class

Lustra::Reflection::Column

Inherits Lustra::Model < Lustra::Model::FullTextSearchable < Lustra::Model::JSONDeserialize < Lustra::Model::Initializer < Lustra::Model::HasFactory < Lustra::Model::Introspection < Lustra::Model::ClassMethods < Lustra::Model::HasScope < Lustra::Model::HasRelations < Lustra::Model::HasValidation < Lustra::Validation::Helper < Lustra::Model::HasSaving < Lustra::Model::HasSerialPkey < Lustra::Model::HasTimestamps < Lustra::Model::HasColumns < Lustra::Model::HasHooks < Lustra::Model::Connection < Lustra::ErrorMessages < Reference < Object

Reflection of the columns using information_schema in postgreSQL. TODO: Usage of view instead of model

Constants

COLUMNS = {"table_catalog" => {type: String, primary: false, converter: "String", db_column_name: "table_catalog", crystal_variable_name: table_catalog, presence: true, mass_assign: true}, "table_schema" => {type: String, primary: false, converter: "String", db_column_name: "table_schema", crystal_variable_name: table_schema, presence: true, mass_assign: true}, "table_name" => {type: ::Union(String, ::Nil), primary: false, converter: "String", db_column_name: "table_name", crystal_variable_name: table_name, presence: true, mass_assign: true}, "column_name" => {type: String, primary: true, converter: "String", db_column_name: "column_name", crystal_variable_name: column_name, presence: true, mass_assign: true}} of Nil => Nil

POLYMORPHISM_SETTINGS = {} of Nil => Nil

Constructors

build(x : NamedTuple) : self

Build a new empty model and fill the columns using the NamedTuple in argument.

Returns the new model

build(x : NamedTuple, &block : self -> Nil) : self

Build a new empty model and fill the columns using the NamedTuple in argument.

Returns the new model

create(x : NamedTuple, &block : self -> Nil) : self

Build and new model and save it. Returns the model.

The model may not be saved due to validation failure; check the returned model errors? and persisted? flags.

create

Build and new model and save it. Returns the model.

The model may not be saved due to validation failure; check the returned model errors? and persisted? flags.

create(x : NamedTuple) : self

Build and new model and save it. Returns the model.

The model may not be saved due to validation failure; check the returned model errors? and persisted? flags.

create

Build and new model and save it. Returns the model.

The model may not be saved due to validation failure; check the returned model errors? and persisted? flags.

create!(x : NamedTuple, &block : self -> Nil) : self

Build and new model and save it. Returns the model.

Returns the newly inserted model Raises an exception if validation failed during the saving process.

create!

Build and new model and save it. Returns the model.

Returns the newly inserted model Raises an exception if validation failed during the saving process.

create!(x : NamedTuple) : self

Build and new model and save it. Returns the model.

Returns the newly inserted model Raises an exception if validation failed during the saving process.

create!

Build and new model and save it. Returns the model.

Returns the newly inserted model Raises an exception if validation failed during the saving process.

new(h : Hash(String, _), cache : Lustra::Model::QueryCache | Nil = nil, persisted = false, fetch_columns = false)

new(json : JSON::Any, cache : Lustra::Model::QueryCache | Nil = nil, persisted = false)

new(t : NamedTuple, persisted = false)

new

Class methods

build

Build a new empty model and fill the columns using the NamedTuple in argument.

Returns the new model

build

Build a new empty model and fill the columns using the NamedTuple in argument.

Returns the new model

build

Build a new empty model and fill the columns using the NamedTuple in argument.

Returns the new model

columns

connection

Define on which connection the model is living. Useful in case of models living in different databases.

Is set to "default" by default.

See Lustra::SQL#init(URI, *opts) for more information about multi-connections.

Example:

Lustra::SQL.init("postgres://postgres@localhost/database_1")
Lustra::SQL.init("secondary", "postgres://postgres@localhost/database_2")

class ModelA
  include Lustra::Model

  # Performs all the queries on `database_1`
  # self.connection = "default"
  column id : Int32, primary: true, presence: false
  column title : String
end

class ModelB
  include Lustra::Model

  # Performs all the queries on `database_2`
  self.connection = "secondary"

  column id : Int32, primary: true, presence: false
end

connection=(connection : String)

Define on which connection the model is living. Useful in case of models living in different databases.

Is set to "default" by default.

See Lustra::SQL#init(URI, *opts) for more information about multi-connections.

