model.constant.ts

The constant file defines the data shape (schema) of your domain. It ensures consistency between database storage, API communication, and frontend usage by creating a single source of truth.

📐Type Safety

Schema definitions are shared across backend and frontend, ensuring type consistency throughout your application.

🔄Auto Generation

From your constant definitions, the framework auto-generates MongoDB schemas, GraphQL types, REST APIs, and client-side types.

Class Hierarchy Pattern

Akan.js uses a hierarchical class pattern to organize schema definitions. Each layer has a specific purpose and builds upon the previous one.

Input

Fields required to create a new document (user-provided data)

Object

Input + system-generated fields (status, timestamps, computed values)

Light

Lightweight version with selected fields for list queries (prevents overfetching)

Full (Model)

Complete model with all fields and static/instance utility methods

Insight

Statistical aggregation fields for analytics (count, sum, avg)

Defining Enums with enumOf()

Before defining your model classes, you typically need to define enums for status fields and other categorical values. The enumOf() function creates type-safe enum classes.

product.constant.ts

Key points about enumOf():

field	Description
as const	Required for TypeScript to infer literal types instead of string[]
First argument	The enum name used in dictionary translations
Generated type	Access values via ProductStatus["value"] type (e.g., "active" \| "soldOut" \| ...)

as const

Required for TypeScript to infer literal types instead of string[]

First argument

The enum name used in dictionary translations

Generated type

Access values via ProductStatus["value"] type (e.g., "active" | "soldOut" | ...)

Input Class - Creation Fields

The Input class defines fields required when creating a new document. These are the values a user or system must provide.

product.constant.ts

Understanding the via() pattern:

method	Description
via((field) => ({...}))	Creates a new class with the specified fields. The field() function takes a type and optional options.
field(Type, options?)	Type can be: String, Int, Float, Boolean, Date, JSON, ID, [Array], custom classes, or enums.
.optional()	Chain .optional() to make a field nullable. The field can then be omitted or set to null.

via((field) => ({...}))

Creates a new class with the specified fields. The field() function takes a type and optional options.

field(Type, options?)

Type can be: String, Int, Float, Boolean, Date, JSON, ID, [Array], custom classes, or enums.

.optional()

Chain .optional() to make a field nullable. The field can then be omitted or set to null.

💡Usage in Code

Input fields are used in: fetch.createProduct(input), st.do.setNameOnProduct(value), productForm state

Object Class

The Object class extends Input and adds fields that are automatically set by the system. These fields are typically modified through service methods, not through direct user input.

product.constant.ts

Common Object fields:

field	Description
status	Document lifecycle state, controlled by service methods
stock	Inventory count, default to 0, modified by sell operations
soldCount	Accumulated sales count for analytics

status

Document lifecycle state, controlled by service methods

stock

Inventory count, default to 0, modified by sell operations

soldCount

Accumulated sales count for analytics

Note: The via(BaseClass, (field) => ({...})) syntax means: 'extend BaseClass and add these new fields'. All fields from ProductInput are inherited.

Light Class - Optimized for Lists

The Light class picks only necessary fields from Object for list views. This prevents overfetching when displaying multiple items in tables or cards.

product.constant.ts

field	Description
Performance Benefits	When fetching 100 products for a list, Light ensures only essential fields are transferred, reducing payload size and database load.
Field Selection	Choose fields that appear in list UIs: identifiers, status, key metadata. Exclude: content, files, detailed nested objects.
Resolve Function	The (resolve) => ({}) function can add computed/virtual fields that are calculated server-side but not stored in DB.

Performance Benefits

When fetching 100 products for a list, Light ensures only essential fields are transferred, reducing payload size and database load.

Field Selection

Choose fields that appear in list UIs: identifiers, status, key metadata. Exclude: content, files, detailed nested objects.

Resolve Function

The (resolve) => ({}) function can add computed/virtual fields that are calculated server-side but not stored in DB.

Full Model - Complete with Methods

The Full model (often named after the domain like 'Product') combines Object and Light, and adds static utility methods for data manipulation.

product.constant.ts

method	Description
Static Methods	Static methods operate on arrays of items. Use for filtering, grouping, or computing derived data from lists.
Usage Example	const activeProducts = cnst.Product.getActiveList(productList);

Static Methods

Static methods operate on arrays of items. Use for filtering, grouping, or computing derived data from lists.

Usage Example

const activeProducts = cnst.Product.getActiveList(productList);

Insight Class

The Insight class defines aggregation fields for analytics. When querying, the framework calculates these statistics using MongoDB aggregation pipeline.

product.constant.ts

Understanding the accumulate option:

field	Description
{ $sum: 1 }	Simple count - adds 1 for each matching document
{ $sum: "$fieldName" }	Sum of field values across all matching documents
{ $cond: [...] }	Conditional aggregation - count only documents matching a condition

{ $sum: 1 }

Simple count - adds 1 for each matching document

{ $sum: "$fieldName" }

Sum of field values across all matching documents

{ $cond: [...] }

Conditional aggregation - count only documents matching a condition

Usage in code:

Using Insight

Scalar - Embedded Objects

Scalars are embedded objects that don't have their own collection. They're stored within the parent document and defined using via() without the full class hierarchy.

milestone.constant.ts

Using scalars in parent models:

bizContract.constant.ts

field	Description
When to use Scalars	Data that belongs to a single parent (1:N embedded). No need for separate queries or references. Data always accessed with parent document.
Scalar Location	Scalars are typically placed in __scalar/ directory to indicate they're embedded types, not standalone collections.

