"""
api.py
======
FastAPI wrapper around the 3-phase AI code detector.

Endpoints:
    GET  /          → welcome + available endpoints
    GET  /health    → status check
    POST /detect    → run detection on submitted code

Usage:
    cd C:\\Users\\DEY STORE\\ai-code-detector
    uvicorn api:app --reload --port 8000

    Then visit http://localhost:8000/docs for the Swagger UI.

Test from PowerShell:
    $body = @{ code = "def hello():\n    print('hi')\n    return 42\n\nhello()\nprint('done')\nprint('end')\nprint('really')\nprint('yes')"; language = "python" } | ConvertTo-Json
    Invoke-RestMethod -Uri "http://localhost:8000/detect" -Method POST -Body $body -ContentType "application/json"
"""

import sys
import os
from pathlib import Path

# Make project root importable
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel, Field

from orchestrator import detect, build_report


app = FastAPI(
    title="AI Code Detector",
    description=(
        "Three-phase detector for AI-generated source code.\n\n"
        "Phase 1: Stylometric analysis (Python only, AUC 0.966)\n"
        "Phase 2: Multi-LLM rewrite similarity (AUC 0.653)\n"
        "Phase 3: Neural code classifier (AUC 0.753)\n\n"
        "Based on: 'Uncovering LLM-Generated Code' (AAAI 2025)"
    ),
    version="1.0.0",
)


# ── Request / Response models ─────────────────────────────────────────────────

class DetectRequest(BaseModel):
    code: str = Field(
        ...,
        min_length=1,
        description="Source code to analyze",
        json_schema_extra={
            "example": "import sys\nfrom collections import Counter\n\ndef solve(arr):\n    if not arr:\n        return 0\n    return max(Counter(arr).values())\n\nif __name__ == '__main__':\n    arr = list(map(int, sys.stdin.read().split()))\n    print(solve(arr))"
        },
    )
    language: str = Field(
        "python",
        description="Programming language of the code",
        json_schema_extra={"example": "python"},
    )


class DetectResponse(BaseModel):
    verdict: str
    confidence: float | None
    p1: float | None = None
    p2: float | None = None
    p3: float | None = None
    suspected_model: str | None = None
    language: str
    phases_used: int | None = None
    report: str
    timings: dict | None = None
    total_time_seconds: float | None = None
    errors: dict | None = None


# ── Endpoints ─────────────────────────────────────────────────────────────────

@app.get("/")
def root():
    return {
        "name": "AI Code Detector",
        "version": "1.0.0",
        "status": "running",
        "endpoints": {
            "/detect": "POST — submit code for detection",
            "/health": "GET — health check",
            "/docs": "GET — Swagger UI (interactive API docs)",
        },
    }


@app.get("/health")
def health():
    return {"status": "ok"}


@app.post("/detect", response_model=DetectResponse)
def detect_endpoint(req: DetectRequest):
    """
    Analyze a code snippet for AI authorship.

    Runs up to 3 detection phases in parallel and returns a combined verdict.
    Phase 2 requires a Groq API key and may take 5-15 seconds.
    """
    if len(req.code.strip()) < 10:
        raise HTTPException(status_code=400, detail="Code too short (< 10 chars)")

    try:
        result = detect(req.code, req.language)
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"Detection failed: {e}")

    result["report"] = build_report(result)
    return result


# ── Run directly ──────────────────────────────────────────────────────────────

if __name__ == "__main__":
    import uvicorn
    print("Starting API server at http://localhost:8000")
    print("Swagger UI at http://localhost:8000/docs")
    uvicorn.run("api:app", host="0.0.0.0", port=8000, reload=True)