Files
igny8/ActiveDocs/03-BACKEND.md
2025-11-09 18:08:23 +05:00

861 lines
24 KiB
Markdown

# IGNY8 Backend Documentation
**Version:** 1.0
**Last Updated:** 2025-01-XX
**Purpose:** Complete backend documentation covering models, views, APIs, modules, serializers, tasks, and structure.
---
## Table of Contents
1. [Backend Overview](#backend-overview)
2. [Tech Stack](#tech-stack)
3. [Project Structure](#project-structure)
4. [Models](#models)
5. [ViewSets](#viewsets)
6. [Serializers](#serializers)
7. [Celery Tasks](#celery-tasks)
8. [API Endpoints](#api-endpoints)
9. [Base Classes](#base-classes)
10. [Middleware](#middleware)
11. [Utilities](#utilities)
12. [Modules](#modules)
---
## Backend Overview
The IGNY8 backend is a Django 5.2+ application using Django REST Framework (DRF) for API endpoints. The backend follows a modular architecture with clear separation of concerns, automatic account isolation, and support for asynchronous task processing via Celery.
### Key Features
- **Multi-Tenancy**: Complete account isolation with automatic filtering
- **RESTful API**: DRF ViewSets with consistent response format
- **Celery Integration**: Asynchronous task processing for long-running operations
- **Account/Site/Sector Hierarchy**: Hierarchical data organization
- **AI Integration**: Unified AIProcessor for all AI operations
- **Progress Tracking**: Real-time progress updates for Celery tasks
---
## Tech Stack
### Core Technologies
- **Django 5.2+**: Web framework
- **Django REST Framework**: API framework
- **PostgreSQL**: Database
- **Celery**: Asynchronous task queue
- **Redis**: Celery broker and caching
### Key Libraries
- **django-filter**: Advanced filtering
- **djangorestframework-simplejwt**: JWT authentication
- **requests**: HTTP client for external APIs
- **python-dotenv**: Environment variable management
---
## Project Structure
```
backend/igny8_core/
├── auth/ # Multi-tenancy and authentication
│ ├── models.py # Account, User, Plan, Site, Sector, Industry models
│ ├── views.py # Account, User, Site, Sector ViewSets
│ ├── serializers.py # Account, User, Plan serializers
│ └── urls.py # Auth module URLs
├── modules/ # Feature modules
│ ├── planner/ # Keywords, Clusters, Ideas
│ │ ├── models.py # Keywords, Clusters, ContentIdeas models
│ │ ├── views.py # KeywordViewSet, ClusterViewSet, ContentIdeasViewSet
│ │ ├── tasks.py # Celery tasks for AI operations
│ │ ├── serializers.py # Model serializers
│ │ └── urls.py # Planner module URLs
│ ├── writer/ # Tasks, Content, Images
│ │ ├── models.py # Tasks, Content, Images models
│ │ ├── views.py # TasksViewSet
│ │ ├── tasks.py # Celery tasks for content/image generation
│ │ └── urls.py # Writer module URLs
│ ├── system/ # Settings, Prompts, Integration
│ │ ├── models.py # AIPrompt, IntegrationSettings, AuthorProfile, Strategy
│ │ ├── views.py # AIPromptViewSet, AuthorProfileViewSet
│ │ ├── integration_views.py # IntegrationSettingsViewSet, task_progress
│ │ ├── utils.py # Default prompts, prompt loading
│ │ └── urls.py # System module URLs
│ └── billing/ # Credits, Transactions, Usage
│ ├── models.py # CreditTransaction, UsageLog models
│ ├── views.py # Billing ViewSets
│ └── services.py # CreditService
├── api/ # API base classes
│ ├── base.py # AccountModelViewSet, SiteSectorModelViewSet
│ └── pagination.py # CustomPageNumberPagination
├── utils/ # Shared utilities
│ ├── ai_processor.py # Unified AI interface
│ └── content_normalizer.py # Content processing utilities
├── middleware/ # Custom middleware
│ ├── account.py # AccountContextMiddleware (sets request.account)
│ └── resource_tracker.py # ResourceTrackerMiddleware (API metrics)
