Deep Dive: Repositories
Technical reference for repository classes - the data access layer in Ignis.
Files: packages/core/src/base/repositories/core/*.ts
Quick Reference
| Class | Capabilities | Use Case |
|---|---|---|
| AbstractRepository | Base class with properties | Extend for custom repositories |
| ReadableRepository | Read-only operations | Views, external tables |
| PersistableRepository | Read + Write operations | Rarely used directly |
| DefaultCRUDRepository | Full CRUD operations | Standard data tables ✅ |
AbstractRepository
Base class for all repositories - sets up fundamental properties and dependencies for data access.
File: packages/core/src/base/repositories/core/base.ts
Key Properties
entity(BaseEntity): An instance of the model class associated with this repository. It provides access to the Drizzle schema. Auto-resolved from@repositorymetadata or passed in constructor.dataSource(IDataSource): The datasource instance injected into the repository, which holds the database connection. Auto-injected from@repositorydecorator or passed in constructor.connector: A getter that provides direct access to the Drizzle ORM instance from the datasource.filterBuilder(DrizzleFilterBuilder): An instance of the filter builder responsible for convertingIgnis's filter objects into Drizzle-compatible query options.operationScope(TRepositoryOperationScope): Defines whether the repository is read-only or read-write.defaultLimit(number): Default limit for queries (default: 10).
Key Methods
getEntityRelations(): Returns a map of relation configurations from the entity's staticrelationsproperty.
Abstract Methods
AbstractRepository defines the method signatures for standard CRUD operations that concrete repository classes must implement:
count()existsWith()find()findOne()findById()create()updateById()deleteById()- (and
...Allvariants)
ReadableRepository
The ReadableRepository provides a read-only implementation of the repository pattern. It is ideal for data sources that should not be modified, such as views or tables from an external system.
- File:
packages/core/src/base/repositories/core/readable.ts
Implemented Methods
ReadableRepository provides concrete implementations for all read operations:
find(opts): Returns an array of entities matching the filter.findOne(opts): Returns the first entity matching the filter.findById(opts): A convenience method that callsfindOnewith an ID-basedwhereclause.count(opts): Returns the number of entities matching thewhereclause.existsWith(opts): Returnstrueif at least one entity matches thewhereclause.
Write Operations
ReadableRepository throws a "NOT ALLOWED" error for all write operations (create, update, delete).
PersistableRepository
The PersistableRepository extends ReadableRepository and adds write operations. It provides the core logic for creating, updating, and deleting records with built-in safety mechanisms.
- File:
packages/core/src/base/repositories/core/persistable.ts
Implemented Methods
create(opts)createAll(opts)updateById(opts)updateAll(opts)updateBy(opts)- Alias forupdateAlldeleteById(opts)deleteAll(opts)deleteBy(opts)- Alias fordeleteAll
Safety Features
Empty Where Clause Protection
The PersistableRepository includes safety mechanisms to prevent accidental mass updates or deletions:
Update Operations (updateAll):
// ❌ Throws error: Empty where condition without force flag
await repository.updateAll({
data: { status: 'inactive' },
where: {}, // Empty condition
});
// ✅ Warning logged: Explicitly allow mass update with force flag
await repository.updateAll({
data: { status: 'inactive' },
where: {},
options: { force: true }, // Force flag allows empty where
});Delete Operations (deleteAll):
// ❌ Throws error: Empty where condition without force flag
await repository.deleteAll({
where: {}, // Empty condition
});
// ✅ Warning logged: Explicitly allow mass delete with force flag
await repository.deleteAll({
where: {},
options: { force: true }, // Force flag allows empty where
});Behavior Summary
| Scenario | force: false (default) | force: true |
|---|---|---|
Empty where clause | ❌ Throws error | ✅ Logs warning and proceeds |
Valid where clause | ✅ Executes normally | ✅ Executes normally |
Warning Messages:
When performing operations with empty where conditions and force: true, the repository logs a warning:
[_update] Entity: MyEntity | Performing update with empty condition | data: {...} | condition: {}
[_delete] Entity: MyEntity | Performing delete with empty condition | condition: {}This helps track potentially dangerous operations in your logs.
