Spaces:

alihamzach
/

skinmodel

Sleeping

skinmodel / API_SPECIFICATION.md
alihamzach
Update API docs with current HF deployment and new endpoints
603a537 4 months ago
47.7 kB
# Skin Disease Detection API — Full Specification

> **Base URL:** `https://alihamzach-skinmodel.hf.space`  
> **API Prefix:** `/v1`  
> **Model:** EfficientNetB0 (224×224, 22 disease classes)  
> **Auth:** JWT (access + refresh tokens)  
> **Rate Limiting:** IP-based + user-based tiered quotas  
> **Database:** PostgreSQL (Neon) via SQLAlchemy async + asyncpg

---

## Table of Contents

1. [Authentication & User Management](#1-authentication--user-management)
2. [Image Scanning & Detection](#2-image-scanning--detection)
3. [Recommendations & Personalization](#3-recommendations--personalization)
4. [Monetization & Subscriptions](#4-monetization--subscriptions)
5. [Database Schema](#5-database-schema)
6. [Rate Limiting Logic](#6-rate-limiting-logic)
7. [Error Handling](#7-error-handling)
8. [Middleware & Configuration](#8-middleware--configuration)

---

## 1. Authentication & User Management

### 1.1 Signup

Creates a new user account with email + password.

```
POST /auth/signup
```

**Request Body:**
```json
{
  "email": "user@example.com",
  "password": "Str0ng!Pass123",
  "full_name": "Jane Doe",
  "medical_history": {
    "allergies": ["penicillin", "sulfa"],
    "chronic_conditions": ["eczema"],
    "current_medications": ["cetirizine"]
  }
}
```

**Validation Rules:**
- `email`: valid email format, max 255 chars, unique in system
- `password`: min 8 chars, at least 1 uppercase, 1 lowercase, 1 digit, 1 special char
- `full_name`: 1–100 chars, letters/spaces/hyphens only
- `medical_history`: optional JSON object

**Response `201 Created`:**
```json
{
  "id": "uuid",
  "email": "user@example.com",
  "full_name": "Jane Doe",
  "is_verified": false,
  "created_at": "2026-05-17T10:30:00Z"
}
```

**Error `409 Conflict`:**
```json
{
  "detail": "An account with this email already exists."
}
```

**Error `422 Unprocessable Entity`:**
```json
{
  "detail": [
    { "loc": ["body", "password"], "msg": "Password must contain at least one uppercase letter" }
  ]
}
```

---

### 1.2 Login

Authenticates user and returns JWT tokens.

```
POST /auth/login
```

**Request Body:**
```json
{
  "email": "user@example.com",
  "password": "Str0ng!Pass123"
}
```

**Response `200 OK`:**
```json
{
  "access_token": "eyJhbGciOiJIUzI1NiIs...",
  "refresh_token": "eyJhbGciOiJIUzI1NiIs...",
  "token_type": "bearer",
  "expires_in": 3600,
  "user": {
    "id": "uuid",
    "email": "user@example.com",
    "full_name": "Jane Doe",
    "subscription_plan": "free",
    "scans_remaining_today": 5
  }
}
```

**Error `401 Unauthorized`:**
```json
{
  "detail": "Invalid email or password."
}
```

**Token Specs:**
| Token | Lifetime | Payload |
|-------|----------|---------|
| Access | 1 hour | `sub` (user_id), `exp`, `iat`, `type: "access"` |
| Refresh | 30 days | `sub` (user_id), `exp`, `iat`, `type: "refresh"` |

---

### 1.3 Social Login (OAuth 2.0)

Initiates or completes social login via Google or Facebook.

```
POST /auth/social/{provider}
```

**Path Parameters:**
- `provider`: `"google"` | `"facebook"`

**Request Body:**
```json
{
  "access_token": "ya29.a0AfH6SMC..."
}
```

**Server-Side Flow:**
1. Receive OAuth access token from client
2. Verify token with Google/Facebook tokeninfo endpoint
3. Extract email, name, profile photo from provider
4. If email exists → link social account + login
5. If email new → create user with `auth_provider` = provider + `social_id` = provider user ID

**Response `200 OK`** (same shape as login response)

**Error `401 Unauthorized`:**
```json
{
  "detail": "Invalid or expired social token."
}
```

---

### 1.4 Forgot Password

Initiates password reset email flow.

```
POST /auth/forgot-password
```

**Request Body:**
```json
{
  "email": "user@example.com"
}
```

**Flow:**
1. Validate email exists (silent fail if not — security best practice)
2. Generate cryptographically secure 6-digit code (or UUIDv4 token)
3. Store hashed token in `password_reset_tokens` table with 15-min expiry
4. Send email with reset link/code

**Response `200 OK` (always):**
```json
{
  "message": "If an account with that email exists, a password reset link has been sent."
}
```

---

### 1.5 Reset Password

Completes password reset using token from email.

```
POST /auth/reset-password
```

**Request Body:**
```json
{
  "token": "a1b2c3d4-e5f6-...",
  "new_password": "NewStr0ng!Pass456"
}
```

**Validation:**
- Token must exist in DB, not expired, not used
- Token is marked `is_used = true` after successful reset (one-time use)
- Password follows same rules as signup

