multitenetsaas/backend/src/core/middleware/tenant_middleware.py

"""
Multi-tenant middleware for Django applications.

Provides tenant isolation and context management for all requests.
Supports multiple tenant identification methods and security features.
"""

import logging
from django.db import connection
from django.http import Http404, HttpResponseForbidden
from django.conf import settings
from django.utils.deprecation import MiddlewareMixin
from django.contrib.auth.middleware import get_user
from django.core.cache import cache
from django.core.exceptions import PermissionDenied
from django.urls import resolve
from django.utils import timezone
from django.db import transaction
import re
import uuid

from ..models.tenant import Tenant
from ..models.user import User

logger = logging.getLogger(__name__)


class TenantMiddleware(MiddlewareMixin):
    """
    Middleware to identify and isolate tenants for each request.
    """

    def __init__(self, get_response):
        self.get_response = get_response
        self.tenant_model = Tenant
        self.user_model = User

        # Configure tenant identification methods
        self.host_based_domains = getattr(settings, 'TENANT_HOST_BASED_DOMAINS', True)
        self.path_based_prefix = getattr(settings, 'TENANT_PATH_BASED_PREFIX', '/tenant/')
        self.header_based = getattr(settings, 'TENANT_HEADER_BASED', True)
        self.tenant_header = getattr(settings, 'TENANT_HEADER_NAME', 'X-Tenant-ID')

        # Security settings
        self.enforce_tenant_isolation = getattr(settings, 'ENFORCE_TENANT_ISOLATION', True)
        self.public_paths = getattr(settings, 'PUBLIC_PATHS', [
            '/api/v1/auth/',
            '/api/v1/public/',
            '/health/',
            '/metrics/',
            '/api/v1/tenants/register',
            '/admin/login/',
            '/static/',
            '/media/',
        ])

        # Cache settings
        self.cache_enabled = getattr(settings, 'TENANT_CACHE_ENABLED', True)
        self.cache_timeout = getattr(settings, 'TENANT_CACHE_TIMEOUT', 300)  # 5 minutes

        # Admin paths that bypass tenant isolation
        self.admin_paths = getattr(settings, 'ADMIN_PATHS', [
            '/admin/',
            '/api/v1/admin/',
        ])

        # API prefix that requires tenant context
        self.api_prefix = getattr(settings, 'API_PREFIX', '/api/v1/')

        # Initialize regex patterns
        self.public_path_patterns = [re.compile(path) for path in self.public_paths]
        self.admin_path_patterns = [re.compile(path) for path in self.admin_paths]

    def process_request(self, request):
        """
        Process incoming request to identify tenant.
        """
        request.tenant = None
        request.tenant_id = None
        request.tenant_context = {}
        request.is_tenant_request = False

        # Skip tenant processing for public paths
        if self._is_public_path(request.path):
            logger.debug(f"Skipping tenant processing for public path: {request.path}")
            return None

        # Skip tenant processing for admin paths (handled by authentication)
        if self._is_admin_path(request.path):
            logger.debug(f"Skipping tenant processing for admin path: {request.path}")
            return None

        # Try to identify tenant using various methods
        tenant = self._identify_tenant(request)

        if not tenant:
            # For API paths that require tenant context, return 404
            if request.path.startswith(self.api_prefix):
                logger.warning(f"Tenant not found for API request: {request.path}")
                raise Http404("Tenant not found")

            # For other paths, allow but without tenant context
            return None

        # Set tenant context
        request.tenant = tenant
        request.tenant_id = tenant.id
        request.tenant_context = self._build_tenant_context(tenant)
        request.is_tenant_request = True

        # Set database connection tenant context
        self._set_database_tenant_context(tenant)

        # Set cache tenant context
        if self.cache_enabled:
            self._set_cache_tenant_context(tenant)

        # Validate tenant status
        if not self._validate_tenant_status(tenant):
            logger.warning(f"Tenant {tenant.id} is not active: {tenant.status}")
            return HttpResponseForbidden("Tenant account is not active")

        # Validate user access to tenant if authenticated
        user = get_user(request)
        if user.is_authenticated and not user.is_superuser:
            if not self._validate_user_tenant_access(user, tenant):
                logger.warning(f"User {user.id} attempted to access tenant {tenant.id}")
                return HttpResponseForbidden("Access denied to this tenant")

        logger.debug(f"Tenant context set: {tenant.name} ({tenant.id})")
        return None

    def process_response(self, request, response):
        """
        Clean up tenant context after request processing.
        """
        # Clear database tenant context
        self._clear_database_tenant_context()