You will typically not use this class directly, but rather the DefaultCRUDRepository.
DefaultCRUDRepository
This is the primary class you should extend for repositories that require full Create, Read, Update, and Delete (CRUD) capabilities. It extends PersistableRepository and serves as the standard, full-featured repository implementation.
- File:
packages/core/src/base/repositories/core/default-crud.ts
@repository Decorator Requirements
IMPORTANT: Both model AND dataSource are required in the @repository decorator for schema auto-discovery. Without both, the model won't be registered and relational queries will fail.
// ❌ WRONG - Will throw error
@repository({ model: User }) // Missing dataSource!
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {}
// ❌ WRONG - Will throw error
@repository({ dataSource: PostgresDataSource }) // Missing model!
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {}
// ✅ CORRECT - Both model and dataSource provided
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {}Injection Patterns
The @repository decorator supports two injection patterns:
Pattern 1: Zero Boilerplate (Recommended)
DataSource is auto-injected from metadata - no constructor needed:
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {
// No constructor needed - datasource auto-injected at param index 0
async findByEmail(email: string) {
return this.findOne({ filter: { where: { email } } });
}
}Pattern 2: Explicit @inject
When you need constructor control, use explicit @inject. Important: The first parameter must extend AbstractDataSource - this is enforced via reflection:
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {
constructor(
@inject({ key: 'datasources.PostgresDataSource' })
dataSource: PostgresDataSource, // ✅ Must be concrete DataSource type, NOT 'any'
) {
super(dataSource);
}
}Note: When @inject is at param index 0, auto-injection is skipped (your @inject takes precedence).
Constructor Type Validation
The framework validates constructor parameters at decorator time:
- First parameter must extend
AbstractDataSource- Usingany,object, or non-DataSource types will throw an error - Type compatibility check - The constructor parameter type must be compatible with the
dataSourcespecified in@repository
// ❌ Error: First parameter must extend AbstractDataSource | Received: 'Object'
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {
constructor(
@inject({ key: 'datasources.PostgresDataSource' })
dataSource: any, // Will cause runtime error!
) {
super(dataSource);
}
}
// ❌ Error: Type mismatch | Constructor expects 'MongoDataSource' but @repository specifies 'PostgresDataSource'
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {
constructor(
@inject({ key: 'datasources.MongoDataSource' })
dataSource: MongoDataSource, // Wrong type!
) {
super(dataSource);
}
}Example Implementation
// src/repositories/configuration.repository.ts
import { Configuration, TConfigurationSchema } from '@/models/entities';
import { PostgresDataSource } from '@/datasources';
import { inject, repository, DefaultCRUDRepository } from '@venizia/ignis';
// Pattern 1: Zero boilerplate (recommended)
@repository({ model: Configuration, dataSource: PostgresDataSource })
export class ConfigurationRepository extends DefaultCRUDRepository<TConfigurationSchema> {
// No constructor needed - datasource and entity auto-resolved!
// Custom data access methods
async findByCode(code: string): Promise<Configuration | undefined> {
const result = await this.connector.query.Configuration.findFirst({
where: (table, { eq }) => eq(table.code, code)
});
return result;
}
}
// Pattern 2: With explicit constructor (when you need custom initialization)
@repository({ model: Configuration, dataSource: PostgresDataSource })
export class ConfigurationRepository extends DefaultCRUDRepository<TConfigurationSchema> {
constructor(
@inject({ key: 'datasources.PostgresDataSource' })
dataSource: PostgresDataSource,
) {
super(dataSource); // Just pass dataSource - entity and relations auto-resolved!
}
}This architecture provides a clean and powerful abstraction for data access, separating the "how" of data fetching (Drizzle logic) from the "what" of business logic (services).