├── settings.py # Django settings
├── urls.py # Root URL configuration
└── celery.py # Celery configuration
```
---
## Models
### Base Models
#### AccountBaseModel
**File**: `auth/models.py`
**Purpose**: Base model for all account-isolated models.
**Fields**:
- `account`: ForeignKey to Account
- `created_at`: DateTimeField (auto_now_add)
- `updated_at`: DateTimeField (auto_now)
**Usage**: All models that need account isolation inherit from this.
#### SiteSectorBaseModel
**File**: `auth/models.py`
**Purpose**: Base model for models that belong to Site and Sector.
**Fields**:
- Inherits from `AccountBaseModel`
- `site`: ForeignKey to Site
- `sector`: ForeignKey to Sector
**Methods**:
- `save()`: Automatically sets `account` from `site.account` and validates sector belongs to site
**Usage**: Models like Keywords, Clusters, ContentIdeas, Tasks inherit from this.
### Auth Models
#### Account
**File**: `auth/models.py`
**Table**: `igny8_accounts`
**Fields**:
- `name`: CharField
- `slug`: SlugField (unique)
- `owner`: ForeignKey to User
- `stripe_customer_id`: CharField (optional)
- `plan`: ForeignKey to Plan
- `credits`: IntegerField (default: 0)
- `status`: CharField (choices: active, suspended, trial, cancelled)
**Methods**:
- `is_system_account()`: Returns True if account is system account (aws-admin, default-account, default)
#### User
**File**: `auth/models.py`
**Table**: `igny8_users`
**Fields**:
- Inherits from `AbstractUser`
- `email`: EmailField (unique, USERNAME_FIELD)
- `account`: ForeignKey to Account
- `role`: CharField (choices: developer, owner, admin, editor, viewer, system_bot)
**Methods**:
- `has_role(*roles)`: Check if user has any of the specified roles
- `is_owner_or_admin()`: Returns True if role is owner or admin
- `is_developer()`: Returns True if role is developer or is_superuser
- `is_admin_or_developer()`: Returns True if role is admin or developer
- `is_system_account_user()`: Returns True if user belongs to system account
- `get_accessible_sites()`: Returns QuerySet of sites user can access
#### Plan
**File**: `auth/models.py`
**Table**: `igny8_plans`
**Fields**:
- `name`: CharField
- `slug`: SlugField (unique)
- `price`: DecimalField
- `billing_cycle`: CharField (choices: monthly, annual)
- `features`: JSONField (array of feature strings)
- `max_users`: IntegerField
- `max_sites`: IntegerField
- `max_keywords`: IntegerField
- `max_clusters`: IntegerField
- `max_content_ideas`: IntegerField
- `daily_cluster_limit`: IntegerField
- `monthly_cluster_ai_credits`: IntegerField
- `daily_content_tasks`: IntegerField
- `daily_ai_requests`: IntegerField
- `monthly_word_count_limit`: IntegerField
- `monthly_content_ai_credits`: IntegerField
- `monthly_image_count`: IntegerField
- `daily_image_generation_limit`: IntegerField
- `monthly_image_ai_credits`: IntegerField
- `max_images_per_task`: IntegerField
- `image_model_choices`: JSONField
- `included_credits`: IntegerField
- `extra_credit_price`: DecimalField
- And more...