When to use Scalars

Data that belongs to a single parent (1:N embedded). No need for separate queries or references. Data always accessed with parent document.

Scalar Location

Scalars are typically placed in __scalar/ directory to indicate they're embedded types, not standalone collections.

Relation Patterns

There are three main ways to establish relationships between models in Akan.js, following MongoDB schema design principles.

method	Description	Example
Scalar Embedding	Embed scalar objects directly in the document. Best for 1:N relationships where child data is always accessed with parent.
Reference by ID	Store only the ObjectID of related document. Best when you need to query the referenced collection separately.
Light Model Reference	Embed a Light version of another model. Best when you need key fields without full population.

Scalar Embedding

Embed scalar objects directly in the document. Best for 1:N relationships where child data is always accessed with parent.

Reference by ID

Store only the ObjectID of related document. Best when you need to query the referenced collection separately.

Light Model Reference

Embed a Light version of another model. Best when you need key fields without full population.

⚠️ Import Warning: Direct imports between constant files can cause circular references. Always import from the specific file path, not from barrel files (index.ts).

Field Options Reference

The field() function accepts various options to configure validation, defaults, database behavior, and more. Here's the complete reference:

Option	Type	Default	Description
default	any \| (() => any)	-	Default value when no value is provided. Can be a static value or a function for dynamic defaults.
nullable	boolean	false	If true, the field can be null or omitted. Equivalent to using .optional() chain method.
ref	string	-	Reference collection name for ID fields. Used for MongoDB population.
refPath	string	-	Dynamic reference path - uses another field's value to determine the referenced collection.
refType	"child" \| "parent" \| "relation"	-	Type of reference relationship. Affects how population and cascading operations work.
type	"email" \| "password" \| "url"	-	Preset type that applies default validation and example values for common patterns.
fieldType	"property" \| "hidden" \| "resolve"	"property"	property: normal field, hidden: backend-only (not sent to frontend), resolve: computed field (not stored in DB).
immutable	boolean	false	If true, field cannot be modified after document creation. Only set during create.
min	number	-	Minimum value constraint for Int or Float fields. Validation runs on save.
max	number	-	Maximum value constraint for Int or Float fields. Validation runs on save.
minlength	number	-	Minimum string length constraint for String fields.
maxlength	number	-	Maximum string length constraint for String fields.
enum	EnumClass	-	Restrict field values to enum values. Typically used with enumOf() defined classes.
select	boolean	true	If false, field is excluded from default queries. Use for sensitive or large fields.
accumulate	object	-	MongoDB aggregation expression for Insight fields. Calculates statistics across matched documents.
example	any	-	Example value for API documentation generation (Swagger/OpenAPI).
of	Type	-	Value type for Map fields. The key is always string, value type is specified by 'of'.
validate	(value, model) => boolean	-	Custom validation function. Receives the field value and full model, returns boolean.
text	"search" \| "filter"	-	Enable text indexing for full-text search (search) or filtering (filter) capabilities.
meta	object	-	Custom metadata object for additional field information used by UI components or plugins.

default:any | (() => any):-

Default value when no value is provided. Can be a static value or a function for dynamic defaults.

nullable:boolean:false

If true, the field can be null or omitted. Equivalent to using .optional() chain method.

ref:string:-

Reference collection name for ID fields. Used for MongoDB population.

refPath:string:-

Dynamic reference path - uses another field's value to determine the referenced collection.

refType:"child" | "parent" | "relation":-

Type of reference relationship. Affects how population and cascading operations work.

type:"email" | "password" | "url":-

Preset type that applies default validation and example values for common patterns.

fieldType Options

field	Description
property	Normal field - stored in DB, sent to frontend
hidden	Backend only - stored in DB but never sent to frontend (e.g., OTP codes)
resolve	Virtual field - computed on query, not stored in DB (e.g., view count from other collection)

property

Normal field - stored in DB, sent to frontend

hidden

Backend only - stored in DB but never sent to frontend (e.g., OTP codes)

resolve

Virtual field - computed on query, not stored in DB (e.g., view count from other collection)

refType Options

field	Description
parent	This document belongs to the referenced document
child	Referenced document belongs to this document
relation	Many-to-many relationship without ownership

parent

This document belongs to the referenced document

child

Common field patterns:

Common Field Patterns

Constant Best Practices

field	Description
Class Ordering	Define in order: Enums → Input → Object → Light → Full → Insight. This prevents forward reference issues.
Light Field Selection	Include: id, status, key identifiers, timestamps. Exclude: large content fields, files, detailed nested objects.
Default Values	Always provide defaults for Object fields. Use functions for dynamic values like dates: () => dayjs()
Static Methods	Add static utility methods to Full class for filtering/grouping logic. Keep them pure (no side effects).
Import Paths	Always use direct file imports for other constants to avoid circular references. Never import from barrel files.

