Conversation
|
An early comment here: I haven't reviewed the code, but as an early comment, I suggest removing all unnecessary features from this proof of concept (for example, assertions) and updating the PR cover to reflect your recent changes. Again, thanks for this PR ✨ I'll be reviewing it in more depth soon. |
|
I would say that what I'm missing the most in this PR is documentation on how to use it. Since I already know how to install the dependencies from the original AI extensions plugin, I'm I was able to, but I think a readme explaining how you get the UI components to connect that would be highly needed. |
There was a problem hiding this comment.
I think the package.json is missing, something simple, the reasoning is 1) we are using a dependency need to specify it and second to use this as a module, this is the important things:
{
"name": "@openedx/ai-badges-creator",
// ....
"peerDependencies": {
"@openedx/openedx-ai-extensions-ui": "^0.0.1"
},
//.....
}| @@ -1,10 +1,35 @@ | |||
| import RedLine from './plugin'; | |||
| import AIRequestBadgesComponent from './components/AIRequestBadgesComponent'; | |||
| animation="border" | ||
| variant="secondary" | ||
| size="sm" | ||
| className="me-2" |
There was a problem hiding this comment.
for all the me classes replace them with mr (me is Bootstrap no Paragon)
| onClick={onAskAI} | ||
| disabled={disabled || isLoading} | ||
| iconBefore={Send} | ||
| style={{ |
4 participants
This pull request introduces a foundational implementation of the AI-powered badges feature, integrating backend workflow orchestration, API structure, and frontend component registration. The changes establish the core architecture for badge generation using LLM processors and enable frontend-backend communication for badge requests and responses.
Backend architecture and workflow orchestration:
BadgeOrchestrator) that manages the badge generation workflow, including session management and mock course context for testing. This orchestrator uses a processor to generate badge definitions and serializes the result for API consumption. [1] [2] [3]BadgeProcessor) that leverages an LLM to generate badge definitions following the Open Badges 2.0 spec, with clear constraints and deterministic output. [1] [2]BadgeClassPublicSerializer) for badge data, ensuring output conforms to a static, model-independent JSON structure.API and URL structure:
Frontend integration and component registration:
AIRequestBadgesComponent) for initiating badge requests, with loading and disabled states, and registered it with the AI extensions UI registry for dynamic backend-driven UI configuration. [1] [2]Workflow configuration and settings:
badges_base.json) that defines the orchestrator, processor, and frontend actuator configuration for badge generation workflows.Other foundational changes:
These changes lay the groundwork for AI-driven badge generation, with clear separation of concerns and extensibility for both backend workflows and frontend UI components.