**Methods**:
- `clean()`: Validates plan limits
- `get_effective_credits_per_month()`: Returns included_credits or credits_per_month
#### Site
**File**: `auth/models.py`
**Table**: `igny8_sites`
**Fields**:
- Inherits from `AccountBaseModel`
- `name`: CharField
- `slug`: SlugField (unique per account)
- `domain`: URLField (optional)
- `industry`: ForeignKey to Industry (optional)
- `is_active`: BooleanField
- `status`: CharField (choices: active, inactive, suspended)
- `wp_url`: URLField (optional, WordPress integration)
- `wp_username`: CharField (optional)
- `wp_app_password`: CharField (optional)
**Methods**:
- `get_active_sectors_count()`: Get count of active sectors
- `can_add_sector()`: Check if site can add another sector (max 5)
#### Sector
**File**: `auth/models.py`
**Table**: `igny8_sectors`
**Fields**:
- Inherits from `AccountBaseModel`
- `site`: ForeignKey to Site
- `industry_sector`: ForeignKey to IndustrySector (optional, template reference)
- `name`: CharField
- `slug`: SlugField (unique per site)
- `is_active`: BooleanField
- `status`: CharField (choices: active, inactive)
### Planner Models
#### Keywords
**File**: `modules/planner/models.py`
**Table**: `igny8_keywords`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `keyword`: CharField
- `volume`: IntegerField
- `difficulty`: IntegerField
- `intent`: CharField (choices: informational, navigational, commercial, transactional)
- `cluster`: ForeignKey to Clusters (optional)
- `status`: CharField (choices: active, pending, archived)
#### Clusters
**File**: `modules/planner/models.py`
**Table**: `igny8_clusters`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `name`: CharField (unique)
- `description`: TextField
- `keywords_count`: IntegerField
- `volume`: IntegerField
- `mapped_pages`: IntegerField
- `status`: CharField
#### ContentIdeas
**File**: `modules/planner/models.py`
**Table**: `igny8_content_ideas`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `idea_title`: CharField
- `description`: TextField
- `content_structure`: CharField (choices: cluster_hub, landing_page, pillar_page, supporting_page)
- `content_type`: CharField (choices: blog_post, article, guide, tutorial)
- `target_keywords`: CharField (comma-separated, legacy)
- `keyword_objects`: ManyToManyField to Keywords
- `keyword_cluster`: ForeignKey to Clusters
- `status`: CharField (choices: new, scheduled, published)
- `estimated_word_count`: IntegerField
### Writer Models
#### Tasks
**File**: `modules/writer/models.py`
**Table**: `igny8_tasks`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `title`: CharField
- `description`: TextField
- `keywords`: CharField (comma-separated, legacy)
- `keyword_objects`: ManyToManyField to Keywords
- `cluster`: ForeignKey to Clusters
- `idea`: ForeignKey to ContentIdeas
- `content_structure`: CharField
- `content_type`: CharField
- `status`: CharField (choices: queued, in_progress, draft, review, published, completed)
- `content`: TextField (generated content)
- `word_count`: IntegerField
- `meta_title`: CharField
- `meta_description`: TextField
- `assigned_post_id`: IntegerField (WordPress post ID)
- `post_url`: URLField
#### Content
**File**: `modules/writer/models.py`
**Table**: `igny8_content`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `task`: OneToOneField to Tasks
- `html_content`: TextField
- `word_count`: IntegerField
- `metadata`: JSONField
**Methods**:
- `save()`: Automatically sets account, site, sector from task
#### Images
**File**: `modules/writer/models.py`
**Table**: `igny8_images`
**Fields**:
- Inherits from `SiteSectorBaseModel`
- `task`: ForeignKey to Tasks
- `image_type`: CharField (choices: featured, desktop, mobile, in_article)
- `image_url`: URLField
- `image_path`: CharField (local path)
- `prompt`: TextField
- `status`: CharField
- `position`: IntegerField
**Methods**:
- `save()`: Automatically sets account, site, sector from task
### System Models
#### AIPrompt
**File**: `modules/system/models.py`
**Table**: `igny8_ai_prompts`
**Fields**:
- Inherits from `AccountBaseModel`
- `prompt_type`: CharField (choices: clustering, ideas, content_generation, image_prompt_extraction, image_prompt_template, negative_prompt)
- `prompt_value`: TextField
- `default_prompt`: TextField
- `is_active`: BooleanField
**Unique Constraint**: `(account, prompt_type)`
#### IntegrationSettings
**File**: `modules/system/models.py`
**Table**: `igny8_integration_settings`
**Fields**:
- Inherits from `AccountBaseModel`
- `integration_type`: CharField (choices: openai, runware, gsc, image_generation)
- `config`: JSONField (API keys, settings, etc.)