Example:

Lustra::SQL.init("postgres://postgres@localhost/database_1")
Lustra::SQL.init("secondary", "postgres://postgres@localhost/database_2")

class ModelA
  include Lustra::Model

  # Performs all the queries on `database_1`
  # self.connection = "default"
  column id : Int32, primary: true, presence: false
  column title : String
end

class ModelB
  include Lustra::Model

  # Performs all the queries on `database_2`
  self.connection = "secondary"

  column id : Int32, primary: true, presence: false
end

create_from_json(string_or_io : String | IO, trusted : Bool = false)

Create a new model from json and save it. Returns the model.

The model may not be saved due to validation failure; check the returned model errors? and persisted? flags. Trusted flag set to true will allow mass assignment without protection, FALSE by default

create_from_json!(string_or_io : String | IO, trusted : Bool = false)

Create a new model from json and save it. Returns the model.

Returns the newly inserted model Raises an exception if validation failed during the saving process. Trusted flag set to true will allow mass assignment without protection, FALSE by default

find(ids : Array)

Find multiple models by an array of primary keys. Returns an array of models (may be empty if none found).

users = User.find([1, 2, 3])
users.size # => 0..3 depending on how many were found

find(x)

Returns a model using primary key equality Returns nil if not found.

find!(ids : Array)

Find multiple models by an array of primary keys. Raises error if ANY of the IDs are not found.

users = User.find!([1, 2, 3]) # Raises if any ID is not found

find!(x)

Returns a model using primary key equality. Raises error if the model is not found.

find_by(tuple : NamedTuple)

Find a model by column values. Returns nil if not found. This is an alias for query.find(**tuple) with better naming.

user = User.find_by(email: "test@example.com")
user = User.find_by(first_name: "John", last_name: "Doe")

find_by

Find a model by column values. Returns nil if not found. This is an alias for query.find(**tuple) with better naming.

user = User.find_by(email: "test@example.com")
user = User.find_by(first_name: "John", last_name: "Doe")

find_by!(tuple : NamedTuple)

Find a model by column values. Raises error if not found. This is an alias for query.find!(**tuple) with better naming.

user = User.find_by!(email: "test@example.com")

find_by!

Find a model by column values. Raises error if not found. This is an alias for query.find!(**tuple) with better naming.

user = User.find_by!(email: "test@example.com")

from_json(string_or_io : String | IO, trusted : Bool = false)

Create a new empty model and fill the columns from json. Returns the new model

Trusted flag set to true will allow mass assignment without protection, FALSE by default

full_table_name

returns the fully qualified and escaped name for this table. add schema if schema is different from 'public' (default schema)

ex: "schema"."table"

import(array : Enumerable(self), on_conflict : Lustra::SQL::InsertQuery -> | Nil = nil)

Import a bulk of models in one SQL insert query. Each model must be non-persisted.

on_conflict callback can be optionnaly turned on to manage constraints of the database.

Note: Old models are not modified. This method return a copy of the models as saved in the database.

Example:

users = [User.new(id: 1), User.new(id: 2), User.new(id: 3)]
users = User.import(users)

none

Return an empty, chainable collection (Rails-like .none). Useful for conditional branches where no records should be returned while keeping query chaining intact.

User.none.where { active == true }.count # => 0

polymorphic?

query

Return a new query SELECT * FROM [my_model_table]. Can be refined after that. Automatically applies default_scope if defined.

read_only=(read_only : Bool)

read_only?

register_counter_cache(association_model : Class, counter_column : String, foreign_key : String)

Register a counter cache for this model using model class

reset_counters(id, *counter_models)

Reset counter cache columns to their correct values. This is useful when counter caches become out of sync due to direct SQL operations.

Example:

User.reset_counters(user.id, Post)
User.reset_counters(user.id, Post, Comment)

schema

Define the current schema used in PostgreSQL. The value is nil by default, which lead to non-specified schema during the querying, and usage of "public" by PostgreSQL.

This property can be redefined on initialization. Example:

class MyModel
  include Lustra::Model

  self.schema = "my_schema"
end
MyModel.query.to_sql # SELECT * FROM "my_schema"."my_models"

schema=(schema : Lustra::SQL::Symbolic | Nil)

Define the current schema used in PostgreSQL. The value is nil by default, which lead to non-specified schema during the querying, and usage of "public" by PostgreSQL.

