How can I reload the table schema in sequel? - ruby

Given I have the following migration:
Sequel.migration do
up do
alter_table :users do
add_column :is_admin, :default => false
end
# Sequel runs a DESCRIBE table statement, when the model is loaded.
# At this point, it does not know that users have a is_admin flag.
# So it fails.
#user = User.find(:email => "admin#fancy-startup.example")
#user.is_admin = true
#user.save!
end
end
Then sequel does not automatically reload the table structure (see comment inline).
I am using this ugly hack to work around it:
# deep magic begins here. If you remove a single line, it will
# break the migration.
User.db.schema("users", :reload => true)
User.instance_variable_set(:#db_schema, nil)
User.columns
User.new.respond_to?(:is_admin=)
sleep 1
Is there a better way?

Much simpler than your hack is this hack: (re)set the dataset to the table name:
User.set_dataset :users
Seen in action:
require 'sequel'
DB = Sequel.sqlite
DB.create_table :users do
primary_key :id
String :name
end
class User < Sequel::Model; end
User << { name:"Bob" }
DB.alter_table :users do
add_column :is_admin, :boolean, default:false
end
p User.first #=> #<User #values={:id=>1, :name=>"Bob", :is_admin=>false}>
p User.setter_methods #=> ["name="]
User.set_dataset :users # Make the magic happen
p User.setter_methods #=> ["name=", "is_admin="]
#user = User.first
#user.is_admin = true
#user.save
p User.first #=> #<User #values={:id=>1, :name=>"Bob", :is_admin=>true}>
Note that there is no Sequel::Model#save! method; I changed it to save so that it would work.

Related

rom-rb form validation when using multiple relations

I'm trying out http://rom-rb.org/ and can't figure out how to get a presence validation to pass in the presence of multiple source models. I would expect the following script to save a new event and organiser, but instead it says that event_name is not present.
What am I missing?
require 'bundler/inline'
gemfile do
source 'https://rubygems.org'
gem 'rom'
gem 'rom-sql'
gem 'rom-rails'
gem 'activemodel'
gem 'sqlite3'
gem 'activesupport'
end
require 'rom'
require 'rom-rails'
`rm -Rf /tmp/romtest.sqlite`
ROM.setup(:sql, 'sqlite:///tmp/romtest.sqlite')
class Events < ROM::Relation[:sql]
end
class Organisers < ROM::Relation[:sql]
end
class CreateEvent < ROM::Commands::Create[:sql]
relation :events
register_as :create
result :one
associates :organiser, key: [:organiser_id, :id]
end
class CreateOrganiser < ROM::Commands::Create[:sql]
relation :organisers
register_as :create
result :one
end
class CreateEventWithOrganiser < ROM::Model::Form
commands organisers: :create, events: :create
input do
attribute :email
attribute :event_name
end
validations do
validates :event_name, presence: true
end
def commit!
command = organisers.create.with(
email: email,
) >> events.create.with(
name: event_name,
)
command.transaction do
command.call
end
end
end
ROM.finalize
rom = ROM.env
gateway = rom.gateways.fetch(:default)
migration = gateway.migration do
change do
create_table :organisers do
primary_key :id
column :email, String, null: false
end
create_table :events do
primary_key :id
column :name, String, null: false
column :organiser_id, Integer, null: false
end
end
end
migration.apply(gateway.connection, :up)
f = CreateEventWithOrganiser.build(
email: 'test#example.com',
event_name: 'Test Event'
)
# Unexpectedly fails
f.save
puts f.errors.full_messages
# => "Event name can't be blank"
Here's an updated version of your script which works:
require 'rom'
require 'rom-rails'
`rm -Rf /tmp/romtest.sqlite`
ROM.setup(:sql, 'sqlite:///tmp/romtest.sqlite')
class Events < ROM::Relation[:sql]
end
class Organisers < ROM::Relation[:sql]
end
class CreateEvent < ROM::Commands::Create[:sql]
relation :events
register_as :create
result :one
associates :organiser, key: [:organiser_id, :id]
end
class CreateOrganiser < ROM::Commands::Create[:sql]
relation :organisers
register_as :create
result :one
end
class CreateEventWithOrganiser < ROM::Model::Form
inject_commands_for :organisers, :events
input do
attribute :email
attribute :event_name
end
validations do
validates :event_name, presence: true
end
def commit!
validate!
return if errors.any?
command = organisers.create.with(
email: email
) >> events.create.with(
name: event_name
)
command.transaction do
command.call
end
end
end
ROM.finalize
rom = ROM.env
gateway = rom.gateways.fetch(:default)
migration = gateway.migration do
change do
create_table :organisers do
primary_key :id
column :email, String, null: false
end
create_table :events do
primary_key :id
column :name, String, null: false
column :organiser_id, Integer, null: false
end
end
end
migration.apply(gateway.connection, :up)
f = CreateEventWithOrganiser.build(
email: 'test#example.com',
event_name: 'Test Event'
)
puts f.save.result.inspect
# #<ROM::Commands::Result::Success:0x007fa92b589ea0 #value={:id=>1, :name=>"Test Event", :organiser_id=>1}>
The reason why it didn't work with commands is because this method will generate command objects for your form and set provided validations for each command, which will only work correctly if you used a single command. Otherwise same validator is used for each command which doesn't make sense. When you use inject_commands_for it will grab your own commands where validators are not set so you are free to handle validations yourself.
I think we should stop setting validators on commands which would make your original sample work but notice that you need to call validate! yourself.
I hope this helps.
I also created a gist showing how to do the same without a form: https://gist.github.com/solnic/3b68342482cf1414f719

