HOTFIX return run details

[vas3a]

This pull request introduces improvements to the AI workflow dispatch and processing logic, focusing on more robust handling of workflow run metadata, cancellation detection, and code simplification. The changes enhance the way workflow runs are tracked and how their statuses are determined and updated, while also improving error handling and logging for better maintainability.

AI workflow dispatch and metadata handling:

• The GiteaService.runDispatchWorkflow method now returns detailed metadata about dispatched workflow runs, including IDs and URLs, which are logged and stored in the database for better traceability. (src/shared/modules/global/gitea.service.ts, src/shared/modules/global/workflow-queue.handler.ts) [1] [2] [3] [4] [5] [6]

AI workflow run status and error handling:

• Improved logic for handling workflow job events: now, the handler checks for unexpected workflow run statuses and logs errors accordingly, and ignores specific job events like dump-workflow-context to avoid unnecessary processing. (src/shared/modules/global/workflow-queue.handler.ts)
• Simplified and clarified the workflow job event handling by removing redundant or obsolete code related to job count lookups and status updates. (src/shared/modules/global/workflow-queue.handler.ts) [1] [2]

AI reviewer decision logic:

• The AI reviewer decision logic now detects and handles cancelled workflow runs: if any run is cancelled, the total score is set to zero, and the submission is automatically marked as not passed with an appropriate reason. This ensures more accurate and reliable decision-making. (src/shared/modules/global/ai-reviewer-decision-maker.service.ts) [1] [2] [3] [4]