for A and B");
-
-// Batch includes: [A, B]
-// Stops at: C (needs_input)
-// Next call will continue from C
-```
-
-### Error Stopping Conditions
-
-Errors during batch execution have different behaviors:
-
-```typescript
-// Prepare hook error - stops immediately
-const stepWithPrepare = {
-  prompt: "Processing...",
-  prepare: async (context, data) => {
-    if (!data.valid) throw new Error("Invalid data");
-  },
-};
-// stoppedReason: "prepare_error"
-// Session state: last successful state preserved
-
-// LLM error - stops with preserved state
-// stoppedReason: "llm_error"
-// Session state: preserved from before LLM call
-
-// Validation error - continues but reports error
-// stoppedReason: "validation_error"
-// Collected data: partial data preserved
-
-// Finalize hook error - logged, continues
-// stoppedReason: original reason (not changed)
-// Error: included in response for logging
-```
-
-### Checking Stop Reason in Response
-
-```typescript
-const response = await agent.respond("Book a hotel");
-
-switch (response.stoppedReason) {
-  case 'needs_input':
-    console.log("Waiting for user input");
-    break;
-  case 'max_steps_reached':
-    console.log("Batch limit reached, more steps pending");
-    break;
-  case 'end_route':
-  case 'route_complete':
-    console.log("Route finished:", response.isRouteComplete);
-    break;
-  case 'prepare_error':
-  case 'llm_error':
-    console.error("Error occurred:", response.error);
-    break;
-  case 'validation_error':
-    console.warn("Validation issues:", response.error);
-    break;
-}
-```
-
-## Conditional Skipping
-
-### SkipIf Logic
-
-The `skipIf` function determines whether a step should be bypassed:
-
-```typescript
-const smartRoute = agent
-  .createRoute({
-    title: "Smart Onboarding",
-
-    initialStep: {
-      prompt: "What's your name?",
-      collect: ["name"],
-      skipIf: (data) => data.name !== undefined, // Skip if name already collected by any route
-    },
-  })
-  .nextStep({
-    prompt: "What's your email, {{name}}?",
-    collect: ["email"],
-    requires: ["name"], // Must have name from agent data to proceed
-    skipIf: (data) => data.email !== undefined, // Skip if email already collected by any route
-  });
-```
-
-### Common SkipIf Patterns
-
-```typescript
-// Skip based on user preferences
-skipIf: (data) => data.skipNewsletterSignup === true;
-
-// Skip if data already exists
-skipIf: (data) => data.userProfile?.completed === true;
-
-// Skip based on external conditions
-skipIf: (data) => {
-  const now = new Date();
-  const lastContact = data.lastContactDate;
-  return lastContact && now - lastContact < 24 * 60 * 60 * 1000; // Within 24 hours
-};
-
-// Skip based on user type
-skipIf: (data) => data.userType === "premium" && data.onboardingComplete;
-```
-
-## Dependency Enforcement
-
-### Requires Validation
-
-The `requires` array ensures prerequisite data exists:
-
-```typescript
-const dependentRoute = agent
-  .createRoute({
-    title: "Dependent Collection",
-
-    initialStep: {
-      prompt: "What's your name?",
-      collect: ["name"],
-    },
-  })
-  .nextStep({
-    prompt: "What's your email address?",
-    collect: ["email"],
-    requires: ["name"], // Cannot proceed without name
-  })
-  .nextStep({
-    prompt: "Should I send updates to {{email}}?",
-    collect: ["emailConsent"],
-    requires: ["name", "email"], // Needs both name and email
-  });
-```
-
-### Complex Requirements
-
-```typescript
-const complexRequirements = {
-  prompt: "Let's schedule your appointment.",
-  collect: ["appointmentDate", "appointmentTime"],
-  requires: [
-    "name", // Basic info
-    "email", // Contact info
-    "serviceType", // What service they want
-    "availability", // When they're available
-  ],
-  skipIf: (data) => {
-    // Skip if appointment already scheduled
-    return data.appointmentId !== undefined;
-  },
-};
-```
-
-## Step Traversal Logic
-