Uninitialized constant (NameError) when using FactoryGirl in module

Here's the error I'm getting when I try to run my tests with RSpec:
C:/Ruby193/lib/ruby/gems/1.9.1/gems/activesupport-3.2.11/lib/active_support/infl
ector/methods.rb:230:in `block in constantize': uninitialized constant User (Nam
eError)
I'm trying to run FactoryGirl with RSpec but without Rails. Here are the files that take part in the testing:
user_spec.rb
require 'spec_helper'
module Bluereader
describe User do
describe 'login' do
user = FactoryGirl.build(:user)
end
describe 'logout' do
end
describe 'create_account' do
end
describe 'delete_account' do
end
end
end
spec/spec_helper
$LOAD_PATH.unshift(File.join(File.dirname(__FILE__), '..'))
$LOAD_PATH.unshift(File.dirname(__FILE__))
require 'rspec'
require 'lib/bluereader'
require 'factory_girl'
FactoryGirl.find_definitions
spec/factories.rb
require 'digest/sha1'
FactoryGirl.define do
sequence(:username) { |n| "user-#{n}" }
factory :user do
username
encrypted_password Digest::SHA1.hexdigest('password')
full_name 'John Doe'
logged_in_at Time.now
logged_out_at 0
end
end
At this point I know that the factories.rb file is being loaded (I tried with the moronic print-debugging). When I remove the user = FactoryGirl.build(:user) line from user_spec.rb I get no errors (and the normal RSpec feedback telling me there are no tests, but no errors). If you are interested, here's my model:
require 'digest/sha1'
module Bluereader
class User < ActiveRecord::Base
has_many :categories, :foreign_key => :user_id
has_many :news, :foreign_key => :user_id
has_many :settings, :foreign_key => :user_id
attr_reader :full_name
class << self
def login(username, password)
encrypted_password = Digest::SHA1.hexdigest(password)
if not User.exists?(:username => username, :encrypted_password => encrypted_password)
user_id = User.id_from_username(username)
update(user_id, :logged_in_at => Time.now, :logged_out_at => 0)
end
end
def logout
update(current_user.id, :logged_out_at => Time.now)
end
def validate_account(username, password, full_name)
if username.empty? or password.empty or full_name.empty?
return 'Please fill in all the fields.'
end
if User.exists?(:username => username)
return 'That username is already in use.'
end
unless username =~ /^\w+$/
return 'Username field should contain only letters, numbers and underscores.'
end
''
end
def create_account(username, password, full_name)
encrypted_password = Digest::SHA1.hexdigest(password)
User.create(:username => username,
:encrypted_password => encrypted_password,
:full_name => full_name,
:logged_in_at => Time.now,
:logged_out_at => 0)
end
def delete_account
current_user.destroy
end
private
def id_from_username(username)
user = where(:username => username).first
user.nil? ? 0 : user.id
end
def current_user
where(:logged_out_at => 0).first
end
end
end
end
SOLUTION
The problem was that the class User was in a module, here's the solution:
factory :user, class: Bluereader::User do
You need to require the rails environment in your spec helper file. Add the following to spec/spec_helper.rb:
require File.expand_path("../../config/environment", __FILE__)
Update
Even if you're not using Rails, you'll still need to require the models in your spec helper.
Taken from the bottom of the question
The problem was that the class User was in a module, here's the solution:
factory :user, class: Bluereader::User do
For anyone clumsy like me, you may have FactoryGirl in your code where you meant to have FactoryBot

