seoh0711/gemini-antigravity-study

Purpose and Topology

This repository builds a curriculum-driven learning and exam-practice application. It contains a React learner UI, two alternative backend implementations, and a staged content-production workspace. The deployable primary stack appears to be a Cloudflare Worker API plus static frontend; a Python backend provides a parallel FastAPI-style implementation.

Application Entry Points

• Frontend bootstraps from app/frontend/src/main.tsx, renders application routing in app/frontend/src/App.tsx, and defines page routes in app/frontend/src/router.tsx.

• Cloudflare API entry is app/worker/src/index.ts, with application wiring in app/worker/src/app.ts.

• Python API entry is app/backend/main.py; database and seed bootstrapping reside in app/backend/database.py and app/backend/seed.py.

• Worker deployment/runtime settings are defined in app/worker/wrangler.toml.

Learner Frontend

The Vite/React TypeScript client implements public/authenticated learning, quizzes, challenges, notes, and progress views. It centralizes API communication and domain types under app/frontend/src/api/client.ts and app/frontend/src/api/types.ts; authentication state and route protection are handled by app/frontend/src/auth/AuthContext.tsx and app/frontend/src/components/RequireAuth.tsx.

Core user journeys are represented by pages:

• Learning syllabus/topic delivery: app/frontend/src/pages/Learn.tsx, app/frontend/src/pages/LearnTopic.tsx

• Quiz execution: app/frontend/src/pages/Quiz.tsx

• Challenge discovery/detail: app/frontend/src/pages/Challenges.tsx, app/frontend/src/pages/ChallengeDetail.tsx

• Aggregated learner progress: app/frontend/src/pages/Stats.tsx

API and Domain Boundary

The Worker API separates HTTP route handlers by learner domain: app/worker/src/routes/content.ts, app/worker/src/routes/quiz.ts, app/worker/src/routes/progress.ts, and app/worker/src/routes/challenges.ts. Authentication is implemented independently in app/worker/src/auth.ts and exposed through app/worker/src/routes/auth.ts.

The Python backend mirrors this domain partition—app/backend/routers/content.py, app/backend/routers/quiz.py, app/backend/routers/progress.py, and app/backend/routers/auth.py—with shared models, request/response schemas, and serializers in app/backend/models.py, app/backend/schemas.py, and app/backend/serializers.py. This duplication is an important implementation boundary: clients should target one deployed API contract, not combine both backends.

Assessment and Progress Flow

A learner requests content or a question set, submits answers to quiz/challenge endpoints, then receives grading and persistence-backed progress/statistics updates. Grading logic is explicit in app/worker/src/grading.ts and app/backend/grading.py; questions, quizzes, challenges, and progress each have dedicated routes. The frontend’s app/frontend/src/pages/Quiz.tsx and app/frontend/src/components/QuestionCard.tsx form the main answer-capture and feedback surface.

Data and Infrastructure

The Cloudflare Worker uses database access through app/worker/src/db.ts, with schema initialization in app/worker/migrations/0001_init.sql. Seed data is generated for that runtime by app/worker/scripts/gen-seed.mjs. The Python alternative has its own persistence layer at app/backend/database.py. Cloudflare deployment guidance is present in DEPLOY.md and .claude/skills/cloudflare-deploy/SKILL.md.

Curriculum Content Pipeline

_workspace is a staged authoring/validation pipeline rather than application runtime code:

• Curriculum structure and plans: _workspace/00_architect/syllabus-tree.json, _workspace/00_architect/content-plan.md

• Authored lessons: _workspace/10_content/content.json

• Questions and challenges: _workspace/20_questions/questions.json, _workspace/20_questions/challenges.json

• Content and integration QA outputs: _workspace/30_qa/content-qa-report.md, _workspace/40_qa/integration-qa-report.md

This pipeline produces structured curriculum and assessment assets intended to seed or populate the application’s content-serving and assessment domains.