| layout |
|---|
default |
DateTimePicker can be used to display and select date and time.
{% highlight haml %} .span3= f.input :created_at, as: :string, input_html: {value: @student.created_at? ? @student.created_at.to_s(:hk) : Time.now, class: 'datetimepicker'} {% endhighlight%}
{% highlight javascript %} $('.datetimepicker').datetimepicker({ format: 'YYYY/MM/DD HH:mm', sideBySide: true, inline: true, locale: 'zh-tw' }); {% endhighlight%}
This 'datetimepicker' gem is designed for bootstrap3. It raises javascript error Uncaught Error: datetimepicker component should be placed within a relative positioned container. To solve this issue, define one of its parent elements as a relative positioned container. For example:
{% highlight haml %} .span3{style: 'position: relative'} = f.input :created_at, as: :string, input_html: {value: @student.created_at? ? @student.created_at.to_s(:localdb) : Time.now, class: 'datetimepicker'} {% endhighlight%}
After deployed to staging server, the project raised an error when uploading file using paperclip: Errno::ENOMEM (Cannot allocate memory - file -b --mime '/tmp/d5243f00b80eed0af24ebac9046b66af20160429-30082-ynxdk2.pdf'). This error has never occured in my development environment. The reason is, paperclip uses ImageMagick for image processing, gem cocaine is the way it invokes ImageMagick. It needs to use fork in linux system, which means if the parent process uses 500 MB memory, the child process will use exact the same size. This is widely used in Rails apps. It makes Rails cost more memory. To solve this problem, we can try using gem 'posix-spawn'. In my project, the error raises again after using this gem. We ended up updating memory of server.
ref: blog
sidekiq is a gem that can run jobs in background. It can be deployed using gem 'capistrano-sidekiq'. In our project, sidekiq is not started on server after each deploy. The reason is to be determined. One way to start sidekiq on server manually is:
{% highlight shell %} bundle exec sidekiq -e staging -d -L log/sidekiq.log -C config/sidekiq.yml {% endhighlight%}
Meaning of parameters:
-d, Daemonize process
-L, path to writable logfile
-C, path to YAML config file
-e, Application environment
My task is to craw the page content of a Facebook page of a closed group. The input is the url of the group. After some research, this is what I have found.
First of all, to access such a page in browser, we need to login with the Facebook username and password. Normal crawler gems like Nokogiri do not support login well. The best way to log into Facebook is to use Facebook Graph API. However, using Facebook API means we have to follow Facebook rules and cannot get information from a closed group. The solution to this task is to use phantomjs to simulate a browser programmatically. Here I choose gem 'poltergeist' and gem 'capybara'. Below is the code:
Create a file 'setup-capybara.rb': {% highlight ruby %}
require 'capybara/poltergeist'
See more options at https://github.com/teampoltergeist/poltergeist#customization
Capybara.register_driver :poltergeist do |app| Capybara::Poltergeist::Driver.new(app, js_errors: false, timeout: 120) end
Capybara.default_driver = :poltergeist {% endhighlight%}
Create another file 'test.rb': {% highlight ruby %} require './setup-capybara.rb'
url = 'https://www.facebook.com/groups/1468981113348561/' browser = Capybara.current_session browser.visit url
browser.fill_in 'email', with: '[email protected]' browser.fill_in 'pass', with: '1234' browser.click_on 'loginbutton'
img = browser.find('.coverPhotoImg') p img['src'] number = browser.find_by_id('count_text') p number.text {% endhighlight%}
This method is used to test website, so it can perfectly simulate a browser and all the actions like fill in form and click button. It can also get html elements by id, class or other selectors. A great online example is here
Rails flash message is delete after redirect not render, so sometimes it is still shown after another redirect.
To use Rails console like development environment, simply run bundle exec rails c staging. Note that this command needs to state which environment to run, like "staging", "production".
Sidekiq will by default retry a job if it fails. To change this setting, we need to run rails console:
{% highlight ruby %}
require 'sidekiq/api'
Sidekiq::RetrySet.new.size # check the size of retry queue
Sidekiq::RetrySet.new.clear # clear the retry queue
Sidekiq::Queue.new.size # check the size of jobs queue
Sidekiq::Queeu.new.clear # clear the jobs queue
{% endhighlight%}
Access the remote server using SSH, then access mysql using command like mysql -u root.
In mysql:
-
show databases;to list all the databases. -
use mydb_name;to change current database. -
show tables;to list all the tables in the current database. -
drop database mydb_name;to drop a database. -
create database mydb_name;to create a database. -
source /path_to/mydump.sqlto import the sql file into current database. -
ALTER DATABASE mydb_name CHARACTER SET utf8 COLLATE utf8_general_cibefore import data to avoid the errorMysql2::Error: Illegal mix of collations (latin1_swedish_ci,IMPLICIT) and (utf8_general_ci,COERCIBLE)
ssh [email protected] to access the remote server.
/etc/init.d/my_project stop
/etc/init.d/my_project start
Session is only available in controllers and views, cannot be used in models.
session[:job_params].deep_merge!(job_params) if job_params merge the hash values to session.
A fundamental tutorial can be found here. This tutorial uses session to store the current step and the form values to pass through different steps.
It can also be done by using gems.
To display Facebook page in a iframe on web, simply go to Facebook developer website and get the embedded iframe code from Facebook.
Google maps can be loaded into web in many ways. It now requires an API key to use geocoder and other services.