ActiveRecord::Migration - Referencing a table in another schema

I'm using Ruby and PostgreSQL and have created 3 distinct DB schemas: billing (for billing related data), customer (for customer related data) and edocs (for electronic documents related data).
I'm not using Rails so I have a stand-alone migration code like this:
#migrate.rb
if ARGV[0] =~ /VERSION=\d+/
version = ARGV[0].split('=')[1].to_i
else
version = nil
end
ActiveRecord::Base.default_timezone = :utc
#logger = Logger.new $stderr
ActiveRecord::Base.logger = #logger
ActiveSupport::LogSubscriber.colorize_logging = false
#config = YAML.load_file(File.join(File.dirname(__FILE__), 'database.yml'))
ActiveRecord::Base.establish_connection(#config["edocs"])
ActiveRecord::Migrator.migrate(".", version)
I have already realized that I probably have to create a different directories to contain the migration for the different schemas, and changing connection info for each migrate.rb.
But I'm not sure how I'm going to make a table reference another table that is in another schema.
For example:
class CreateBillingEventsTable < ActiveRecord::Migration
def self.up
create_table :billing_events do |t|
t.references :customer, :null => false
t.references :service_type, :null => false
t.timestamps
end
change_table :billing_events do |t|
t.index [:customer_id, :created_at]
end
end
def self.down
remove_index :billing_events, :column => [:customer_id, :created_at]
drop_table :billing_events
end
end
In the above example, "customer" and "billing_events" are in different schemas.
How can I code that?
Thanks.

How does the :save hook relate to the :update hook in DataMapper?

If I define the following model...
class Foo
include DataMapper::Resource
property :name, String, :key => true
before :save, do
puts 'save'
end
before :update, do
puts 'update'
end
end
Why does the 2nd save also trigger the 'update' hook?
ruby :001 > f = Foo.new
=> #<Foo #name=nil>
ruby :002 > f.name = 'Bob'
=> "Bob"
ruby :003 > f.save
save
=> true
ruby :004 > f.name = 'Joe'
=> "Joe"
ruby :005 > f.save
save
update
=> true
Of course I could dive into the source and answer the question of what code drives this behavior. More importantly, I want to understand the right way to use each of these hooks in practice.
require 'rubygems'
require 'data_mapper'
class Foo
include DataMapper::Resource
property :name, String, :key => true
before :create, do
puts 'Create: Only happens when saving a new object.'
end
before :update, do
puts 'Update: Only happens when saving an existing object.'
end
before :save, do
puts 'Save: Happens when either creating or updating an object.'
end
before :destroy, do
puts 'Destroy: Only happens when destroying an existing object.'
end
end
DataMapper.setup :default, 'sqlite::memory:'
DataMapper.finalize
DataMapper.auto_migrate!
puts "New Foo:"
f = Foo.new :name => "Fighter"
f.save
puts "\nUpdate Foo:"
f.name = "Bar"
f.save
puts "\nUpdate Foo again:"
f.update :name => "Baz"
puts "\nDestroy Foo:"
f.destroy
Which returns:
New Foo:
Save: Happens when either creating or updating an object.
Create: Only happens when saving a new object.
Update Foo:
Save: Happens when either creating or updating an object.
Update: Only happens when saving an existing object.
Update Foo again:
Save: Happens when either creating or updating an object.
Update: Only happens when saving an existing object.
Destroy Foo:
Destroy: Only happens when destroying an existing object.
So as you can see you'd want to use :save hooks whenever something should happen after either a create or an update, and :create and/or :update when you want a finer level of control.