**Response `200 OK`:**
```json
{
  "message": "Password has been reset successfully."
}
```

**Error `400 Bad Request`:**
```json
{
  "detail": "Invalid or expired reset token."
}
```

---

### 1.6 Change Password (Authenticated)

```
PUT /auth/change-password
```

**Headers:** `Authorization: Bearer <access_token>`

**Request Body:**
```json
{
  "current_password": "OldPass123!",
  "new_password": "NewStr0ng!Pass456"
}
```

**Response `200 OK`:**
```json
{
  "message": "Password changed successfully."
}
```

**Error `401 Unauthorized`:**
```json
{
  "detail": "Current password is incorrect."
}
```

---

### 1.7 Refresh Token

```
POST /auth/refresh
```

**Request Body:**
```json
{
  "refresh_token": "eyJhbGciOiJIUzI1NiIs..."
}
```

**Response `200 OK`:**
```json
{
  "access_token": "eyJhbGciOiJIUzI1NiIs...",
  "expires_in": 3600
}
```

---

### 1.8 Update Profile (Authenticated)

```
PUT /auth/profile
```

**Headers:** `Authorization: Bearer <access_token>`

**Request Body:**
```json
{
  "full_name": "Jane Smith",
  "profile_photo": "data:image/png;base64,iVBOR...",
  "medical_history": {
    "allergies": ["penicillin"],
    "chronic_conditions": ["eczema", "psoriasis"],
    "current_medications": ["diphenhydramine"],
    "skin_type": "oily",
    "known_sensitivities": ["fragrance", "parabens"]
  }
}
```

**Response `200 OK`:**
```json
{
  "id": "uuid",
  "email": "user@example.com",
  "full_name": "Jane Smith",
  "profile_photo_url": "https://cdn.skindetect.com/users/uuid/photo.jpg",
  "medical_history": { ... },
  "updated_at": "2026-05-17T11:00:00Z"
}
```

---

## 2. Image Scanning & Detection

### 2.1 Scan Image

Upload a skin image for AI-based disease detection.

```
POST /scans
```

**Unauthenticated users:** IP-based rate limit of **3 total scans per IP**.  
**Authenticated users:** Daily limit based on subscription plan (default: **5/day** for free tier).

**Headers:** `Authorization: Bearer <access_token>` *(optional for unauthenticated)*

**Request Body** (`multipart/form-data`):
| Field | Type | Required | Description |
|-------|------|----------|-------------|
| `image` | file | yes | JPEG or PNG, max 10MB, min 224×224px |
| `location` | string | no | Lat,lng for geo-recommendations (e.g. `"40.7128,-74.0060"`) |

**Processing Pipeline:**
1. Validate file type and size
2. Check rate limits (IP for unauth, user quota for auth)
3. Upload image to CDN (S3/GCS) with UUID filename
4. Preprocess: resize to 224×224, normalize to [0,1]
5. Run model inference → softmax probabilities over 22 classes
6. Apply confidence threshold (default: 15%) to filter low-confidence results
7. Log scan record in database
8. Increment quota counters
9. Return results

**Response `200 OK`:**
```json
{
  "items": [
    {
      "scan_id": "uuid",
      "title": "skin_photo.jpg",
      "image_url": "https://cdn.skindetect.com/scans/uuid/thumb.jpg",
      "primary_detection": {
        "disease": "Eczema",
        "confidence": 87.32,
        "urgency_level": "consult_recommended"
      },
      "all_results": [
        { "disease": "Eczema", "confidence": 87.32, "severity": "moderate" },
        { "disease": "Psoriasis", "confidence": 8.15 },
        { "disease": "Acne", "confidence": 2.41 }
      ],
      "created_at": "2026-05-17T10:30:00Z"
    }
  ],
  "total": 42,
  "page": 1,
  "page_size": 20,
  "total_pages": 3
}
```

**Error `401 Unauthorized`:**
```json
{
  "detail": "Authentication required to view scan history."
}
```

---

### 2.3 Get Single Scan Detail (Authenticated)

```
GET /scans/{scan_id}
```

**Headers:** `Authorization: Bearer <access_token>`

**Response `200 OK`:** Full scan result (same shape as scan response)

**Error `404 Not Found`:**
```json
{
  "detail": "Scan not found."
}
```

---

### 2.4 Rename Scan (Authenticated)

Updates the title of an existing scan.

```
PATCH /v1/scans/{scan_id}
```

**Headers:** `Authorization: Bearer <access_token>`

**Request Body:**
```json
{
  "title": "My updated scan name"
}
```

**Response `200 OK`:**
```json
{
  "scan_id": "uuid",
  "title": "My updated scan name"
}
```

**Error `404 Not Found`:**
```json
{
  "detail": "Scan not found."
}
```

---

### 2.5 Delete Scan (Authenticated)

Permanently deletes a scan record.

```
DELETE /v1/scans/{scan_id}
```

**Headers:** `Authorization: Bearer <access_token>`

**Response `204 No Content`** (no body)

**Error `404 Not Found`:**
```json
{
  "detail": "Scan not found."
}
```

---

## 3. Recommendations & Personalization

### 3.1 Dermatologist Recommendations

Returns a list of dermatologists ranked by proximity to the user's IP location.

```
GET /recommendations/dermatologists
```

**Headers:** `Authorization: Bearer <access_token>` *(optional — wider results if authenticated)*

**Query Parameters:**
| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| `lat` | float | no | IP geo-lookup | Latitude |
| `lng` | float | no | IP geo-lookup | Longitude |
| `radius_km` | int | no | 50 | Search radius |
| `specialization` | string | no | — | Filter by specialty (e.g. "pediatric_dermatology") |
| `page` | int | no | 1 | — |
| `page_size` | int | no | 20 | — |

**Response `200 OK`:**
```json
{
  "items": [
    {
      "id": "uuid",
      "name": "Dr. Sarah Johnson",
      "specialization": ["general_dermatology", "pediatric_dermatology"],
      "address": "123 Medical Plaza, Suite 200, New York, NY 10001",
      "phone": "+1-212-555-0198",
      "email": "dr.johnson@dermclinic.com",
      "website": "https://dermclinic.com/dr-johnson",
      "rating": 4.8,
      "review_count": 124,
      "distance_km": 2.3,
      "coordinates": { "lat": 40.7580, "lng": -73.9855 },
      "available_slots": ["2026-05-18T09:00", "2026-05-18T10:30"],
      "insurance_accepted": ["Aetna", "Cigna", "UnitedHealthcare"]
    }
  ],
  "user_location": { "lat": 40.7484, "lng": -73.9967, "source": "ip" },
  "total": 12,
  "page": 1,
  "page_size": 20
}
```