- `is_active`: BooleanField
**Unique Constraint**: `(account, integration_type)`
#### AuthorProfile
**File**: `modules/system/models.py`
**Table**: `igny8_author_profiles`
**Fields**:
- Inherits from `AccountBaseModel`
- `name`: CharField
- `description`: TextField
- `tone`: CharField
- `language`: CharField
- `structure_template`: JSONField
- `is_active`: BooleanField
#### Strategy
**File**: `modules/system/models.py`
**Table**: `igny8_strategies`
**Fields**:
- Inherits from `AccountBaseModel`
- `name`: CharField
- `description`: TextField
- `sector`: ForeignKey to Sector (optional)
- `prompt_types`: JSONField
- `section_logic`: JSONField
- `is_active`: BooleanField
---
## ViewSets
### Base ViewSets
#### AccountModelViewSet
**File**: `api/base.py`
**Purpose**: Base ViewSet with automatic account filtering.
**Methods**:
- `get_queryset()`: Filters queryset by `request.account` (with admin/developer override)
- `perform_create()`: Sets account on created objects
- `get_serializer_context()`: Adds account to serializer context
**Access Control**:
- Admin/Developer users: Bypass account filtering
- System account users: Bypass account filtering
- Regular users: Only see data from their account
#### SiteSectorModelViewSet
**File**: `api/base.py`
**Purpose**: Base ViewSet with site/sector filtering and access control.
**Inherits**: `AccountModelViewSet`
**Methods**:
- `get_queryset()`: Filters by account, accessible sites, and optional site_id/sector_id
- `perform_create()`: Validates site access and sector-site relationship
- `get_serializer_context()`: Adds accessible sites and sectors to context
**Access Control**:
- Developers: All active sites
- System account users: All active sites
- Owners/Admins: All sites in their account
- Editors/Viewers: Only sites granted via `SiteUserAccess`
### Planner ViewSets
#### KeywordViewSet
**File**: `modules/planner/views.py`
**Inherits**: `SiteSectorModelViewSet`
**Actions**:
- `list()`: List keywords with filtering
- `create()`: Create keyword
- `retrieve()`: Get keyword details
- `update()`: Update keyword
- `destroy()`: Delete keyword
- `auto_cluster()`: Auto-cluster keywords using AI
- `bulk_delete()`: Bulk delete keywords
- `bulk_update_status()`: Bulk update keyword status
- `export_csv()`: Export keywords to CSV
- `import_csv()`: Import keywords from CSV
**Filtering**:
- Search: `keyword` field
- Filters: `status`, `cluster_id`, `intent`
- Custom: `difficulty_min`, `difficulty_max`, `volume_min`, `volume_max`
- Ordering: `created_at`, `volume`, `difficulty`
#### ClusterViewSet
**File**: `modules/planner/views.py`
**Inherits**: `SiteSectorModelViewSet`
**Actions**:
- `list()`: List clusters
- `create()`: Create cluster
- `retrieve()`: Get cluster details
- `update()`: Update cluster
- `destroy()`: Delete cluster
- `auto_generate_ideas()`: Auto-generate content ideas for clusters
#### ContentIdeasViewSet
**File**: `modules/planner/views.py`
**Inherits**: `SiteSectorModelViewSet`
**Actions**:
- `list()`: List content ideas
- `create()`: Create content idea
- `retrieve()`: Get content idea details
- `update()`: Update content idea
- `destroy()`: Delete content idea
### Writer ViewSets
#### TasksViewSet
**File**: `modules/writer/views.py`
**Inherits**: `SiteSectorModelViewSet`
**Actions**:
- `list()`: List tasks
- `create()`: Create task
- `retrieve()`: Get task details
- `update()`: Update task
- `destroy()`: Delete task
- `auto_generate_content()`: Auto-generate content for tasks
- `auto_generate_images()`: Auto-generate images for tasks
- `bulk_delete()`: Bulk delete tasks
- `bulk_update()`: Bulk update task status
**Filtering**:
- Search: `title`, `keywords`
- Filters: `status`, `cluster_id`, `content_type`, `content_structure`
- Ordering: `title`, `created_at`, `word_count`, `status`
### System ViewSets
#### IntegrationSettingsViewSet
**File**: `modules/system/integration_views.py`
**Inherits**: `viewsets.ViewSet`
**Actions**:
- `list()`: List integrations
- `retrieve()`: Get integration settings
- `update()`: Save integration settings
- `save_post()`: Save integration settings (POST)
- `test_connection()`: Test API connection
- `test_openai()`: Test OpenAI connection
- `test_runware()`: Test Runware connection
- `generate_image()`: Test image generation
- `task_progress()`: Get Celery task progress
#### AIPromptViewSet
**File**: `modules/system/views.py`
**Inherits**: `AccountModelViewSet`
**Actions**:
- `list()`: List prompts