-### Candidate Step Algorithm
-
-The system uses a sophisticated algorithm to find valid next steps:
-
-```typescript
-// 1. Start with current step's direct transitions
-const transitions = currentStep.getTransitions();
-
-// 2. For each transition, evaluate:
-for (const transition of transitions) {
-  // Check for END_ROUTE marker
-  if (transition.id === END_ROUTE_ID) {
-    return { isRouteComplete: true };
-  }
-
-  // Skip steps that should be skipped
-  if (transition.shouldSkip(data)) {
-    // Recursively traverse to find next valid step
-    continue;
-  }
-
-  // Check if step has required data
-  if (!transition.hasRequires(data)) {
-    // Step cannot be entered yet
-    continue;
-  }
-
-  // Found valid step
-  return { step: transition };
-}
-```
-
-### Recursive Traversal
-
-For complex flows with skipIf conditions:
-
-```typescript
-// Example flow: A -> B -> C -> END
-// If user has already completed B, skip to C
-// If B and C are complete, route is done
-
-const flow = agent
-  .createRoute({
-    title: "Complex Flow",
-
-    initialStep: {
-      /* Step A */
-    },
-  })
-  .nextStep({
-    prompt: "Step B content",
-    skipIf: (data) => data.stepBComplete,
-  })
-  .nextStep({
-    prompt: "Step C content",
-    skipIf: (data) => data.stepCComplete,
-    requires: ["stepBComplete"],
-  })
-  .endRoute();
-```
-
-## Branching Logic
-
-### Branch Evaluation
-
-When multiple branches exist, the system evaluates each:
-
-```typescript
-const branchingRoute = agent
-  .createRoute({
-    title: "Multi-Path Flow",
-
-    initialStep: {
-      prompt: "What type of help do you need?",
-      collect: ["helpType"],
-    },
-  })
-  .branch([
-    {
-      name: "technical",
-      step: {
-        prompt: "Technical support question?",
-        collect: ["techIssue"],
-        requires: ["helpType"],
-        when: "if helpType indicates technical issue",
-      },
-    },
-    {
-      name: "billing",
-      step: {
-        prompt: "Billing question?",
-        collect: ["billingIssue"],
-        requires: ["helpType"],
-        when: "if helpType indicates billing issue",
-      },
-    },
-    {
-      name: "general",
-      step: {
-        prompt: "General question?",
-        collect: ["generalQuestion"],
-        requires: ["helpType"],
-        // No 'when' condition - fallback branch
-      },
-    },
-  ]);
-```
-
-### Branch Resolution
-
-```typescript
-// The routing engine evaluates branches and selects the best path
-// 1. Check explicit 'when' conditions (AI-evaluated)
-// 2. Fall back to branches without conditions
-// 3. Allow multiple branches if no clear winner
-
-// Access specific branches for chaining
-const techBranch = branchingRoute.branches.technical.nextStep({
-  prompt: "Let me help with that technical issue...",
-  requires: ["techIssue"],
-});
-```
-
-## Route Completion
-
-### End Route Detection
-
-Routes complete when reaching END_ROUTE:
-
-```typescript
-const completeRoute = agent
-  .createRoute({
-    title: "Task Completion",
-
-    initialStep: { prompt: "What task?", collect: ["task"] },
-  })
-  .nextStep({
-    prompt: "I'll help with {{task}}",
-    requires: ["task"],
-  })
-  .nextStep({
-    prompt: "Task details?",
-    collect: ["details"],
-    requires: ["task"],
-  })
-  .endRoute({
-    prompt: "Task completed! Anything else?",
-    collect: ["satisfaction"], // Optional final collection
-  });
-```
-
-### Completion Handlers
-
-```typescript
-const routeWithCompletion = agent.createRoute({
-  title: "Workflow",
-
-  // Simple string transition
-  onComplete: "feedback-route",
-
-  // Or conditional transition
-  // onComplete: {
-  //   nextStep: "success-route",
-  //   condition: "if workflow succeeded"
-  // }
-
-  // Or dynamic transition
-  // onComplete: async (session, context) => {
-  //   if (session.data?.success) return "success-route";
-  //   return "retry-route";
-  // }
-});
-```
-
-## Advanced Transition Patterns