**Location Resolution Priority:**
1. Explicit `lat`/`lng` query params (highest priority)
2. User's saved location from profile (if authenticated)
3. IP geolocation lookup (MaxMind GeoIP or similar)

---

### 3.2 Skin Care Product Recommendations

Returns product/medicine recommendations. May include affiliate links.

```
GET /recommendations/products
```

**Headers:** `Authorization: Bearer <access_token>` *(optional)*

**Query Parameters:**
| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| `condition` | string | no | — | Target condition (e.g. "Eczema") |
| `product_type` | string | no | — | `medicine` \| `moisturizer` \| `cleanser` \| `sunscreen` \| `treatment` |
| `lat` | float | no | IP geo | For pharmacy availability |
| `lng` | float | no | IP geo | — |
| `page` | int | no | 1 | — |
| `page_size` | int | no | 20 | — |

**Response `200 OK`:**
```json
{
  "items": [
    {
      "id": "uuid",
      "name": "Cetaphil Moisturizing Cream",
      "brand": "Cetaphil",
      "category": "moisturizer",
      "type": "over_the_counter",
      "description": "Non-greasy, fragrance-free moisturizer for sensitive skin.",
      "key_ingredients": ["glycerin", "petrolatum", "dimethicone"],
      "recommended_for": ["Eczema", "Psoriasis", "dry_skin"],
      "average_rating": 4.6,
      "price": {
        "amount": 14.99,
        "currency": "USD",
        "retailer": "Walmart"
      },
      "affiliate_link": "https://www.amazon.com/dp/B001B4...?tag=skindetect-20",
      "image_url": "https://cdn.skindetect.com/products/uuid/image.jpg",
      "medical_rating": 4.5,
      "side_effects": ["rare: mild irritation"],
      "prescription_required": false
    }
  ],
  "total": 8,
  "page": 1,
  "page_size": 20
}
```

**Affiliate Link Strategy:**
- Append tracking tag `?tag=skindetect-{source}-{user_id_hash}` to all product links
- Link to Amazon Pharmacy, Walmart, CVS, Walgreens affiliate programs
- Track clicks via redirect service (`/r/{product_id}`) for analytics

---

### 3.3 Clinic Recommendations

Returns nearby skin clinics/hospitals.

```
GET /recommendations/clinics
```

**Headers:** `Authorization: Bearer <access_token>` *(optional)*

**Query Parameters:** (same location params as dermatologists)

**Response `200 OK`:**
```json
{
  "items": [
    {
      "id": "uuid",
      "name": "Manhattan Dermatology Center",
      "type": "clinic",
      "address": "500 Fifth Avenue, Floor 15, New York, NY 10110",
      "phone": "+1-212-555-0246",
      "distance_km": 0.8,
      "rating": 4.5,
      "review_count": 89,
      "services": ["general_dermatology", "mohs_surgery", "cosmetic"],
      "open_hours": {
        "monday": "08:00-18:00",
        "tuesday": "08:00-18:00",
        "wednesday": "08:00-20:00",
        "thursday": "08:00-18:00",
        "friday": "08:00-16:00",
        "saturday": "09:00-13:00",
        "sunday": "closed"
      },
      "accepts_insurance": true,
      "walkins_accepted": false,
      "coordinates": { "lat": 40.7527, "lng": -73.9822 }
    }
  ],
  "total": 5,
  "page": 1,
  "page_size": 20
}
```

---

## 4. Monetization & Subscriptions

### 4.1 List Subscription Plans

```
GET /subscriptions/plans
```

**Response `200 OK`:**
```json
{
  "plans": [
    {
      "id": "plan_free",
      "name": "Free",
      "price_cents": 0,
      "currency": "USD",
      "billing_interval": "month",
      "features": {
        "daily_scans": 5,
        "scan_history_days": 30,
        "detailed_results": true,
        "product_recommendations": true,
        "dermatologist_referrals": false,
        "priority_support": false,
        "data_export": false
      }
    },
    {
      "id": "plan_basic",
      "name": "Basic",
      "price_cents": 499,
      "currency": "USD",
      "billing_interval": "month",
      "features": {
        "daily_scans": 20,
        "scan_history_days": 365,
        "detailed_results": true,
        "product_recommendations": true,
        "dermatologist_referrals": true,
        "priority_support": false,
        "data_export": false
      },
      "stripe_price_id": "price_1ABC2Def3Ghi4Jkl"
    },
    {
      "id": "plan_premium",
      "name": "Premium",
      "price_cents": 999,
      "currency": "USD",
      "billing_interval": "month",
      "features": {
        "daily_scans": 100,
        "scan_history_days": null,
        "detailed_results": true,
        "product_recommendations": true,
        "dermatologist_referrals": true,
        "priority_support": true,
        "data_export": true
      },
      "stripe_price_id": "price_5Mno6Pqr7Stu8Vwx"
    }
  ]
}
```

---

### 4.2 Create Checkout Session

Initiates Stripe Checkout Session for subscription purchase.

```
POST /subscriptions/create-checkout-session
```

**Headers:** `Authorization: Bearer <access_token>`

**Request Body:**
```json
{
  "plan_id": "plan_basic",
  "success_url": "https://skindetect.com/subscriptions/success?session_id={CHECKOUT_SESSION_ID}",
  "cancel_url": "https://skindetect.com/subscriptions/cancel"
}
```

**Response `200 OK`:**
```json
{
  "checkout_url": "https://checkout.stripe.com/c/pay/cs_test_..."
}
```

**Error `400 Bad Request`:**
```json
{
  "detail": "Invalid plan ID."
}
```

---

### 4.3 Stripe Webhook

Receives Stripe webhook events for subscription lifecycle management.

```
POST /subscriptions/webhook
```

