app.py

"""
Maria PT - HuggingFace Spaces Backend
======================================
Auth: SHA-256 hash check (primary) or Cloudflare Referer/Domain check (fallback)
Endpoints: GET /ping  |  POST /base_start
"""

import os
import hashlib
import logging
import httpx
from fastapi import FastAPI, Request, HTTPException, status
from fastapi.responses import JSONResponse
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from pymongo import MongoClient
from pymongo.errors import ConnectionFailure, OperationFailure
from typing import Optional

# ─────────────────────────────────────────────
# Logging
# ─────────────────────────────────────────────
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger("maria_pt")

# ─────────────────────────────────────────────
# Environment / Secrets  (set in HF Space Settings → Secrets)
# ─────────────────────────────────────────────
def _require_secret(key: str) -> str:
    value = os.environ.get(key)
    if not value:
        raise RuntimeError(f"Missing required secret: {key}")
    return value

EXPECTED_HASH    = _require_secret("EXPECTED_HASH")
MONGO_PASSWORD   = _require_secret("MONGO_PASSWORD")
MONGO_DB         = os.environ.get("MONGO_DB",   "MariaPTDB")
MONGO_COLLECTION = os.environ.get("MONGO_COLL", "MariaPTColl")
MONGO_URI        = os.environ.get("MONGO_URI") or \
                   f"mongodb+srv://testuser:{MONGO_PASSWORD}@cluster0.ntz2mpi.mongodb.net/"

# Cloudflare Turnstile secret (placeholder – paste your real secret in HF Secrets)
CF_TURNSTILE_SECRET = os.environ.get("CF_TURNSTILE_SECRET", "PLACEHOLDER_CF_TURNSTILE_SECRET")
ALLOWED_DOMAIN      = os.environ.get("ALLOWED_DOMAIN", "buildwithsupratim.github.io")

# ─────────────────────────────────────────────
# FastAPI App
# ─────────────────────────────────────────────
app = FastAPI(title="Maria PT API", version="1.0.0", docs_url=None, redoc_url=None)

app.add_middleware(
    CORSMiddleware,
    allow_origins=["https://buildwithsupratim.github.io"],
    allow_credentials=True,
    allow_methods=["GET", "POST"],
    allow_headers=["*"],
)

# ─────────────────────────────────────────────
# MongoDB Client (lazy singleton)
# ─────────────────────────────────────────────
_mongo_client: Optional[MongoClient] = None

def get_mongo_collection():
    global _mongo_client
    if _mongo_client is None:
        _mongo_client = MongoClient(MONGO_URI, serverSelectionTimeoutMS=5000)
    db   = _mongo_client[MONGO_DB]
    coll = db[MONGO_COLLECTION]
    # Ensure capped (100 MB) – silently ignored if collection already exists
    try:
        db.create_collection(
            MONGO_COLLECTION,
            capped=True,
            size=100_000_000,
        )
    except Exception:
        pass
    return coll


# ─────────────────────────────────────────────
# Auth Helpers
# ─────────────────────────────────────────────
def _hash_auth_code(auth_code: str) -> str:
    return hashlib.sha256(auth_code.encode()).hexdigest()


def _primary_auth(request: Request) -> bool:
    """Check SHA-256 of auth_code header against EXPECTED_HASH."""
    auth_code = request.headers.get("auth_code") or request.headers.get("Auth-Code")
    if not auth_code:
        return False
    return _hash_auth_code(auth_code) == EXPECTED_HASH


async def _cloudflare_domain_check(request: Request) -> bool:
    """
    Cloudflare-backed fallback:
    1. Extract the Referer / Origin header.
    2. Verify it belongs to ALLOWED_DOMAIN (domain-level check).
    3. Optionally verify a Cloudflare Turnstile token if present in headers.
    """
    # ── Step 1: Domain/Referer check ──────────────────────────────────────
    referer = request.headers.get("referer", "")
    origin  = request.headers.get("origin", "")

    referer_ok = ALLOWED_DOMAIN in referer
    origin_ok  = ALLOWED_DOMAIN in origin

    if not (referer_ok or origin_ok):
        logger.warning("Blocked – domain not allowed. Referer=%s Origin=%s", referer, origin)
        return False