        # Clear cache tenant context
        if self.cache_enabled:
            self._clear_cache_tenant_context()

        return response

    def process_exception(self, request, exception):
        """
        Handle exceptions and clean up tenant context.
        """
        # Clear database tenant context
        self._clear_database_tenant_context()

        # Clear cache tenant context
        if self.cache_enabled:
            self._clear_cache_tenant_context()

        return None

    def _is_public_path(self, path):
        """
        Check if path is public and doesn't require tenant context.
        """
        return any(pattern.match(path) for pattern in self.public_path_patterns)

    def _is_admin_path(self, path):
        """
        Check if path is admin path.
        """
        return any(pattern.match(path) for pattern in self.admin_path_patterns)

    def _identify_tenant(self, request):
        """
        Identify tenant using various methods.
        """
        tenant = None

        # Method 1: Host-based tenant identification
        if self.host_based_domains and not tenant:
            tenant = self._identify_tenant_by_host(request)

        # Method 2: Path-based tenant identification
        if not tenant:
            tenant = self._identify_tenant_by_path(request)

        # Method 3: Header-based tenant identification
        if self.header_based and not tenant:
            tenant = self._identify_tenant_by_header(request)

        # Method 4: Subdomain-based tenant identification
        if not tenant:
            tenant = self._identify_tenant_by_subdomain(request)

        # Method 5: User-based tenant identification (if authenticated)
        if not tenant:
            tenant = self._identify_tenant_by_user(request)

        return tenant

    def _identify_tenant_by_host(self, request):
        """
        Identify tenant by hostname.
        """
        host = request.get_host().split(':')[0]  # Remove port

        # Try to find tenant by custom domain
        tenant = self._get_tenant_from_cache(f'tenant:host:{host}')
        if not tenant:
            try:
                tenant = self.tenant_model.objects.filter(
                    domain_mappings__contains=[host],
                    status__in=['ACTIVE', 'PENDING']
                ).first()

                if tenant:
                    self._set_tenant_cache(f'tenant:host:{host}', tenant)
            except Exception as e:
                logger.error(f"Error identifying tenant by host {host}: {e}")

        return tenant

    def _identify_tenant_by_path(self, request):
        """
        Identify tenant by URL path.
        """
        if request.path.startswith(self.path_based_prefix):
            tenant_slug = request.path[len(self.path_based_prefix):].split('/')[0]
            if tenant_slug:
                tenant = self._get_tenant_from_cache(f'tenant:slug:{tenant_slug}')
                if not tenant:
                    try:
                        tenant = self.tenant_model.objects.filter(
                            slug=tenant_slug,
                            status__in=['ACTIVE', 'PENDING']
                        ).first()

                        if tenant:
                            self._set_tenant_cache(f'tenant:slug:{tenant_slug}', tenant)
                    except Exception as e:
                        logger.error(f"Error identifying tenant by slug {tenant_slug}: {e}")

                return tenant

        return None

    def _identify_tenant_by_header(self, request):
        """
        Identify tenant by request header.
        """
        tenant_id = request.headers.get(self.tenant_header) or request.headers.get('X-Tenant-ID')

        if tenant_id:
            # Try to parse as UUID
            try:
                tenant_uuid = uuid.UUID(tenant_id)
                tenant = self._get_tenant_from_cache(f'tenant:id:{tenant_uuid}')
                if not tenant:
                    try:
                        tenant = self.tenant_model.objects.filter(
                            id=tenant_uuid,
                            status__in=['ACTIVE', 'PENDING']
                        ).first()

                        if tenant:
                            self._set_tenant_cache(f'tenant:id:{tenant_uuid}', tenant)
                    except Exception as e:
                        logger.error(f"Error identifying tenant by ID {tenant_id}: {e}")

                return tenant
            except ValueError:
                # Not a UUID, try as slug
                tenant = self._get_tenant_from_cache(f'tenant:slug:{tenant_id}')
                if not tenant:
                    try:
                        tenant = self.tenant_model.objects.filter(
                            slug=tenant_id,
                            status__in=['ACTIVE', 'PENDING']
                        ).first()

                        if tenant:
                            self._set_tenant_cache(f'tenant:slug:{tenant_id}', tenant)
                    except Exception as e:
                        logger.error(f"Error identifying tenant by header {tenant_id}: {e}")

                return tenant

        return None

    def _identify_tenant_by_subdomain(self, request):
        """
        Identify tenant by subdomain.
        """
        host = request.get_host().split(':')[0]