**Headers:** `Stripe-Signature: <webhook_signature>`

**No auth** (verified via Stripe signature).

**Events Handled:**
| Event | Action |
|-------|--------|
| `checkout.session.completed` | Activate subscription, set end_date |
| `invoice.paid` | Renew subscription period |
| `invoice.payment_failed` | Send payment failure email, mark subscription past_due |
| `customer.subscription.updated` | Sync plan changes |
| `customer.subscription.deleted` | Cancel subscription, revert to free plan |
**Response `200 OK`:**
```json
{
  "received": true
}
```
---
### 4.4 Get My Subscription
Returns the authenticated user's current subscription details.
```
GET /subscriptions/my
```
**Headers:** `Authorization: Bearer <access_token>`
**Response `200 OK`:**
```json
{
  "user_id": "uuid",
  "plan": {
    "id": "plan_basic",
    "name": "Basic",
    "daily_scans": 20,
    "features": { ... }
  },
  "status": "active",
  "current_period_start": "2026-05-01T00:00:00Z",
  "current_period_end": "2026-06-01T00:00:00Z",
  "cancel_at_period_end": false,
  "scans_used_today": 3,
  "scans_remaining_today": 17,
  "stripe_subscription_id": "sub_1ABC2Def3Ghi4Jkl"
}
```
---

### 4.5 Cancel Subscription

```
POST /subscriptions/cancel
```

**Headers:** `Authorization: Bearer <access_token>`

**Response `200 OK`:**
```json
{
  "message": "Subscription will be cancelled at the end of the current billing period.",
  "cancel_at_period_end": true,
  "current_period_end": "2026-06-01T00:00:00Z"
}
```

---

## 5. Database Schema

### 5.1 Entity-Relationship Diagram (Textual)

```
users ──1:1── user_subscriptions ──N:1── subscription_plans
users ──1:N── scan_records
users ──1:N── scan_daily_usage
users ──1:N── password_reset_tokens
users ──1:N── payment_transactions
scan_records ──N:1── users (nullable)
ip_scan_tracking (standalone, keyed by IP)
```

