Release Explorer
- Current release
- Release - 10/10/2025
- Release - 11/10/2025
- Release - 12/10/2025
- Release - 13/10/2025
- Release - 14/10/2025
- Release - 15/10/2025
- Release - 16/10/2025
- Release - 17/10/2025
- Release - 18/10/2025
- Release - 19/10/2025
- Release - 20/10/2025
- Release - 21/10/2025
- Release - 22/10/2025
- Release - 23/10/2025
- Release - 24/10/2025
- Release - 25/10/2025
- Release - 26/10/2025
- Release - 27/10/2025
- Release - 28/10/2025
- Release - 29/10/2025
- Release - 30/10/2025
- Release - 31/10/2025
- Release - 01/11/2025
- Release - 02/11/2025
- Release - 03/11/2025
- Release - 04/11/2025
- Release - 05/11/2025
- Release - 07/11/2025
- Release - 07/11/2025
- Release - 08/11/2025
- Release - 09/11/2025
- Release - 10/11/2025
- Release - 11/11/2025
- Release - 12/11/2025
Release Notes - ms-nodejs-backend and ms-nodejs-api-gateway (match-non-experts branch)
Overview
This release introduces a new feature enabling innovation managers and providers to add structured challenges into team funnel templates by step number. It also enhances AI-driven title generation for challenges by refining prompt syntax and response handling. Additional backend improvements include better filtering of audience data, logging enhancements for debugging, and refactoring of chat completion calls to support conditional JSON responses.
New Features
- Added TeamFunnelStepChallenge functionality:
- Innovation managers and providers can now add challenges linked to specific steps of team funnel templates.
- Automatic challenge title generation powered by ChatGPT, creating fitting and potentially catchy titles if none are provided.
- Created new Mongoose model, service, controller, and Express route to support challenge creation, retrieval by step name, and associated validations.
- Implemented strict authorization checks allowing only users with correct roles and ownership of templates to add challenges.
Improvements / Changes
- Refined the ChatGPT prompt for generating challenge titles to explicitly require JSON-formatted output, improving parse reliability.
- Extracted chat completion request options into a variable and introduced conditional setting of
response_formatto JSON, making the AI integration more flexible. - Improved filtering of audience levels by including only string-type IDs before fetching providers, ensuring correct provider queries.
- Enhanced logging around audience level processing and challenge title generation for better traceability during debugging.
- Initialized structures to store AI-generated expert squads and offer currencies, laying groundwork for advanced matching features.
- Small code cleanups including consistent use of await and streamlined service calls supporting the overall maintainability.
Bug Fixes
- Fixed incorrect use of audience levels when fetching providers by filtering to strings, preventing errors or unintended behavior.
- Resolved parsing issues related to challenge title generation prompts by clarifying expected ChatGPT response formats.
Technical / Developer Notes
- New TeamFunnelStepChallenge entities are linked by
stepNameIdand require referencing the existing TeamFunnelStepName model. - When calling the AI chat completions, the options object now conditionally sets response format which must be handled accordingly in consuming code.
- Authorization logic enforces that only users with role 'innovation manager' or 'provider' owning the respective template can manipulate challenges.
- The AI prompt for title generation must return JSON in the format
{title: "Fitting Title"}, which should be parsed safely.
Acknowledgements / Contributors
Thanks to the development team for their work on integrating AI-assisted title generation and enhancing challenge management workflows, improving innovation templates management significantly.