        # Extract subdomain
        parts = host.split('.')
        if len(parts) > 2:
            subdomain = parts[0]

            tenant = self._get_tenant_from_cache(f'tenant:subdomain:{subdomain}')
            if not tenant:
                try:
                    tenant = self.tenant_model.objects.filter(
                        slug=subdomain,
                        status__in=['ACTIVE', 'PENDING']
                    ).first()

                    if tenant:
                        self._set_tenant_cache(f'tenant:subdomain:{subdomain}', tenant)
                except Exception as e:
                    logger.error(f"Error identifying tenant by subdomain {subdomain}: {e}")

            return tenant

        return None

    def _identify_tenant_by_user(self, request):
        """
        Identify tenant by authenticated user.
        """
        user = get_user(request)
        if user.is_authenticated and hasattr(user, 'tenant_id'):
            tenant = self._get_tenant_from_cache(f'tenant:id:{user.tenant_id}')
            if not tenant:
                try:
                    tenant = self.tenant_model.objects.filter(
                        id=user.tenant_id,
                        status__in=['ACTIVE', 'PENDING']
                    ).first()

                    if tenant:
                        self._set_tenant_cache(f'tenant:id:{user.tenant_id}', tenant)
                except Exception as e:
                    logger.error(f"Error identifying tenant by user {user.id}: {e}")

            return tenant

        return None

    def _build_tenant_context(self, tenant):
        """
        Build tenant context dictionary.
        """
        return {
            'tenant_id': tenant.id,
            'tenant_name': tenant.name,
            'tenant_slug': tenant.slug,
            'tenant_status': tenant.status,
            'tenant_plan': tenant.subscription_plan,
            'tenant_industry': tenant.business_type,
            'tenant_timezone': tenant.timezone,
            'tenant_currency': tenant.currency,
            'tenant_locale': tenant.locale,
            'tenant_features': tenant.get_features(),
            'tenant_limits': {
                'users': tenant.get_user_limits(),
                'modules': tenant.get_module_limits(),
            },
            'is_trial': tenant.is_on_trial,
            'trial_ends_at': tenant.trial_ends_at.isoformat() if tenant.trial_ends_at else None,
            'subscription_active': tenant.subscription_active,
            'subscription_ends_at': tenant.subscription_ends_at.isoformat() if tenant.subscription_ends_at else None,
        }

    def _validate_tenant_status(self, tenant):
        """
        Validate that tenant is in acceptable state.
        """
        if tenant.status == 'SUSPENDED':
            return False
        elif tenant.status == 'TERMINATED':
            return False
        elif tenant.status == 'PENDING' and not tenant.is_on_trial:
            return False

        return True

    def _validate_user_tenant_access(self, user, tenant):
        """
        Validate that user has access to the specified tenant.
        """
        # Superusers have access to all tenants
        if user.is_superuser:
            return True

        # Check if user belongs to tenant
        if user.tenant_id == tenant.id:
            return True

        # Check if user is staff with admin access
        if user.is_staff and user.role in ['ADMIN', 'MANAGER']:
            return True

        return False

    def _set_database_tenant_context(self, tenant):
        """
        Set tenant context for database connection.
        """
        if hasattr(connection, 'set_tenant'):
            connection.set_tenant(tenant)

        # Set session variable for PostgreSQL RLS
        if connection.vendor == 'postgresql':
            with connection.cursor() as cursor:
                cursor.execute("SET app.current_tenant_id = %s", [str(tenant.id)])

    def _clear_database_tenant_context(self):
        """
        Clear tenant context from database connection.
        """
        if hasattr(connection, 'clear_tenant'):
            connection.clear_tenant()

        # Clear session variable for PostgreSQL RLS
        if connection.vendor == 'postgresql':
            with connection.cursor() as cursor:
                cursor.execute("RESET app.current_tenant_id")

    def _set_cache_tenant_context(self, tenant):
        """
        Set tenant context for cache keys.
        """
        if hasattr(cache, 'set_tenant'):
            cache.set_tenant(tenant)

    def _clear_cache_tenant_context(self):
        """
        Clear tenant context from cache.
        """
        if hasattr(cache, 'clear_tenant'):
            cache.clear_tenant()

    def _get_tenant_from_cache(self, cache_key):
        """
        Get tenant from cache.
        """
        if not self.cache_enabled:
            return None

        try:
            return cache.get(cache_key)
        except Exception as e:
            logger.error(f"Error getting tenant from cache {cache_key}: {e}")
            return None