### 5.2 Table Definitions

```sql
-- ============================================================
-- USERS
-- ============================================================
CREATE EXTENSION IF NOT EXISTS "uuid-ossp";
CREATE TABLE users (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    email           VARCHAR(255) UNIQUE NOT NULL,
    password_hash   VARCHAR(255),
    full_name       VARCHAR(100) NOT NULL,
    profile_photo_url TEXT,
    medical_history  JSONB DEFAULT '{}',
    auth_provider   VARCHAR(20) DEFAULT 'email'
                    CHECK (auth_provider IN ('email', 'google', 'facebook')),
    social_id       VARCHAR(255),
    is_verified     BOOLEAN DEFAULT false,
    is_active       BOOLEAN DEFAULT true,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
    updated_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_users_email ON users(email);
CREATE INDEX idx_users_social ON users(auth_provider, social_id);
-- ============================================================
-- SUBSCRIPTION PLANS (seed data)
-- ============================================================
CREATE TABLE subscription_plans (
    id              VARCHAR(50) PRIMARY KEY,  -- 'plan_free', 'plan_basic', 'plan_premium'
    name            VARCHAR(100) NOT NULL,
    price_cents     INTEGER NOT NULL,
    currency        VARCHAR(3) DEFAULT 'USD',
    billing_interval VARCHAR(20) DEFAULT 'month',
    daily_scan_limit INTEGER NOT NULL,
    features        JSONB NOT NULL DEFAULT '{}',
    stripe_price_id VARCHAR(255),
    is_active       BOOLEAN DEFAULT true,
    sort_order      INTEGER DEFAULT 0
);
-- Seed data
INSERT INTO subscription_plans (id, name, price_cents, daily_scan_limit, features, sort_order) VALUES
('plan_free',    'Free',    0,   5,  '{"scan_history_days": 30, "detailed_results": true, "product_recommendations": true, "dermatologist_referrals": false, "priority_support": false, "data_export": false}', 1),
('plan_basic',   'Basic',   499, 20, '{"scan_history_days": 365, "detailed_results": true, "product_recommendations": true, "dermatologist_referrals": true, "priority_support": false, "data_export": false}', 2),
('plan_premium', 'Premium', 999, 100,'{"scan_history_days": null, "detailed_results": true, "product_recommendations": true, "dermatologist_referrals": true, "priority_support": true, "data_export": true}', 3);
-- ============================================================
-- USER SUBSCRIPTIONS
-- ============================================================
CREATE TABLE user_subscriptions (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    user_id         UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
    plan_id         VARCHAR(50) NOT NULL REFERENCES subscription_plans(id),
    status          VARCHAR(20) NOT NULL DEFAULT 'active'
                    CHECK (status IN ('active', 'past_due', 'canceled', 'expired')),
    current_period_start TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
    current_period_end   TIMESTAMP WITH TIME ZONE,
    cancel_at_period_end BOOLEAN DEFAULT false,
    stripe_subscription_id VARCHAR(255),
    stripe_customer_id    VARCHAR(255),
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
    updated_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
    UNIQUE(user_id)  -- one active subscription per user
);
CREATE INDEX idx_user_sub_status ON user_subscriptions(user_id, status);
-- ============================================================
-- SCAN RECORDS
-- ============================================================
CREATE TABLE scan_records (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    user_id         UUID REFERENCES users(id) ON DELETE SET NULL,
    ip_address      VARCHAR(45) NOT NULL,
    image_url       TEXT NOT NULL,
    thumbnail_url   TEXT,
    original_filename VARCHAR(255),
    image_size_bytes INTEGER,
    processing_time_ms INTEGER,
    results         JSONB NOT NULL,
    primary_detection JSONB,
    detected_conditions TEXT[],
    model_version   VARCHAR(50),
    is_flagged      BOOLEAN DEFAULT false,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_scan_user ON scan_records(user_id, created_at DESC);
CREATE INDEX idx_scan_ip   ON scan_records(ip_address, created_at DESC);
CREATE INDEX idx_scan_conditions ON scan_records USING GIN(detected_conditions);
-- ============================================================
-- SCAN DAILY USAGE (per-user quota tracking)
-- ============================================================
CREATE TABLE scan_daily_usage (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    user_id         UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
    scan_date       DATE NOT NULL DEFAULT CURRENT_DATE,
    scan_count      INTEGER NOT NULL DEFAULT 0,
    PRIMARY KEY (user_id, scan_date)
);
-- ============================================================
-- IP SCAN TRACKING (anonymous rate limiting)
-- ============================================================
CREATE TABLE ip_scan_tracking (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    ip_address      VARCHAR(45) UNIQUE NOT NULL,
    scan_count      INTEGER NOT NULL DEFAULT 0,
    first_scan_at   TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
    last_scan_at    TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
-- ============================================================
-- PASSWORD RESET TOKENS
-- ============================================================
CREATE TABLE password_reset_tokens (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    user_id         UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
    token_hash      VARCHAR(255) NOT NULL,
    expires_at      TIMESTAMP WITH TIME ZONE NOT NULL,
    is_used         BOOLEAN DEFAULT false,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_reset_token ON password_reset_tokens(user_id, is_used);
-- ============================================================
-- PAYMENT TRANSACTIONS
-- ============================================================
CREATE TABLE payment_transactions (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    user_id         UUID NOT NULL REFERENCES users(id) ON DELETE CASCADE,
    amount_cents    INTEGER NOT NULL,
    currency        VARCHAR(3) DEFAULT 'USD',
    payment_provider VARCHAR(50) DEFAULT 'stripe',
    payment_provider_id VARCHAR(255),
    subscription_id UUID REFERENCES user_subscriptions(id),
    status          VARCHAR(20) NOT NULL
                    CHECK (status IN ('pending', 'completed', 'failed', 'refunded')),
    metadata        JSONB DEFAULT '{}',
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_payment_user ON payment_transactions(user_id, created_at DESC);
-- ============================================================
-- DOCTORS / CLINICS / PRODUCTS (for recommendations)
-- ============================================================
CREATE TABLE doctors (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    name            VARCHAR(200) NOT NULL,
    specializations TEXT[],
    address         TEXT,
    phone           VARCHAR(50),
    email           VARCHAR(255),
    website         TEXT,
    rating          DECIMAL(2,1) DEFAULT 0.0,
    review_count    INTEGER DEFAULT 0,
    coordinates     POINT,
    insurance_accepted TEXT[],
    is_active       BOOLEAN DEFAULT true,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_doctor_location ON doctors USING GIST(coordinates);
CREATE TABLE clinics (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    name            VARCHAR(200) NOT NULL,
    type            VARCHAR(50) DEFAULT 'clinic',
    address         TEXT,
    phone           VARCHAR(50),
    rating          DECIMAL(2,1) DEFAULT 0.0,
    review_count    INTEGER DEFAULT 0,
    services        TEXT[],
    open_hours      JSONB,
    accepts_insurance BOOLEAN,
    walkins_accepted BOOLEAN,
    coordinates     POINT,
    is_active       BOOLEAN DEFAULT true,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_clinic_location ON clinics USING GIST(coordinates);
CREATE TABLE products (
    id              UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
    name            VARCHAR(200) NOT NULL,
    brand           VARCHAR(100),
    category        VARCHAR(50),
    type            VARCHAR(50) CHECK (type IN ('over_the_counter', 'prescription')),
    description     TEXT,
    key_ingredients TEXT[],
    recommended_for TEXT[],
    average_rating  DECIMAL(2,1),
    price_amount    DECIMAL(10,2),
    price_currency  VARCHAR(3) DEFAULT 'USD',
    retailer        VARCHAR(100),
    affiliate_link  TEXT,
    image_url       TEXT,
    prescription_required BOOLEAN DEFAULT false,
    is_active       BOOLEAN DEFAULT true,
    created_at      TIMESTAMP WITH TIME ZONE DEFAULT NOW()
);
CREATE INDEX idx_product_condition ON products USING GIN(recommended_for);
```

---

## 6. Rate Limiting Logic

### 6.1 Architecture Overview

```
                    ┌─────────────────┐
                    │   Client Request │
                    └────────┬────────┘
                             │
                    ┌────────▼────────┐
                    │  Rate Limit     │
                    │  Middleware      │
                    └────────┬────────┘
                             │
              ┌──────────────┼──────────────┐
              │                             │
    ┌─────────▼─────────┐       ┌──────────▼──────────┐
    │ Has valid JWT?     │       │ No JWT / Invalid    │
    │ => User-based      │       │ => IP-based         │
    │    rate limiting   │       │    rate limiting     │
    └─────────┬──────────┘       └──────────┬──────────┘
              │                             │
    ┌─────────▼─────────┐       ┌──────────▼──────────┐
    │ Check scan_daily   │       │ Check ip_scan_track │
    │ _usage table      │       │ _ing table          │
    │ WHERE user_id=X   │       │ WHERE ip=X          │
    │   AND date=TODAY  │       │                     │
    └─────────┬──────────┘       └──────────┬──────────┘
              │                             │
              │      ┌──────────────────────┘
              │      │
          ┌───▼──────▼───┐
          │  Under limit? │───No──→ 429 Too Many Requests
          └───┬──────┬───┘
              │Yes   │Yes
              │      │
    ┌─────────▼┐ ┌──▼────────┐
    │ Increment│ │ Proceed   │
    │ counter  │ │ to scan   │
    └──────────┘ └───────────┘
```

### 6.2 Rate Limit Rules

| Scope | Limit | Window | Storage | Reset |
|-------|-------|--------|---------|-------|
| Anonymous (per IP) | 3 scans total | Lifetime (per IP) | `ip_scan_tracking` table | Never (manual reset) |
| Free tier (per user) | 5 scans | Daily (UTC) | `scan_daily_usage` table | Midnight UTC |
| Basic plan (per user) | 20 scans | Daily (UTC) | `scan_daily_usage` table | Midnight UTC |
| Premium plan (per user) | 100 scans | Daily (UTC) | `scan_daily_usage` table | Midnight UTC |