- `create()`: Create prompt
- `retrieve()`: Get prompt details
- `update()`: Update prompt
- `destroy()`: Delete prompt
- `reset_to_default()`: Reset prompt to default value
#### AuthorProfileViewSet
**File**: `modules/system/views.py`
**Inherits**: `AccountModelViewSet`
**Actions**:
- `list()`: List author profiles
- `create()`: Create author profile
- `retrieve()`: Get author profile details
- `update()`: Update author profile
- `destroy()`: Delete author profile
---
## Serializers
### Planner Serializers
#### KeywordSerializer
**File**: `modules/planner/serializers.py`
**Fields**: All Keyword model fields
**Validation**: Validates keyword uniqueness, cluster belongs to same sector
#### ClusterSerializer
**File**: `modules/planner/cluster_serializers.py`
**Fields**: All Cluster model fields
**Read-Only Fields**: `keywords_count`, `volume` (calculated)
#### ContentIdeasSerializer
**File**: `modules/planner/serializers.py`
**Fields**: All ContentIdeas model fields
### Writer Serializers
#### TasksSerializer
**File**: `modules/writer/serializers.py`
**Fields**: All Tasks model fields
#### ContentSerializer
**File**: `modules/writer/serializers.py`
**Fields**: All Content model fields
#### ImagesSerializer
**File**: `modules/writer/serializers.py`
**Fields**: All Images model fields
### System Serializers
#### AIPromptSerializer
**File**: `modules/system/serializers.py`
**Fields**: All AIPrompt model fields
#### IntegrationSettingsSerializer
**File**: `modules/system/serializers.py`
**Fields**: All IntegrationSettings model fields
---
## Celery Tasks
### Planner Tasks
#### auto_cluster_keywords_task
**File**: `modules/planner/tasks.py`
**Purpose**: Auto-cluster keywords using AI.
**Parameters**:
- `keyword_ids`: List of keyword IDs
- `account_id`: Account ID
- `site_id`: Site ID
- `sector_id`: Sector ID
**Progress Tracking**: Updates progress with request_steps and response_steps
**Calls**: `_auto_cluster_keywords_core()`
#### auto_generate_ideas_task
**File**: `modules/planner/tasks.py`
**Purpose**: Auto-generate content ideas for clusters.
**Parameters**:
- `cluster_ids`: List of cluster IDs
- `account_id`: Account ID
**Progress Tracking**: Updates progress for each cluster
**Calls**: `_generate_single_idea_core()` for each cluster
### Writer Tasks
#### auto_generate_content_task
**File**: `modules/writer/tasks.py`
**Purpose**: Auto-generate content for tasks.
**Parameters**:
- `task_ids`: List of task IDs
- `account_id`: Account ID
**Progress Tracking**: Updates progress for each task
**Calls**: `AIProcessor.generate_content()`
#### auto_generate_images_task
**File**: `modules/writer/tasks.py`
**Purpose**: Auto-generate images for tasks.
**Parameters**:
- `task_ids`: List of task IDs
- `account_id`: Account ID
**Progress Tracking**: Updates progress for each task
**Calls**: `AIProcessor.extract_image_prompts()` and `AIProcessor.generate_image()`
---
## API Endpoints
### Base URL
`/api/v1/`
### Planner Endpoints
- `GET /api/v1/planner/keywords/` - List keywords
- `POST /api/v1/planner/keywords/` - Create keyword
- `GET /api/v1/planner/keywords/{id}/` - Get keyword
- `PUT /api/v1/planner/keywords/{id}/` - Update keyword
- `DELETE /api/v1/planner/keywords/{id}/` - Delete keyword
- `POST /api/v1/planner/keywords/auto_cluster/` - Auto-cluster keywords
- `POST /api/v1/planner/keywords/bulk_delete/` - Bulk delete keywords
- `POST /api/v1/planner/keywords/bulk_update_status/` - Bulk update status
- `GET /api/v1/planner/keywords/export_csv/` - Export keywords
- `POST /api/v1/planner/keywords/import_csv/` - Import keywords
- `GET /api/v1/planner/clusters/` - List clusters
- `POST /api/v1/planner/clusters/auto_generate_ideas/` - Auto-generate ideas
- `GET /api/v1/planner/ideas/` - List content ideas
### Writer Endpoints
- `GET /api/v1/writer/tasks/` - List tasks
- `POST /api/v1/writer/tasks/auto_generate_content/` - Auto-generate content
- `POST /api/v1/writer/tasks/auto_generate_images/` - Auto-generate images
### System Endpoints
- `GET /api/v1/system/settings/integrations/{pk}/` - Get integration settings
- `PUT /api/v1/system/settings/integrations/{pk}/` - Save integration settings
- `POST /api/v1/system/settings/integrations/{pk}/test_openai/` - Test OpenAI
- `POST /api/v1/system/settings/integrations/{pk}/test_runware/` - Test Runware
- `POST /api/v1/system/settings/integrations/{pk}/generate_image/` - Test image generation