    def _set_tenant_cache(self, cache_key, tenant):
        """
        Set tenant in cache.
        """
        if not self.cache_enabled:
            return

        try:
            cache.set(cache_key, tenant, self.cache_timeout)
        except Exception as e:
            logger.error(f"Error setting tenant cache {cache_key}: {e}")

    def _tenant_cache_key(self, prefix, tenant):
        """
        Generate cache key for tenant.
        """
        return f"tenant:{prefix}:{tenant.id}"


class TenantIsolationMiddleware(MiddlewareMixin):
    """
    Middleware to enforce tenant isolation at the application level.
    """

    def __init__(self, get_response):
        self.get_response = get_response
        self.enforce_isolation = getattr(settings, 'ENFORCE_TENANT_ISOLATION', True)
        self.debug_mode = getattr(settings, 'DEBUG', False)

    def process_view(self, request, view_func, view_args, view_kwargs):
        """
        Validate tenant isolation before view execution.
        """
        if not self.enforce_isolation:
            return None

        # Skip isolation checks for public paths
        if not hasattr(request, 'tenant') or not request.tenant:
            return None

        # Validate that view has access to tenant data
        if hasattr(view_func, 'tenant_required') and view_func.tenant_required:
            if not request.tenant:
                raise PermissionDenied("Tenant context required for this view")

        # Validate tenant-specific view permissions
        if hasattr(view_func, 'tenant_permissions'):
            required_permissions = view_func.tenant_permissions
            user = get_user(request)

            if not user.is_authenticated:
                raise PermissionDenied("Authentication required")

            if not user.is_superuser:
                # Check user permissions against required permissions
                user_permissions = user.get_tenant_permissions()
                for resource, permissions in required_permissions.items():
                    if resource not in user_permissions:
                        raise PermissionDenied(f"No access to {resource}")

                    user_perms = user_permissions[resource]
                    for perm in permissions:
                        if perm not in user_perms:
                            raise PermissionDenied(f"Missing {perm} permission for {resource}")

        return None

    def process_response(self, request, response):
        """
        Add tenant isolation headers to response.
        """
        if hasattr(request, 'tenant') and request.tenant:
            response['X-Tenant-ID'] = str(request.tenant.id)
            response['X-Tenant-Slug'] = request.tenant.slug

            if self.debug_mode:
                response['X-Tenant-Name'] = request.tenant.name
                response['X-Tenant-Plan'] = request.tenant.subscription_plan

        return response


class TenantActivityMiddleware(MiddlewareMixin):
    """
    Middleware to track tenant activity and usage.
    """

    def __init__(self, get_response):
        self.get_response = get_response
        self.tracking_enabled = getattr(settings, 'TENANT_ACTIVITY_TRACKING_ENABLED', True)

    def process_response(self, request, response):
        """
        Track tenant activity after request completion.
        """
        if not self.tracking_enabled:
            return response

        if hasattr(request, 'tenant') and request.tenant:
            # Track API calls
            if request.path.startswith('/api/'):
                self._track_api_usage(request, response)

            # Track user activity
            user = get_user(request)
            if user.is_authenticated:
                self._track_user_activity(request, user, response)

        return response

    def _track_api_usage(self, request, response):
        """
        Track API usage for tenant.
        """
        try:
            # This would integrate with your usage tracking system
            # For now, we'll just log the activity
            logger.info(f"API usage tracked for tenant {request.tenant.id}: {request.method} {request.path}")
        except Exception as e:
            logger.error(f"Error tracking API usage: {e}")

    def _track_user_activity(self, request, user, response):
        """
        Track user activity within tenant.
        """
        try:
            # Update user last login time
            if not user.last_login or (timezone.now() - user.last_login).seconds > 300:  # 5 minutes
                user.last_login = timezone.now()
                user.save(update_fields=['last_login'])

            # Log user activity
            logger.info(f"User activity tracked: {user.id} in tenant {request.tenant.id}")
        except Exception as e:
            logger.error(f"Error tracking user activity: {e}")


def tenant_required(view_func):
    """
    Decorator to mark view as requiring tenant context.
    """
    view_func.tenant_required = True
    return view_func


def tenant_permissions(**permissions):
    """
    Decorator to specify required tenant permissions for view.

    Usage:
        @tenant_permissions(users=['read', 'write'], billing=['read'])
        def my_view(request):
            ...
    """
    def decorator(view_func):
        view_func.tenant_permissions = permissions
        return view_func
    return decorator