### 6.3 Quota Check Algorithm (Pseudocode)

```python
async def check_scan_quota(request: Request, user: User | None) -> QuotaResult:
    if user:
        # ── Authenticated user ──────────────────────────────
        plan = await get_user_plan(user.id)  # default: free plan
        daily_limit = plan.daily_scan_limit
        today = date.today()
        usage = await ScanDailyUsage.objects.get_or_create(
            user_id=user.id, scan_date=today,
            defaults={"scan_count": 0}
        )
        if usage.scan_count >= daily_limit:
            raise HTTPException(
                status_code=429,
                detail={
                    "message": "Daily scan limit reached.",
                    "limit": daily_limit,
                    "plan": plan.name,
                    "reset_at": f"{today + timedelta(days=1)}T00:00:00Z",
                    "upgrade_url": "/subscriptions/plans"
                }
            )
        usage.scan_count += 1
        await usage.save()
        return QuotaResult(remaining=daily_limit - usage.scan_count)
    else:
        # ── Anonymous user ───────────────────────────────────
        ip = get_client_ip(request)
        ANON_LIMIT = 3
        tracking = await ip_scan_tracking.objects.get_or_create(
            ip_address=ip,
            defaults={"scan_count": 0, "first_scan_at": now()}
        )
        if tracking.scan_count >= ANON_LIMIT:
            raise HTTPException(
                status_code=429,
                detail={
                    "message": "Anonymous scan limit reached (3 scans). Sign up to continue.",
                    "limit": ANON_LIMIT,
                    "reset_at": None  # never resets for anonymous
                }
            )
        tracking.scan_count += 1
        tracking.last_scan_at = now()
        await tracking.save()
        return QuotaResult(remaining=ANON_LIMIT - tracking.scan_count)
```

### 6.4 IP Address Detection

```python
def get_client_ip(request: Request) -> str:
    # Priority: X-Forwarded-For > X-Real-IP > direct remote
    forwarded = request.headers.get("X-Forwarded-For")
    if forwarded:
        return forwarded.split(",")[0].strip()
    real_ip = request.headers.get("X-Real-IP")
    if real_ip:
        return real_ip
    return request.client.host
```

### 6.5 IP Geolocation (for Recommendations)

```python
# Use MaxMind GeoLite2 database
from geoip2.database import Reader
geo_reader = Reader("data/GeoLite2-City.mmdb")
async def get_location_from_ip(ip: str) -> GeoLocation:
    if ip in ("127.0.0.1", "::1", "localhost"):
        return GeoLocation(lat=40.7128, lng=-74.0060)  # default fallback
    response = geo_reader.city(ip)
    return GeoLocation(
        lat=response.location.latitude,
        lng=response.location.longitude,
        city=response.city.name,
        country=response.country.name
    )
```

---

## 7. Error Handling

### 7.1 Global Error Schema

All error responses follow this structure:

```json
{
  "detail": "Human-readable error message.",
  "code": "ERROR_CODE",
  "timestamp": "2026-05-17T10:30:00Z",
  "path": "/v1/scans",
  "extra": {}
}
```

### 7.2 Standard Error Codes

| HTTP Code | Code String | When |
|-----------|-------------|------|
| 400 | `VALIDATION_ERROR` | Invalid input (Pydantic validation) |
| 400 | `INVALID_IMAGE` | Bad image format/dimensions/size |
| 401 | `UNAUTHORIZED` | Missing/invalid/expired JWT |
| 403 | `FORBIDDEN` | Valid JWT but insufficient permissions |
| 404 | `NOT_FOUND` | Resource doesn't exist |
| 409 | `CONFLICT` | Duplicate email, etc. |
| 422 | `UNPROCESSABLE_ENTITY` | Request body fails schema validation |
| 429 | `RATE_LIMIT_EXCEEDED` | Quota exceeded |
| 500 | `INTERNAL_ERROR` | Unexpected server error |
| 502 | `AI_MODEL_ERROR` | Model inference failure |
| 503 | `SERVICE_UNAVAILABLE` | Model not loaded / maintenance |

### 7.3 FastAPI Exception Handler Setup

```python
@app.exception_handler(HTTPException)
async def http_exception_handler(request, exc):
    return JSONResponse(
        status_code=exc.status_code,
        content={
            "detail": exc.detail,
            "code": ERROR_CODE_MAP.get(exc.status_code, "UNKNOWN"),
            "timestamp": datetime.utcnow().isoformat() + "Z",
            "path": request.url.path,
        }
    )
@app.exception_handler(RequestValidationError)
async def validation_exception_handler(request, exc):
    return JSONResponse(
        status_code=422,
        content={
            "detail": exc.errors(),
            "code": "VALIDATION_ERROR",
            "timestamp": datetime.utcnow().isoformat() + "Z",
            "path": request.url.path,
        }
    )
```

---

## 8. Middleware & Configuration

### 8.1 CORS Configuration

```python
origins = [
    "https://skindetect.com",
    "https://app.skindetect.com",
    "https://admin.skindetect.com",
    "http://localhost:3000",     # dev
    "http://localhost:5173",     # vite dev
]
app.add_middleware(
    CORSMiddleware,
    allow_origins=origins,
    allow_credentials=True,
    allow_methods=["GET", "POST", "PUT", "DELETE", "OPTIONS"],
    allow_headers=["Authorization", "Content-Type", "X-Request-ID"],
    expose_headers=["X-RateLimit-Remaining", "X-RateLimit-Reset"],
    max_age=600,
)
```

