2 Commits

Author SHA1 Message Date
wangdl
f1eb99b6fa fix(quiz): complete cross-platform attempt and question type contract
Some checks failed
Deploy API Server / build-and-unit (push) Failing after 42s
Deploy API Server / current-integration (push) Has been skipped
Deploy API Server / backward-compat (push) Has been skipped
Deploy API Server / deploy (push) Has been skipped
- Add attempt lifecycle with per-attempt answer visibility (#336)
- Add atomic submission with updateMany CAS + $transaction (#337)
- Add iOS attempt-based question loading + local answer cache (#338)
- Freeze choice/fill/judge cross-platform contract with fill grading (#339)
- Add per-question-type learning analysis accuracyByQuestionType (#340)
- Add GATE E2E test suite with 17 scenarios (#341)
- CLEANUP-01: Remove illegal type → choice silent degradation
- CLEANUP-02: Integrate QuizQuestionType enum into iOS quiz views
- CLEANUP-03: Test verification, git delivery, targeted gate review

Co-Authored-By: Claude <noreply@anthropic.com>
2026-06-24 19:55:05 +08:00
wangdl
d368f1a97f feat(M-AI-07-03): quiz generation Executor + Validator + prompt/schema
Some checks failed
Deploy API Server / build-and-unit (push) Successful in 37s
Deploy API Server / backward-compat (push) Has been cancelled
Deploy API Server / deploy (push) Has been cancelled
Deploy API Server / current-integration (push) Has been cancelled
Resolves A1/A2: bring quiz prompt + output schema in-house from Runtime.
- quiz-generation.prompt.ts: system prompt with question design principles
- quiz-generation.schema.ts: Zod schema (choice/fill/judge types)
- Registered in PromptTemplateService with key 'quiz-generation' v1.0.0
- Executor: AiGateway call with knowledge items + generation params
- Validator: per-type validation (choice options/index, judge true/false)
- Engine dispatch for quiz_generation job type

Co-Authored-By: Claude <noreply@anthropic.com>
2026-06-22 20:40:12 +08:00