-
-### Circular Flows
-
-Handle repeating patterns with proper termination:
-
-```typescript
-const iterativeRoute = agent
-  .createRoute({
-    title: "Iterative Collection",
-
-    initialStep: {
-      prompt: "Add an item (or 'done' to finish)",
-      collect: ["currentItem"],
-    },
-  })
-  .nextStep({
-    prompt: "Item added: {{currentItem}}. Add another?",
-    collect: ["continue"],
-    requires: ["currentItem"],
-    skipIf: (data) => data.continue === false || data.currentItem === "done",
-  })
-  .nextStep({
-    prompt: "Add another item?",
-    collect: ["currentItem"],
-    requires: ["continue"],
-    // Loops back to previous step
-  });
-```
-
-### State-Based Transitions
-
-```typescript
-const statefulRoute = agent
-  .createRoute({
-    title: "State Machine",
-
-    initialStep: {
-      prompt: "What's your current status?",
-      collect: ["status"],
-    },
-  })
-  .branch([
-    {
-      name: "new",
-      step: {
-        prompt: "Welcome! Let's get you started.",
-        collect: ["onboardingData"],
-        when: "if status is 'new'",
-        skipIf: (data) => data.status !== "new",
-      },
-    },
-    {
-      name: "returning",
-      step: {
-        prompt: "Welcome back! What can I help with?",
-        collect: ["request"],
-        when: "if status is 'returning'",
-        skipIf: (data) => data.status !== "returning",
-      },
-    },
-  ]);
-```
-
-## Lifecycle Integration
-
-### Prepare and Finalize Hooks
-
-```typescript
-const hookedRoute = agent
-  .createRoute({
-    title: "Lifecycle Demo",
-
-    initialStep: {
-      prompt: "Starting process...",
-      prepare: async (context, data) => {
-        // Setup before AI responds
-        console.log("Preparing step with context:", context);
-        console.log("Current data:", data);
-      },
-      finalize: async (context, data) => {
-        // Cleanup after AI responds
-        console.log("Step completed with data:", data);
-        await saveProgress(data);
-      },
-    },
-  })
-  .nextStep({
-    prompt: "Process step...",
-    prepare: (context, data) => {
-      // Validate state before proceeding
-      if (!data.requiredField) {
-        throw new Error("Missing required field");
-      }
-    },
-  });
-```
-
-### Route-Level Hooks
-
-```typescript
-const routeWithHooks = agent.createRoute({
-  title: "Route Lifecycle",
-
-  hooks: {
-    onDataUpdate: (newData, previousData) => {
-      // Validate data changes across the route
-      if (newData.criticalField && !isValid(newData.criticalField)) {
-        throw new Error("Invalid critical field");
-      }
-      return newData;
-    },
-
-    onContextUpdate: (newContext, previousContext) => {
-      // React to context changes
-      console.log("Route context updated");
-    },
-  },
-});
-```
-
-## Error Handling & Recovery
-
-### Transition Validation
-
-```typescript
-const robustRoute = agent
-  .createRoute({
-    title: "Error Handling",
-
-    initialStep: {
-      prompt: "What would you like to do?",
-      collect: ["action"],
-      finalize: async (context, data) => {
-        try {
-          await validateAction(data.action);
-        } catch (error) {
-          // Set error state for recovery
-          data.actionError = error.message;
-          data.needsRecovery = true;
-        }
-      },
-    },
-  })
-  .nextStep({
-    prompt:
-      "There was an issue with your request: {{actionError}}. Please try again.",
-    collect: ["action"],
-    skipIf: (data) => !data.needsRecovery,
-    finalize: (context, data) => {
-      // Clear error state on successful retry
-      delete data.actionError;
-      delete data.needsRecovery;
-    },
-  });
-```
-
-### Recovery Patterns
-
-```typescript
-const recoveryRoute = agent
-  .createRoute({
-    title: "Recovery Flow",
-
-    // Normal flow
-    initialStep: {
-      /* normal steps */
-    },
-
-    // Recovery branch
-  })
-  .branch([
-    {