This property can be redefined on initialization. Example:

class MyModel
  include Lustra::Model

  self.schema = "my_schema"
end
MyModel.query.to_sql # SELECT * FROM "my_schema"."my_models"

schema_description

Return the table description without printing.

table

Return the table name setup for this model. By convention, the class name is by default equals to the pluralized underscored string form of the model name. Example:

MyModel => "my_models"
Person => "people"
Project::Info => "project_infos"

The property can be updated at initialization to a custom table name:

class MyModel
  include Lustra::Model

  self.table = "another_table_name"
end
MyModel.query.to_sql # SELECT * FROM "another_table_name"

table=(table : Lustra::SQL::Symbolic)

Return the table name setup for this model. By convention, the class name is by default equals to the pluralized underscored string form of the model name. Example:

MyModel => "my_models"
Person => "people"
Project::Info => "project_infos"

The property can be updated at initialization to a custom table name:

class MyModel
  include Lustra::Model

  self.table = "another_table_name"
end
MyModel.query.to_sql # SELECT * FROM "another_table_name"

Instance methods

_cached_table

add_built_association(association_name : String, model : Lustra::Model)

Add a built association to track for autosave

attributes

Attributes, used when fetch_columns is true

built_associations

Track built associations for autosave functionality

built_associations=(built_associations : Hash(String, Array(Lustra::Model)))

Track built associations for autosave functionality

cache

changed

Returns an array of names of all changed attributes.

user.email = "new@test.com"
user.first_name = "John"
user.changed  # => ["email", "first_name"]

changed?

Return true if the model is dirty (e.g. one or more fields have been changed.). Return false otherwise.

changes

Returns a hash of all changed attributes with their [old_value, new_value].

user.email = "new@test.com"
user.first_name = "John"
user.changes  # => {"email" => ["old@test.com", "new@test.com"], "first_name" => [nil, "John"]}

clear_built_associations

Clear all built associations (called after successful save)

clear_change_flags

Reset the changed? flag on all columns

The model behave like its not dirty anymore and call to save would apply no changes.

Returns self

column_name

Returns the value of column_name column or throw an exception if the column is not defined.

column_name=(x : String)

Setter for column_name column.

column_name_column

Returns the column object used to manage column_name field

See Lustra::Model::Column

has_built_associations?

Check if there are any pending built associations

reset(h : Hash(Symbol, _))

Set the columns from hash

reset(h : Hash(String, _))

Set the model fields from hash

reset(t : NamedTuple)

reset(from_json : JSON::Any)

reset

reset flavors

reset_counters(*counter_models)

Reset counter cache columns

Example:

user = User.find(1)
user.reset_counters(Post)
user.reset_counters(Post, Comment)

set(h : Hash(Symbol, _))

Set the columns from hash

set(h : Hash(String, _))

Set the model fields from hash

set(t : NamedTuple)

set(from_json : JSON::Any)

set

Set one or multiple columns to a specific value This two are equivalents:

model.set(a: 1)
model.a = 1

set_from_json(string_or_io : String | IO, trusted : Bool = false)

Set the fields from json passed as argument Trusted flag set to true will allow mass assignment without protection, FALSE by default

table

The method table is a belongs_to relation to Lustra::Reflection::Table

table!

table=(model : Lustra::Reflection::Table | Nil)

table_catalog

Returns the value of table_catalog column or throw an exception if the column is not defined.

table_catalog=(x : String)

Setter for table_catalog column.

table_catalog_column

Returns the column object used to manage table_catalog field

See Lustra::Model::Column

table_name

Returns the value of table_name column or throw an exception if the column is not defined.

table_name=(x : Union(String, Nil))

Setter for table_name column.

table_name_column

Returns the column object used to manage table_name field

See Lustra::Model::Column

table_schema

Returns the value of table_schema column or throw an exception if the column is not defined.

table_schema=(x : String)

Setter for table_schema column.

table_schema_column

Returns the column object used to manage table_schema field

See Lustra::Model::Column

to_h(full = false) : Hash(String, Lustra::SQL::Any)

Return a hash version of the columns of this model.

to_json(emit_nulls : Bool = false)

to_json(json, emit_nulls = false)

update_from_json(string_or_io : String | IO, trusted : Bool = false)

Set the fields from json passed as argument and call save on the object Trusted flag set to true will allow mass assignment without protection, FALSE by default

update_from_json!(string_or_io : String | IO, trusted : Bool = false)

Set the fields from json passed as argument and call save! on the object Trusted flag set to true will allow mass assignment without protection, FALSE by default

update_h

Generate the hash for update request (like during save)

validate_fields_presence

For each column, ensure than when needed the column has present information into it.

This method is called on validation.

Nested types

Lustra::Reflection::Column::Collection