Rails audit system with both ActiveResource and ActiveRecord

I have a huge project with both of ActiveRecord and ActiveResource models. I need to implement logging of user activity with these models and also to log changes of model attributes (save object state or somthing like that). Changes can made by users or cron rake tasks.
I also must have possibility to search any data by date , any field ..etc
Will be nice also to generate readable messages with last activity , for example
User Bob change his password to * and email to ** at 2011-08-12 08:12
Staff Jeff added new partner: Company name at 2011-08-12 08:13
Admin Jack deleted product : Product name at 2011-09-12 11:11
Client Sam ordered new service : Service name at 2011-09-12 11:12
Does anybody implement such logging? Ideas? Advices?
should I use gems or can I do all the logic with observers not changing models?
I liked gem https://github.com/airblade/paper_trail can anybody say how can I make it work with activeresource ?
You are looking for
https://github.com/collectiveidea/acts_as_audited
Few open source projects use that plugin I think Red Mine as well as The Foreman.
Edit: Unfortunately it can do only ActiveRecord, not ActiveResource.
Fivell, I just saw this question and don't have time to work up alterations this evening before the bounty expires, so I'll give you my auditing code that works with ActiveRecord and should work with ActiveResource, perhaps with a few tweaks (I don't use ARes often enough to know offhand). I know the callbacks we use are there, but I'm not sure if ARes has ActiveRecord's dirty attribute changes tracking.
This code logs each CREATE/UPDATE/DELETE on all models (excepting CREATEs on the audit log model and any other exceptions you specify) with the changes stored as JSON. A cleaned backtrace is also stored so you can determine what code made the change (this captures any point in your MVC as well as rake tasks and console usage).
This code works for console usage, rake tasks, and http requests, although generally only the last one logs the current user. (If I recall correctly, the ActiveRecord observer that this replaced did not work in rake tasks or the console.) Oh, this code comes from a Rails 2.3 app - I have a couple Rails 3 apps, but I haven't needed this kind of auditing for them yet.
I don't have code that builds a nice display of this information (we only dig into the data when we need to look into an issue), but since the changes are stored as JSON it should be fairly straightforward.
First, we store the current user in User.current so it is accessible everywhere, so in app/models/user.rb:
Class User < ActiveRecord::Base
cattr_accessor :current
...
end
The current user is set in the application controller for each request like so (and does not cause concurrency issues):
def current_user
User.current = session[:user_id] ? User.find_by_id(session[:user_id]) : nil
end
You could set User.current in your rake tasks if it made sense to.
Next, we define the model to store the audit info app/models/audit_log_entry.rb - you'll want to customize IgnoreClassesRegEx to fit any models you don't want audited:
# == Schema Information
#
# Table name: audit_log_entries
#
# id :integer not null, primary key
# class_name :string(255)
# entity_id :integer
# user_id :integer
# action :string(255)
# data :text
# call_chain :text
# created_at :datetime
# updated_at :datetime
#
class AuditLogEntry < ActiveRecord::Base
IgnoreClassesRegEx = /^ActiveRecord::Acts::Versioned|ActiveRecord.*::Session|Session|Sequence|SchemaMigration|CronRun|CronRunMessage|FontMetric$/
belongs_to :user
def entity (reload = false)
#entity = nil if reload
begin
#entity ||= Kernel.const_get(class_name).find_by_id(entity_id)
rescue
nil
end
end
def call_chain
return if call_chain_before_type_cast.blank?
if call_chain_before_type_cast.instance_of?(Array)
call_chain_before_type_cast
else
JSON.parse(call_chain_before_type_cast)
end
end
def data
return if data_before_type_cast.blank?
if data_before_type_cast.instance_of?(Hash)
data_before_type_cast