- `GET /api/v1/system/settings/task_progress/{task_id}/` - Get task progress
---
## Base Classes
### AccountModelViewSet
**File**: `api/base.py`
**Purpose**: Base ViewSet with automatic account filtering.
**Features**:
- Automatic account filtering
- Admin/Developer override
- Account context in serializers
### SiteSectorModelViewSet
**File**: `api/base.py`
**Purpose**: Base ViewSet with site/sector filtering.
**Features**:
- Account filtering (inherited)
- Site access control
- Sector validation
- Accessible sites/sectors in serializer context
---
## Middleware
### AccountContextMiddleware
**File**: `middleware/account.py`
**Purpose**: Sets `request.account` from JWT token.
**Functionality**:
- Extracts account ID from JWT token
- Loads Account object
- Sets `request.account`
### ResourceTrackerMiddleware
**File**: `middleware/resource_tracker.py`
**Purpose**: Tracks API request metrics.
**Functionality**:
- Tracks CPU, memory, I/O usage
- Stores metrics in cache
- Provides metrics endpoint
---
## Utilities
### AIProcessor
**File**: `utils/ai_processor.py`
**Purpose**: Unified AI interface for all AI operations.
**Methods**:
- `cluster_keywords()`: Cluster keywords using AI
- `generate_ideas()`: Generate content ideas
- `generate_content()`: Generate text content
- `extract_image_prompts()`: Extract image prompts from content
- `generate_image()`: Generate images using OpenAI DALL-E or Runware
**See**: AI Functions documentation for complete details
### Content Normalizer
**File**: `utils/content_normalizer.py`
**Purpose**: Content processing utilities.
**Functions**:
- `_extract_body_content()`: Extract body content from HTML
---
## Modules
### Planner Module
**Purpose**: Keyword management and content planning.
**Models**: Keywords, Clusters, ContentIdeas
**ViewSets**: KeywordViewSet, ClusterViewSet, ContentIdeasViewSet
**Tasks**: auto_cluster_keywords_task, auto_generate_ideas_task
### Writer Module
**Purpose**: Content generation and management.
**Models**: Tasks, Content, Images
**ViewSets**: TasksViewSet
**Tasks**: auto_generate_content_task, auto_generate_images_task
### System Module
**Purpose**: System settings, prompts, and integrations.
**Models**: AIPrompt, IntegrationSettings, AuthorProfile, Strategy
**ViewSets**: AIPromptViewSet, AuthorProfileViewSet, IntegrationSettingsViewSet
**Utilities**: Default prompts, prompt loading
### Billing Module
**Purpose**: Credits, transactions, and usage tracking.
**Models**: CreditTransaction, UsageLog
**ViewSets**: CreditTransactionViewSet, UsageLogViewSet
**Services**: CreditService
---
## Summary
The IGNY8 backend is built on:
1. **Django + DRF**: Robust web framework with RESTful API
2. **Multi-Tenancy**: Complete account isolation with automatic filtering
3. **Modular Architecture**: Clear module boundaries with shared utilities
4. **Celery Integration**: Asynchronous task processing for long-running operations
5. **Base ViewSets**: Consistent access control and filtering
6. **AI Integration**: Unified AIProcessor for all AI operations
7. **Progress Tracking**: Real-time progress updates for Celery tasks
8. **Account/Site/Sector Hierarchy**: Hierarchical data organization
This architecture ensures scalability, maintainability, and extensibility while providing a robust API for the frontend.