Class Ordering

Define in order: Enums → Input → Object → Light → Full → Insight. This prevents forward reference issues.

Light Field Selection

Include: id, status, key identifiers, timestamps. Exclude: large content fields, files, detailed nested objects.

Default Values

Always provide defaults for Object fields. Use functions for dynamic values like dates: () => dayjs()

Static Methods

Add static utility methods to Full class for filtering/grouping logic. Keep them pure (no side effects).

Import Paths

Always use direct file imports for other constants to avoid circular references. Never import from barrel files.

model.constant.ts

Class Hierarchy Pattern

Defining Enums with enumOf()

Input Class - Creation Fields

Object Class

Light Class - Optimized for Lists

Full Model - Complete with Methods

Insight Class

Scalar - Embedded Objects

Relation Patterns

Field Options Reference

Constant Best Practices

Overview

model.dictionary.ts

field

Description

as const

Required for TypeScript to infer literal types instead of string[]

First argument

The enum name used in dictionary translations

Generated type

Access values via ProductStatus["value"] type (e.g., "active" | "soldOut" | ...)

method

Description

via((field) => ({...}))

Creates a new class with the specified fields. The field() function takes a type and optional options.

field(Type, options?)

Type can be: String, Int, Float, Boolean, Date, JSON, ID, [Array], custom classes, or enums.

.optional()

Chain .optional() to make a field nullable. The field can then be omitted or set to null.

field

Description

status

Document lifecycle state, controlled by service methods

stock

Inventory count, default to 0, modified by sell operations

soldCount

Accumulated sales count for analytics

field

Description

Performance Benefits

When fetching 100 products for a list, Light ensures only essential fields are transferred, reducing payload size and database load.

Field Selection

Choose fields that appear in list UIs: identifiers, status, key metadata. Exclude: content, files, detailed nested objects.

Resolve Function

The (resolve) => ({}) function can add computed/virtual fields that are calculated server-side but not stored in DB.

method

Description

Static Methods

Static methods operate on arrays of items. Use for filtering, grouping, or computing derived data from lists.

Usage Example

const activeProducts = cnst.Product.getActiveList(productList);

field

Description

{ $sum: 1 }

Simple count - adds 1 for each matching document

{ $sum: "$fieldName" }

Sum of field values across all matching documents

{ $cond: [...] }

Conditional aggregation - count only documents matching a condition

field

Description

When to use Scalars

Data that belongs to a single parent (1:N embedded). No need for separate queries or references. Data always accessed with parent document.

Scalar Location

Scalars are typically placed in __scalar/ directory to indicate they're embedded types, not standalone collections.

method

Description

Example

Scalar Embedding

Embed scalar objects directly in the document. Best for 1:N relationships where child data is always accessed with parent.

Reference by ID

Store only the ObjectID of related document. Best when you need to query the referenced collection separately.

Light Model Reference

Embed a Light version of another model. Best when you need key fields without full population.

Option

Type

Default

Description

Example

default

any | (() => any)

Default value when no value is provided. Can be a static value or a function for dynamic defaults.

nullable

boolean

false

If true, the field can be null or omitted. Equivalent to using .optional() chain method.

ref

string

Reference collection name for ID fields. Used for MongoDB population.

refPath

string

Dynamic reference path - uses another field's value to determine the referenced collection.

refType

"child" | "parent" | "relation"

Type of reference relationship. Affects how population and cascading operations work.

type

"email" | "password" | "url"

Preset type that applies default validation and example values for common patterns.

fieldType

"property" | "hidden" | "resolve"

"property"

property: normal field, hidden: backend-only (not sent to frontend), resolve: computed field (not stored in DB).

immutable

boolean

false

If true, field cannot be modified after document creation. Only set during create.

min

number

Minimum value constraint for Int or Float fields. Validation runs on save.

max

number

Maximum value constraint for Int or Float fields. Validation runs on save.

minlength

number

Minimum string length constraint for String fields.

maxlength

number

Maximum string length constraint for String fields.

enum

EnumClass

Restrict field values to enum values. Typically used with enumOf() defined classes.

select

boolean

true

If false, field is excluded from default queries. Use for sensitive or large fields.

accumulate

object

MongoDB aggregation expression for Insight fields. Calculates statistics across matched documents.

example

any

Example value for API documentation generation (Swagger/OpenAPI).

Type

Value type for Map fields. The key is always string, value type is specified by 'of'.

validate

(value, model) => boolean

Custom validation function. Receives the field value and full model, returns boolean.

text

"search" | "filter"

Enable text indexing for full-text search (search) or filtering (filter) capabilities.