else
JSON.parse(data_before_type_cast)
end
end
def self.debug_entity(class_name, entity_id)
require 'fastercsv'
FasterCSV.generate do |csv|
csv << %w[class_name entity_id date action first_name last_name data]
find_all_by_class_name_and_entity_id(class_name, entity_id,
:order => 'created_at').each do |a|
csv << [a.class_name, a.entity_id, a.created_at, a.action,
(a.user && a.user.first_name), (a.user && a.user.last_name), a.data]
end
end
end
end
Next we add some methods to ActiveRecord::Base to make the audits work. You'll want to look at the audit_log_clean_backtrace method and modify for your needs. (FWIW, we put additions to existing classes in lib/extensions/*.rb which are loaded in an initializer.) In lib/extensions/active_record.rb:
class ActiveRecord::Base
cattr_accessor :audit_log_backtrace_cleaner
after_create :audit_log_on_create
before_update :save_audit_log_update_diff
after_update :audit_log_on_update
after_destroy :audit_log_on_destroy
def audit_log_on_create
return if self.class.name =~ /AuditLogEntry/
return if self.class.name =~ AuditLogEntry::IgnoreClassesRegEx
audit_log_create 'CREATE', self, caller
end
def save_audit_log_update_diff
#audit_log_update_diff = changes.reject{ |k,v| 'updated_at' == k }
end
def audit_log_on_update
return if self.class.name =~ AuditLogEntry::IgnoreClassesRegEx
return if #audit_log_update_diff.empty?
audit_log_create 'UPDATE', #audit_log_update_diff, caller
end
def audit_log_on_destroy
return if self.class.name =~ AuditLogEntry::IgnoreClassesRegEx
audit_log_create 'DESTROY', self, caller
end
def audit_log_create (action, data, call_chain)
AuditLogEntry.create :user => User.current,
:action => action,
:class_name => self.class.name,
:entity_id => id,
:data => data.to_json,
:call_chain => audit_log_clean_backtrace(call_chain).to_json
end
def audit_log_clean_backtrace (backtrace)
if !ActiveRecord::Base.audit_log_backtrace_cleaner
ActiveRecord::Base.audit_log_backtrace_cleaner = ActiveSupport::BacktraceCleaner.new
ActiveRecord::Base.audit_log_backtrace_cleaner.add_silencer { |line| line =~ /\/lib\/rake\.rb/ }
ActiveRecord::Base.audit_log_backtrace_cleaner.add_silencer { |line| line =~ /\/bin\/rake/ }
ActiveRecord::Base.audit_log_backtrace_cleaner.add_silencer { |line| line =~ /\/lib\/(action_controller|active_(support|record)|hoptoad_notifier|phusion_passenger|rack|ruby|sass)\// }
ActiveRecord::Base.audit_log_backtrace_cleaner.add_filter { |line| line.gsub(RAILS_ROOT, '') }
end
ActiveRecord::Base.audit_log_backtrace_cleaner.clean backtrace
end
end
Finally, here are the tests we have on this - you'll need to modify the actual test actions of course. test/integration/audit_log_test.rb
require File.dirname(__FILE__) + '/../test_helper'
class AuditLogTest < ActionController::IntegrationTest
def setup
end
def test_audit_log
u = users(:manager)
log_in u
a = Alert.first :order => 'id DESC'
visit 'alerts/new'
fill_in 'alert_note'
click_button 'Send Alert'
a = Alert.first :order => 'id DESC', :conditions => ['id > ?', a ? a.id : 0]
ale = AuditLogEntry.first :conditions => {:class_name => 'Alert', :entity_id => a.id }
assert_equal 'Alert', ale.class_name
assert_equal 'CREATE', ale.action
end
private
def log_in (user, password = 'test', initial_url = home_path)
visit initial_url
assert_contain 'I forgot my password'
fill_in 'email', :with => user.email
fill_in 'password', :with => password
click_button 'Log In'
end
def log_out
visit logout_path
assert_contain 'I forgot my password'
end
end
And test/unit/audit_log_entry_test.rb:
# == Schema Information
#
# Table name: audit_log_entries
#
# id :integer not null, primary key
# class_name :string(255)
# action :string(255)
# data :text
# user_id :integer
# created_at :datetime
# updated_at :datetime
# entity_id :integer
# call_chain :text
#
require File.dirname(__FILE__) + '/../test_helper'
class AuditLogEntryTest < ActiveSupport::TestCase
test 'should handle create update and delete' do
record = Alert.new :note => 'Test Alert'
assert_difference 'Alert.count' do
assert_difference 'AuditLogEntry.count' do