Advanced Features
Log Option for Debugging
All CRUD operations support a log option for debugging:
// Enable logging for a specific operation
await repo.create({
data: { name: 'John', email: 'john@example.com' },
options: {
log: { use: true, level: 'debug' }
}
});
// Output: [_create] Executing with opts: { data: [...], options: {...} }
// Available log levels: 'debug', 'info', 'warn', 'error'
await repo.updateById({
id: '123',
data: { name: 'Jane' },
options: { log: { use: true, level: 'info' } }
});Available on: create, createAll, updateById, updateAll, deleteById, deleteAll
TypeScript Return Type Inference
Repository methods now have improved type inference based on shouldReturn:
// When shouldReturn: false - TypeScript knows data is null
const result1 = await repo.create({
data: { name: 'John' },
options: { shouldReturn: false }
});
// Type: Promise<TCount & { data: null }>
console.log(result1.data); // null
// When shouldReturn: true (default) - TypeScript knows data is the entity
const result2 = await repo.create({
data: { name: 'John' },
options: { shouldReturn: true }
});
// Type: Promise<TCount & { data: User }>
console.log(result2.data.name); // 'John' - fully typed!
// Same for array operations
const results = await repo.createAll({
data: [{ name: 'John' }, { name: 'Jane' }],
options: { shouldReturn: true }
});
// Type: Promise<TCount & { data: User[] }>Relations Auto-Resolution
Relations are now automatically resolved from the entity's static relations property:
// Define entity with static relations
@model({ type: 'entity' })
export class User extends BaseEntity<typeof User.schema> {
static override schema = userTable;
static override relations = () => userRelations.definitions;
}
// Repository automatically uses entity's relations
@repository({ model: User, dataSource: PostgresDataSource })
export class UserRepository extends DefaultCRUDRepository<typeof User.schema> {
// No need to pass relations in constructor - auto-resolved!
}
// Relations are available for include queries
const users = await repo.find({
filter: {
where: { status: 'active' },
include: [{ relation: 'posts' }], // Works automatically
}
});Query Interface Validation
The getQueryInterface() method validates that the entity's schema is properly registered:
// If schema key doesn't match, you get a helpful error:
// Error: [UserRepository] Schema key mismatch | Entity name 'User' not found in connector.query | Available keys: [Configuration, Post] | Ensure the model's TABLE_NAME matches the schema registration keyPerformance Optimizations
Core API for Flat Queries
The ReadableRepository automatically optimizes flat queries (no relations, no field selection) using Drizzle's Core API instead of Query API. This provides ~15-20% performance improvement for simple queries.
Automatic Optimization:
// This query is automatically optimized to use Core API
const users = await repo.find({
filter: {
where: { status: 'active' },
limit: 10,
order: ['createdAt DESC'],
}
});
// Uses: db.select().from(table).where(...).orderBy(...).limit(10)
// This query uses Query API (relations need relational mapper)
const usersWithPosts = await repo.find({
filter: {
where: { status: 'active' },
include: [{ relation: 'posts' }], // Has relations
}
});
// Uses: db.query.tableName.findMany({ with: { posts: true }, ... })When Core API is used:
| Filter Options | API Used | Reason |
|---|---|---|
where, limit, order, offset only | Core API | Flat query, no overhead |
Has include (relations) | Query API | Needs relational mapper |
Has fields selection | Query API | Core API field syntax differs |
Protected Helper Method:
For advanced use cases, you can directly use the findWithCoreAPI method:
// Available in subclasses
protected async findWithCoreAPI(opts: {
filter: TFilter<DataObject>;
findOne?: boolean;
}): Promise<Array<DataObject>>;
// Check if Core API can be used
protected canUseCoreAPI(filter: TFilter<DataObject>): boolean;WeakMap Cache for Filter Builder
The DrizzleFilterBuilder uses a static WeakMap cache for getTableColumns() results, avoiding repeated reflection calls:
// Internal optimization - no action needed
// First call: getTableColumns(schema) → cached
// Subsequent calls: retrieved from WeakMapThis is especially beneficial for:
- High-concurrency environments
- Queries with nested AND/OR conditions (each recursion reuses cache)
- Multiple queries to the same table
Query Operators
The filter builder supports a comprehensive set of query operators for building complex queries.
