Human-in-the-Loop (HITL) #52
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Add cleaning step to remove newline spaces - Compare cleaned JSON strings for more accurate matching
- Comment out @babel/preset-env and import.meta plugin - Improve Node.js debugging by reducing code transpilation
…tasks - Added state variables and functions for handling feedback submission and task validation - Added UI elements for task selection, feedback entry, and submission - Added UI elements for displaying tasks, approving or requesting revision, and showing feedback history
…alidation, and status monitoring.
…m task execution. Allows also to preserve memory in the same execution. - Modified executeTask method - Added executableAgent getter - Added initialize, workOnTask, workOnFeedback, prepareAgentForTask, and agenticLoop methods.
- Added feedback history processing - Improved error handling
…for providing feedback and validating tasks and preparation of new log entries.
- Replace log.task.status with log.taskStatus to prevent mutation issues - Update related test snapshots
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Human-in-the-Loop (HITL) Implementation for Enhanced Task Oversight
This PR implements Human-in-the-Loop (HITL) support to allow for manual intervention in tasks that require human oversight, addressing issue #35. The changes enhance AgenticJS's ability to handle complex tasks that benefit from human judgment while maintaining automated efficiency.
Key Changes:
AgentsBoardDebugger Component: Added functionality for providing feedback and validating tasks, including UI elements for task selection, feedback entry, and submission.
React Playground: Added a new story for ResumeCreationTeam with HITL integration using OpenAI models.
Agent and Task Classes: Updated with new methods for workflow management, feedback provision, and task validation.
Task and Team Stores: Implemented new enums and logic for handling task completion, errors, validation, and feedback processing.
Workflow Controller: Added helper function to check agent availability and manage task status transitions.
Subscribers: Updated taskSubscriber and teamSubscriber to support new task statuses and provide more specific status messages.
Enums: Added new enums for feedback and task statuses (AWAITING_VALIDATION, VALIDATED, PENDING, PROCESSED).
Test Updates: Modified various test files to reflect new HITL functionality, including updates to snapshots and the addition of new test cases for OpenAI agents with HITL features.
API Mocks: Updated request/response mocks for OpenAI and Gemini to support HITL testing scenarios.
Utility Updates: Enhanced moscaFetch function to improve request body comparison accuracy in tests.
These changes collectively implement a robust HITL system within AgenticJS, allowing for seamless integration of human oversight in AI-driven workflows. This enhancement improves the accuracy, reliability, and ethical considerations of complex or sensitive operations managed by the library.