-      name: "normal",
-      step: {
-        /* normal processing */
-      },
-    },
-    {
-      name: "recovery",
-      step: {
-        prompt: "I encountered an error. Let's start over.",
-        collect: ["resetConfirmation"],
-        when: "if error state is detected",
-        finalize: (context, data) => {
-          // Reset problematic data
-          resetSessionData(data);
-        },
-      },
-    },
-  ]);
-```
-
-## Performance Considerations
-
-### Efficient Traversal
-
-```typescript
-// The system optimizes by:
-// 1. Using visited sets to prevent infinite loops
-// 2. Early termination on END_ROUTE detection
-// 3. Lazy evaluation of skipIf conditions
-// 4. Caching of validation results
-
-// Example of efficient validation
-skipIf: (data) => {
-  // Cache expensive operations
-  if (data._skipCache === undefined) {
-    data._skipCache = expensiveCheck(data);
-  }
-  return data._skipCache;
-};
-```
-
-### Memory Management
-
-```typescript
-// Clean up temporary data
-finalize: (context, data) => {
-  // Remove temporary fields
-  delete data._tempField;
-  delete data._skipCache;
-
-  // Compress large data structures if needed
-  if (data.largeArray?.length > 100) {
-    data.largeArray = data.largeArray.slice(-50); // Keep recent items
-  }
-};
-```
-
-## Debugging Transitions
-
-### Step Inspection
-
-```typescript
-// Inspect route structure
-console.log(route.describe());
-
-// Check step validity
-const step = route.getStep("step-id");
-console.log("Should skip:", step.shouldSkip(session.data));
-console.log("Has requirements:", step.hasRequires(session.data));
-
-// Get all transitions
-const transitions = step.getTransitions();
-console.log(
-  "Available transitions:",
-  transitions.map((s) => s.id)
-);
-```
-
-### Transition Tracing
-
-```typescript
-// Enable debug logging
-const agent = new Agent({
-  name: "Debug Agent",
-  debug: true, // Enables transition logging
-  provider: provider,
-});
-
-// Logs will show:
-// [RoutingEngine] Found valid step after skipping: step-3
-// [RoutingEngine] Route complete: all data collected, END_ROUTE reached
-// [Agent] Entered step: step-3
-```
-
-### Testing Transitions
-
-```typescript
-// Test step conditions
-const testData = { name: "John", email: "john@test.com" };
-
-const shouldSkip = step.shouldSkip(testData);
-const hasRequirements = step.hasRequires(testData);
-
-console.log(`Step ${step.id}:`);
-console.log(`- Should skip: ${shouldSkip}`);
-console.log(`- Has requirements: ${hasRequirements}`);
-console.log(`- Can proceed: ${!shouldSkip && hasRequirements}`);
-```
-
-## Best Practices
-
-### Transition Design
-
-1. **Clear Dependencies**: Make `requires` relationships obvious
-2. **Smart Skipping**: Use `skipIf` to avoid redundant interactions
-3. **Fail Fast**: Validate early to prevent invalid states
-4. **Recovery Paths**: Plan for error conditions and recovery flows
-
-### Performance
-
-1. **Efficient Conditions**: Cache expensive `skipIf` evaluations
-2. **Minimal Dependencies**: Avoid overly complex requirement chains
-3. **Early Termination**: Use END_ROUTE appropriately
-4. **Memory Cleanup**: Remove temporary data in finalize hooks
-
-### Maintainability
-
-1. **Descriptive IDs**: Use meaningful step and route identifiers
-2. **Documentation**: Comment complex transition logic
-3. **Testing**: Cover edge cases in transition logic
-4. **Monitoring**: Track transition success rates
-
-### User Experience
-
-1. **Natural Flow**: Make transitions feel conversational
-2. **Context Preservation**: Maintain conversation context across steps
-3. **Error Recovery**: Handle failures gracefully with clear recovery paths
-4. **Progress Indication**: Show users where they are in multi-step processes
-
-The step transition system enables sophisticated conversation flows that can adapt to user context, handle complex branching logic, and maintain data integrity throughout multi-turn interactions.