File: packages/core/src/base/repositories/operators/query.ts
Available Operators
| Operator | Alias | SQL Equivalent | Description |
|---|---|---|---|
eq | - | = | Equal to |
ne | neq | != | Not equal to |
gt | - | > | Greater than |
gte | - | >= | Greater than or equal |
lt | - | < | Less than |
lte | - | <= | Less than or equal |
like | - | LIKE | Pattern matching (case-sensitive) |
nlike | - | NOT LIKE | Negative pattern matching |
ilike | - | ILIKE | Pattern matching (case-insensitive, PostgreSQL) |
nilike | - | NOT ILIKE | Negative case-insensitive pattern |
in | inq | IN | Value in array |
nin | - | NOT IN | Value not in array |
between | - | BETWEEN | Value between two values |
is | - | IS NULL | Null check |
isn | - | IS NOT NULL | Not null check |
regexp | - | ~ | PostgreSQL POSIX regex (case-sensitive) |
iregexp | - | ~* | PostgreSQL POSIX regex (case-insensitive) |
Logical Operators
| Operator | Description |
|---|---|
and | Combine conditions with AND |
or | Combine conditions with OR |
Usage Examples
Simple equality:
await repo.find({ filter: { where: { status: 'active' } } });
// SQL: WHERE status = 'active'Comparison operators:
await repo.find({
filter: {
where: {
age: { gte: 18, lt: 65 },
score: { gt: 100 }
}
}
});
// SQL: WHERE age >= 18 AND age < 65 AND score > 100Array operators:
// IN operator
await repo.find({ filter: { where: { id: [1, 2, 3] } } });
// SQL: WHERE id IN (1, 2, 3)
// Using explicit IN
await repo.find({ filter: { where: { status: { in: ['active', 'pending'] } } } });
// NOT IN
await repo.find({ filter: { where: { status: { nin: ['deleted', 'archived'] } } } });Pattern matching:
// LIKE (case-sensitive)
await repo.find({ filter: { where: { name: { like: '%john%' } } } });
// ILIKE (case-insensitive, PostgreSQL)
await repo.find({ filter: { where: { email: { ilike: '%@gmail.com' } } } });Regex (PostgreSQL):
// Case-sensitive regex
await repo.find({ filter: { where: { name: { regexp: '^John' } } } });
// Case-insensitive regex
await repo.find({ filter: { where: { name: { iregexp: '^john' } } } });Between:
await repo.find({
filter: {
where: {
createdAt: { between: [new Date('2024-01-01'), new Date('2024-12-31')] }
}
}
});
// SQL: WHERE created_at BETWEEN '2024-01-01' AND '2024-12-31'Logical operators:
// OR conditions
await repo.find({
filter: {
where: {
or: [
{ status: 'active' },
{ isPublished: true }
]
}
}
});
// AND conditions (explicit)
await repo.find({
filter: {
where: {
and: [
{ role: 'admin' },
{ createdAt: { gte: new Date('2024-01-01') } }
]
}
}
});
// Nested conditions
await repo.find({
filter: {
where: {
status: 'active',
or: [
{ role: 'admin' },
{ and: [{ role: 'user' }, { verified: true }] }
]
}
}
});Security Notes
- Empty IN array: Returns
false(no rows), preventing security bypass - Empty NOT IN array: Returns
true(all rows match) - BETWEEN validation: Requires exactly 2 elements in array, throws error otherwise
- Invalid columns: Throws error if column doesn't exist in schema