record.save
ale = AuditLogEntry.first :order => 'created_at DESC'
assert ale
assert_equal 'CREATE', ale.action, 'AuditLogEntry.action should be CREATE'
assert_equal record.class.name, ale.class_name, 'AuditLogEntry.class_name should match record.class.name'
assert_equal record.id, ale.entity_id, 'AuditLogEntry.entity_id should match record.id'
end
end
assert_difference 'AuditLogEntry.count' do
record.update_attribute 'note', 'Test Update'
ale = AuditLogEntry.first :order => 'created_at DESC'
expected_data = {'note' => ['Test Alert', 'Test Update']}
assert ale
assert_equal 'UPDATE', ale.action, 'AuditLogEntry.action should be UPDATE'
assert_equal expected_data, ale.data
assert_equal record.class.name, ale.class_name, 'AuditLogEntry.class_name should match record.class.name'
assert_equal record.id, ale.entity_id, 'AuditLogEntry.entity_id should match record.id'
end
assert_difference 'AuditLogEntry.count' do
record.destroy
ale = AuditLogEntry.first :order => 'created_at DESC'
assert ale
assert_equal 'DESTROY', ale.action, 'AuditLogEntry.action should be CREATE'
assert_equal record.class.name, ale.class_name, 'AuditLogEntry.class_name should match record.class.name'
assert_equal record.id, ale.entity_id, 'AuditLogEntry.entity_id should match record.id'
assert_nil Alert.find_by_id(record.id), 'Alert should be deleted'
end
end
test 'should not log AuditLogEntry create entry and block on update and delete' do
record = Alert.new :note => 'Test Alert'
assert_difference 'Alert.count' do
assert_difference 'AuditLogEntry.count' do
record.save
end
end
ale = AuditLogEntry.first :order => 'created_at DESC'
assert_equal 'CREATE', ale.action, 'AuditLogEntry.action should be CREATE'
assert_equal record.class.name, ale.class_name, 'AuditLogEntry.class_name should match record.class.name'
assert_equal record.id, ale.entity_id, 'AuditLogEntry.entity_id should match record.id'
assert_nil AuditLogEntry.first(:conditions => { :class_name => 'AuditLogEntry', :entity_id => ale.id })
if ale.user_id.nil?
u = User.first
else
u = User.first :conditions => ['id != ?', ale.user_id]
end
ale.user_id = u.id
assert !ale.save
assert !ale.destroy
end
end
https://github.com/collectiveidea/acts_as_audited
and
https://github.com/airblade/paper_trail
are both great solutions for ActiveRecord only, but since much of ActiveRecord has been extracted to ActiveModel, it's likely to be reasonable to extend either to support ActiveResource as well, at least for read-only support. I looked through the Github network graphs and googled around and there doesn't appear to be any ongoing development of such a solution, nevertheless I expect it will be easier to implement on top of one of these two plugins than starting from scratch. paper_trail appears to be under more active development and has some commits for Rails 3.1, so it may be more up to date with Rails internals and easier to extend, but that's just a gut instinct—I'm not familiar with the internals of either.
The acts_as_audited gem should work well for you:
https://github.com/collectiveidea/acts_as_audited
And as far as ActiveResource is considered, it will also be a model in some other application. You can use the gem at the server side, and you don't need to audit it at the client side. All the CRUD operations using ActiveResource would finally translate to CRUD operations on the ActiveRecord (on server side).
So probably you need to look at it from a distance, and the same solution would apply in both the cases, but at different places.
for tracking user activity(CRUD ), i've created a class inherits from Logger, and now I am planing to write a litle plugin for tracking user that i can use for any ROR application built. I have already checked if there is a plugin like that but I didn’t see. I guess there are many gem like paper-trail, acts_as_audited or itslog but i prefer to use a plugin. Any suggestions?
Here is a link that might help you : http://robaldred.co.uk/2009/01/custom-log-files-for-your-ruby-on-rails-applications/comment-page-1/#comment-342
nice coding

Resources