### 8.2 Audit Logging Middleware

```python
@app.middleware("http")
async def audit_log_middleware(request: Request, call_next):
    start_time = time.perf_counter()
    request_id = str(uuid4())
    request.state.request_id = request_id
    response = await call_next(request)
    elapsed = (time.perf_counter() - start_time) * 1000
    logger.info(
        "audit_log",
        extra={
            "request_id": request_id,
            "method": request.method,
            "path": request.url.path,
            "status_code": response.status_code,
            "elapsed_ms": round(elapsed, 2),
            "ip": get_client_ip(request),
            "user_agent": request.headers.get("user-agent"),
            "user_id": getattr(request.state, "user_id", None),
        }
    )
    return response
```

### 8.3 Security Headers Middleware

```python
@app.middleware("http")
async def security_headers(request: Request, call_next):
    response = await call_next(request)
    response.headers["X-Content-Type-Options"] = "nosniff"
    response.headers["X-Frame-Options"] = "DENY"
    response.headers["X-XSS-Protection"] = "1; mode=block"
    response.headers["Strict-Transport-Security"] = "max-age=31536000; includeSubDomains"
    response.headers["Cache-Control"] = "no-store"
    return response
```

### 8.4 Rate Limit Response Headers

Every scan response includes these headers:

```
X-RateLimit-Limit: 5          # Total daily limit
X-RateLimit-Remaining: 3      # Remaining scans today
X-RateLimit-Reset: 1715904000 # Unix timestamp of reset (midnight UTC)
```

---

## 9. Complete API Route Table

| Method | Route | Auth | Rate Limited | Description |
|--------|-------|------|-------------|-------------|
| POST | `/auth/signup` | No | No | Create account |
| POST | `/auth/login` | No | 10/min per IP | Login |
| POST | `/auth/social/{provider}` | No | 10/min per IP | Social login |
| POST | `/auth/forgot-password` | No | 3/hr per email | Request reset |
| POST | `/auth/reset-password` | No | 5/hr per IP | Complete reset |
| PUT | `/auth/change-password` | Yes | No | Change password |
| POST | `/auth/refresh` | No | No | Refresh JWT |
| PUT | `/auth/profile` | Yes | No | Update profile |
| POST | `/scans` | Optional | Yes (IP/user) | Scan image |
| GET | `/scans` | Yes | 60/min | Scan history |
| GET | `/scans/{id}` | Yes | 60/min | Scan detail |
| PATCH | `/scans/{id}` | Yes | 60/min | Rename scan |
| DELETE | `/scans/{id}` | Yes | 60/min | Delete scan |
| GET | `/recommendations/dermatologists` | Optional | 30/min | Find doctors |
| GET | `/recommendations/products` | Optional | 30/min | Find products |
| GET | `/recommendations/clinics` | Optional | 30/min | Find clinics |
| GET | `/subscriptions/plans` | No | No | List plans |
| POST | `/subscriptions/create-checkout-session` | Yes | 5/min | Start checkout |
| POST | `/subscriptions/webhook` | No (sig) | No | Stripe webhook |
| GET | `/subscriptions/my` | Yes | 30/min | My subscription |
| POST | `/subscriptions/cancel` | Yes | 5/min | Cancel subscription |
| GET | `/health` | No | No | Health check |

---

## 10. Environment Variables

```bash
# ── App ──────────────────────────────────────────────
APP_ENV=production
APP_SECRET_KEY=<random-64-char-hex>
APP_DEBUG=false
API_PREFIX=/v1
# ── Database ──────────────────────────────────────────
DATABASE_URL=postgresql+asyncpg://user:pass@localhost:5432/skindetect
# ── JWT ──────────────────────────────────────────────
JWT_ACCESS_SECRET=<random-64-char-hex>
JWT_REFRESH_SECRET=<different-random-64-char-hex>
JWT_ACCESS_EXPIRE_MINUTES=60
JWT_REFRESH_EXPIRE_DAYS=30
JWT_ALGORITHM=HS256
# ── AI Model ─────────────────────────────────────────
MODEL_PATH=best_skin_model_3.keras
CLASS_NAMES_PATH=class_names.json
MODEL_INPUT_SIZE=224
CONFIDENCE_THRESHOLD=15.0
# ── Storage (CDN) ────────────────────────────────────
CDN_PROVIDER=s3
AWS_ACCESS_KEY_ID=...
AWS_SECRET_ACCESS_KEY=...
AWS_S3_BUCKET=skindetect-scans
AWS_REGION=us-east-1
CDN_BASE_URL=https://cdn.skindetect.com
# ── OAuth ────────────────────────────────────────────
GOOGLE_CLIENT_ID=...
GOOGLE_CLIENT_SECRET=...
FACEBOOK_APP_ID=...
FACEBOOK_APP_SECRET=...
# ── Stripe ───────────────────────────────────────────
STRIPE_SECRET_KEY=sk_live_...
STRIPE_WEBHOOK_SECRET=whsec_...
STRIPE_PRICE_FREE=price_1...
STRIPE_PRICE_BASIC=price_2...
STRIPE_PRICE_PREMIUM=price_3...
# ── Rate Limiting ────────────────────────────────────
ANONYMOUS_SCAN_LIMIT=3
FREE_TIER_DAILY_SCANS=5
BASIC_PLAN_DAILY_SCANS=20
PREMIUM_PLAN_DAILY_SCANS=100
# ── Email ────────────────────────────────────────────
SMTP_HOST=smtp.sendgrid.net
SMTP_PORT=587
SMTP_USER=apikey
SMTP_PASSWORD=...
FROM_EMAIL=noreply@skindetect.com
# ── GeoIP ────────────────────────────────────────────
GEOIP_DB_PATH=data/GeoLite2-City.mmdb
# ── Affiliate ────────────────────────────────────────
AMAZON_AFFILIATE_TAG=skindetect-20
WALMART_AFFILIATE_TAG=skindetect-20
CVS_AFFILIATE_TAG=skindetect-20
```

