AI Course Reviewer

Automated SCORM e-learning course testing system using FastAPI, Playwright, and LangChain AI agents.

---

What This Project Does

This tool automatically tests SCORM e-learning courses by:

• Uploading and extracting SCORM ZIP packages
• Opening courses in a headless browser with a fake LMS
• Navigating through all slides automatically
• Running functional tests (buttons, navigation, media)
• Running content checks (spelling, grammar)
• Using AI agents to evaluate content quality
• Generating detailed reports

---

Tech Stack

• FastAPI - REST API framework
• Playwright - Browser automation
• LangChain - AI agent framework
• OpenAI / Anthropic - LLM providers
• pyspellchecker - Spelling checks
• language-tool-python - Grammar checks
• lxml - XML parsing

---

Project Structure

ai_course_reviewer/
│
├── main.py                 # FastAPI app entry point
├── config.py               # Settings and configuration
├── requirements.txt        # Python dependencies
├── .env                    # Secret keys
├── .gitignore              # Git ignore rules
│
├── endpoints/
│   └── review.py           # API routes 
│
└── src/
    ├── schemas.py          # Pydantic data models
    │
    ├── scorm/
    │   ├── extractor.py    # ZIP extraction
    │   └── parser.py       # Manifest XML parsing
    │
    ├── browser/
    │   ├── manager.py      # Playwright browser control
    │   └── scorm_stub.js   # Fake LMS SCORM API
    │
    ├── agents/
    │   ├── orchestrator.py # Main review controller
    │   ├── functional.py   # Functional testing agent
    │   ├── content.py      # Content quality agent
    │   └── ai_agents.py    # LangChain AI agents
    │
    └── reports/
        └── generator.py    # Report generation

---

File Descriptions

Root Files

File	Purpose
main.py	Creates FastAPI app, sets up routes and middleware
config.py	All configuration settings (browser, AI, paths)
requirements.txt	Python package dependencies
.env	Environment variables (API keys) - you create this

API Layer (endpoints/)

File	Purpose
review.py	REST endpoints: POST /upload, POST /start, GET /status, GET /report

Data Models (src/)

File	Purpose
schemas.py	Pydantic models for requests, responses, and internal data

SCORM Handling (src/scorm/)

File	Purpose
extractor.py	Extracts SCORM ZIP files, validates manifest exists
parser.py	Parses imsmanifest.xml to find launch file and course info

Browser (src/browser/)

File	Purpose
manager.py	Controls headless browser (navigate, click, screenshot)
scorm_stub.js	JavaScript injected to fake LMS API (window.API)

Agents (src/agents/)

File	Purpose
orchestrator.py	Main controller - coordinates entire review process
functional.py	Tests buttons, navigation, media, timers
content.py	Checks spelling, grammar, integrates AI agents
ai_agents.py	LangChain agents: Clarity, Accuracy, Tone, Accessibility, Structure

Reports (src/reports/)

File	Purpose
generator.py	Generates JSON, HTML, and text reports

---

API Endpoints

Method	Endpoint	Description
POST	/api/upload	Upload SCORM ZIP file
POST	/api/start/{id}	Start review process
GET	/api/status/{id}	Check review progress
GET	/api/report/{id}	Get final report

---

Setup

# Install dependencies
pip install -r requirements.txt

# Install browser
playwright install chromium

# Create .env file with your API key
echo "OPENAI_API_KEY=sk-your-key" > .env

# Run server
python main.py

---

License

MIT License