    # ── Step 2: Cloudflare Turnstile token (optional header) ─────────────
    # If the front-end sends a CF-Turnstile-Token header we verify it.
    # If no token is sent we fall back to domain-only gating.
    cf_token = request.headers.get("CF-Turnstile-Token")
    if cf_token and CF_TURNSTILE_SECRET != "PLACEHOLDER_CF_TURNSTILE_SECRET":
        try:
            async with httpx.AsyncClient(timeout=5.0) as client:
                resp = await client.post(
                    "https://challenges.cloudflare.com/turnstile/v0/siteverify",
                    data={
                        "secret":   CF_TURNSTILE_SECRET,
                        "response": cf_token,
                        "remoteip": request.client.host,
                    },
                )
            result = resp.json()
            if not result.get("success"):
                logger.warning("Cloudflare Turnstile rejected: %s", result)
                return False
        except Exception as exc:
            logger.error("Turnstile verification error: %s", exc)
            return False

    return True


async def require_auth(request: Request):
    """
    Gate every protected endpoint.
    Primary  → SHA-256 hash check on auth_code header.
    Fallback → Cloudflare domain / Turnstile check.
    Block    → 403 + client IP is logged (extend here for real IP-ban storage).
    """
    if _primary_auth(request):
        return  # ✅ Primary auth passed

    if await _cloudflare_domain_check(request):
        return  # ✅ Cloudflare check passed

    # ❌ Both checks failed → block
    client_ip = request.client.host
    logger.warning("BLOCKED IP: %s", client_ip)
    raise HTTPException(
        status_code=status.HTTP_403_FORBIDDEN,
        detail="Access denied. Invalid credentials or unauthorized origin.",
    )


# ─────────────────────────────────────────────
# Request Schema
# ─────────────────────────────────────────────
class BaseStartRequest(BaseModel):
    request: str        # e.g. "send_code"
    student_name: str   # e.g. "Greti"


# ─────────────────────────────────────────────
# Response Builders
# ─────────────────────────────────────────────
def build_learning_path(doc: dict, student_name: str) -> dict:
    """
    Merge the MongoDB document with the student_name from the API call,
    then build the full learning_path response with blank/initialised fields.
    """
    curriculum_objectives = doc.get("curriculum_objectives", [])

    # ── Normalise curriculum (topic → topics key) ─────────────────────────
    normalised_curriculum = []
    for item in curriculum_objectives:
        normalised_curriculum.append({
            "topics":              item.get("topic", ""),
            "content":             item.get("content", ""),
            "learning_objectives": item.get("learning_objectives", []),
        })

    # ── assessment_stages: initialise with FIRST topic only ───────────────
    first_topic = curriculum_objectives[0] if curriculum_objectives else {}
    first_learning_objectives = [
        {
            "goal":         goal,
            "teach":        "not_complete",
            "re_teach":     "not_complete",
            "show_and_tell":"not_complete",
            "assess":       "not_complete",
        }
        for goal in first_topic.get("learning_objectives", [])
    ]

    current_learning = [
        {
            "topic":               first_topic.get("topic", ""),
            "content":             first_topic.get("content", ""),
            "learning_objectives": first_learning_objectives,
        }
    ] if first_topic else []

    return {
        "learning_path": {
            "board":                doc.get("board", ""),
            "class":                doc.get("class", ""),
            "subject":              doc.get("subject", ""),
            "student_name":         student_name,          # from API call
            "teacher_persona":      doc.get("teacher_persona", ""),
            "curriculum_objectives": normalised_curriculum,
            "chat_history":         [],                    # blank – yet to start
            "scratchpad":           [],                    # blank – yet to start
            "assessment_stages": {
                "current_learning": current_learning,
            },
        }
    }


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

@app.get("/ping")
async def ping(request: Request):
    """Health-check endpoint – wakes the Space if sleeping."""
    await require_auth(request)
    return JSONResponse(content={"status": "alive"})


@app.post("/base_start")
async def base_start(request: Request, body: BaseStartRequest):
    """
    Accepts { request, student_name }, queries MongoDB for matching request,
    and returns an initialised learning_path JSON.
    """
    await require_auth(request)

    # ── MongoDB Lookup ────────────────────────────────────────────────────
    try:
        collection = get_mongo_collection()
        doc = collection.find_one(
            {"request": body.request},
            {"_id": 0},          # exclude Mongo internal _id
        )
    except (ConnectionFailure, OperationFailure) as exc:
        logger.error("MongoDB error: %s", exc)
        raise HTTPException(
            status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
            detail="Database connection error. Please try again later.",
        )

    if doc is None:
        raise HTTPException(
            status_code=status.HTTP_404_NOT_FOUND,
            detail=f"No curriculum found for request '{body.request}'.",
        )

    # ── Build & Return Response ───────────────────────────────────────────
    response_payload = build_learning_path(doc, body.student_name)
    return JSONResponse(content=response_payload)


# ─────────────────────────────────────────────
# Entry point (for local testing)
# ─────────────────────────────────────────────
if __name__ == "__main__":
    import uvicorn
    uvicorn.run("app:app", host="0.0.0.0", port=7860, reload=True)