---

## 11. Pydantic Model Definitions

```python
# ── Schemas ──────────────────────────────────────────────
class SignupRequest(BaseModel):
    email: EmailStr
    password: str = Field(..., min_length=8, max_length=128)
    full_name: str = Field(..., min_length=1, max_length=100, pattern=r"^[a-zA-Z\s\-]+$")
    medical_history: Optional[dict] = None
class LoginRequest(BaseModel):
    email: EmailStr
    password: str
class SocialLoginRequest(BaseModel):
    access_token: str
class ForgotPasswordRequest(BaseModel):
    email: EmailStr
class ResetPasswordRequest(BaseModel):
    token: str
    new_password: str = Field(..., min_length=8, max_length=128)
class ChangePasswordRequest(BaseModel):
    current_password: str
    new_password: str = Field(..., min_length=8, max_length=128)
class RefreshTokenRequest(BaseModel):
    refresh_token: str
class UpdateProfileRequest(BaseModel):
    full_name: Optional[str] = Field(None, min_length=1, max_length=100)
    profile_photo: Optional[str] = None  # base64
    medical_history: Optional[dict] = None
class ScanResponse(BaseModel):
    scan_id: str
    image_url: str
    results: list[DiseaseResult]
    primary_detection: PrimaryDetection
    processing_time_ms: int
    disclaimer: str
    quota_remaining: int
class DiseaseResult(BaseModel):
    disease: str
    scientific_name: Optional[str] = None
    confidence: float  # 0–100
    severity: Optional[str] = None
    description: Optional[str] = None
    common_symptoms: list[str] = []
    urgency_level: str  # low | consult_recommended | urgent | emergency
class PrimaryDetection(BaseModel):
    disease: str
    confidence: float
    severity: Optional[str] = None
    urgency_level: str
class PaginatedResponse(BaseModel, Generic[T]):
    items: list[T]
    total: int
    page: int
    page_size: int
    total_pages: int
class TokenResponse(BaseModel):
    access_token: str
    refresh_token: str
    token_type: str = "bearer"
    expires_in: int
    user: UserSummary
class UserSummary(BaseModel):
    id: str
    email: str
    full_name: str
    subscription_plan: str
    scans_remaining_today: int
class ErrorResponse(BaseModel):
    detail: str
    code: str
    timestamp: str
    path: str
    extra: dict = {}
```

---

## 12. File Structure (Recommended)

```
app/
├── main.py                        # FastAPI app creation, middleware, exception handlers
├── config.py                      # Settings from env vars (pydantic-settings)
├── database.py                    # SQLAlchemy async engine + session
├── models/                        # SQLAlchemy ORM models
│   ├── __init__.py
│   ├── user.py
│   ├── scan.py
│   ├── subscription.py
│   └── recommendation.py
├── schemas/                       # Pydantic request/response schemas
│   ├── __init__.py
│   ├── auth.py
│   ├── user.py
│   ├── scan.py
│   ├── subscription.py
│   └── recommendation.py
├── api/
│   ├── __init__.py
│   ├── deps.py                    # Dependency injection (get_db, get_current_user, etc.)
│   └── v1/
│       ├── __init__.py
│       ├── auth.py                # /auth/* routes
│       ├── users.py               # /auth/profile route
│       ├── scans.py               # /scans/* routes
│       ├── recommendations.py     # /recommendations/* routes
│       ├── subscriptions.py       # /subscriptions/* routes
│       └── payments.py            # payment webhook
├── core/
│   ├── __init__.py
│   ├── security.py                # JWT encode/decode, password hashing
│   ├── rate_limiter.py            # Quota check logic
│   ├── ip_geo.py                  # IP geolocation
│   └── logging.py                 # Logger setup
├── services/
│   ├── __init__.py
│   ├── ai_detector.py             # Model loading + inference
│   ├── payment_gateway.py         # Stripe integration
│   ├── email_service.py           # SendGrid/SMTP
│   └── recommendation_engine.py   # Doctor/product/clinic lookup
├── middleware/
│   ├── __init__.py
│   └── rate_limit.py              # Rate limit middleware
└── tests/
    ├── __init__.py
    ├── conftest.py
    ├── test_auth.py
    ├── test_scans.py
    ├── test_recommendations.py
    └── test_subscriptions.py
```

---

## 13. Security Checklist

- [x] Passwords hashed with bcrypt (≥12 rounds)
- [x] JWT signed with separate secrets for access/refresh tokens
- [x] Refresh tokens are single-use (rotate on each refresh)
- [x] Rate limiting on auth endpoints (3/hr forgot password, 10/min login)
- [x] CORS restricted to specific origins
- [x] Security headers via middleware
- [x] Stripe webhook verified via signature, not auth token
- [x] Uploaded images scanned for malware (ClamAV)
- [x] Image EXIF data stripped before storage
- [x] Input validation on all endpoints (Pydantic)
- [x] SQL injection prevention (parameterized queries via SQLAlchemy)
- [x] No secrets in code (all via env vars)
- [x] Rate limit headers exposed to clients
- [x] Password reset tokens are one-time-use with short expiry (15 min)
- [x] Silent fail on forgot-password email lookup (don't reveal if email exists)

---

## 14. Performance Considerations

| Concern | Strategy |
|---------|----------|
| Model inference latency | Load model at startup, warm up with dummy input, use async executor |
| Image upload | Upload to CDN (S3/GCS) directly via presigned URL, process async |
| Scan history queries | Composite indexes on (user_id, created_at DESC), GIN on detected_conditions |
| Rate limit counters | Use Redis for production; fallback to DB for simplicity |
| Geo-recommendations | PostGIS for spatial queries, cache results (5-min TTL) |
| Database connections | Async SQLAlchemy with connection pooling (min 5, max 20) |
| CDN | Serve scan thumbnails (256×256 WebP) from CDN with 1-year cache |
| Monitoring | Prometheus metrics: scan latency histogram, error rate, quota exhaustion rate |
---
> **Document Version:** 1.0.0  
> **Last Updated:** 2026-05-17  
> **Author:** FastAPI Backend Architecture Team