road to deployment

[hi-ogawa]

• angular: put design which is good enouth to use
  • where to deploy ?
  • not separate with rails, put angularjs in (rails root)/public like this http://blog.benmorgan.io/post/77446075979/setting-up-an-angularjs-app-with-a-rails-api
  • to get cross domain request nicely done.
• chrome:
  • change all the url of api requests to production
• rails:
  • amazon ec2
  • (preferred) database on amazon
  • (preferred) paperclip on amazon s3
• copy my previous data into new server

[hi-ogawa]

this is what I'm gonna do.

• replace the place for paperclip images into cloud (http://www.rubydoc.info/gems/paperclip/Paperclip/Storage/S3)

  • here, I need to change angularjs code to be able to load picture

• change database to mysql (this is my choice) or postgresql

  • go with mysql (default value for text type is not working as it's different from sqlite.)

• change dist generated by grunt to (rails root)/public

  • needed to add directive html sources into ngtemplate src config.

• put rails with production oprion then test locally (secret_key stuff?)

  • just gem 'passenger'
  • before starting server, export SECRET_KEY_BASE=$(rake secret)

• clone railsAPI into ec2

  • (gave up) setup ruby on rails via docker?

• change all request url in chrome extention to the production server domail url

  • prepare one js file for share config across each part of chrome ext, then only thing I do is just edit the file.

• put rails on ec2 (gem passenger, rails_env=production), some tweak on domail name or anything

  • amazon ec2, how to set domain

• test everything

• copy previous data

  • in rails console, request to the api

• drop previous version

  • write down the ideas about design implemented in the previous version
  • chrome ext. client
    • forget it.
  • rails view web client
    • leave the previous page open only with my account so that I can check anytime

• replace master branch with backbone branch

  • drop master branch (keep only omniauth_test branch)
  • http://stackoverflow.com/questions/2862590/how-to-replace-master-branch-in-git-entirely-from-another-branch
  • sakura server
  • don't have to rush to delete sakura account
• done!

[hi-ogawa]

copy previous data by running $ rake run_rb[move_data.rb] on rails root directory.

# lib/tasks/mytask.rake
desc "run .rb file in rails console"
task :run_rb, [:filename] do |t, args|
  sh "bundle exec rails runner \"eval(File.read 'lib/tasks/#{args[:filename]}')\""
end

and

# lib/tasks/move_data.rb
require 'rest-client'

base = "http://ec2-52-69-132-73.ap-northeast-1.compute.amazonaws.com"
headers = {Authorization: "..."}
pic_base = "/home/hiogawa/myapps/wordCollectorRails/public"

me = User.first

me.categories.map do |c|

  # create categories
  resp = RestClient.post "#{base}/api/categories",
                         {category: {name: c.name, description: ""}},
                         headers
  c_ = JSON.parse(resp)

  c.posts.map do |p|

    begin
      # create items
      RestClient.post "#{base}/api/items",
                      {
                        category_id: c_["category"]["id"],
                        item: {
                          word:     p.word,
                          sentence: p.sentence,
                          meaning:  p.meaning,
                          picture:  File.new(pic_base + p.picture.url.sub(/\?.*/, ""), 'rb')
                        }
                      },
                      headers
    rescue => e
      puts p
      puts e.response
    end
  end
end