Add this line to your application's Gemfile:

gem 'impex'

And then execute:

$ bundle

?> rails g impex:install
?> rake db:migrate
?> rake impex:all

By default, impex looks for CSV files in public/csv_import/TABLE_NAME/*.csv. TABLE_NAME stands for the name of table (e.g: users, ...). Each file may require a header that contains the exact name of the field. If it's not the case, a hook is provided to re-organize the mapping of the file before check.

Before the update of a field, impex will check if csv column value has never been a value of the table field (e.g: Update the email only if has never been an email of the person).

However, it's possible to custom the list of field to look up via a whitelisting config.

By default, the reference field (a unique field that's used as identifier to search in history) corresponds to a field named reference. You can customize the reference for each tables in config.

The gem provides 2 tools: a rake tasks and a rails generator.

The rails generator impex:install provides 2 files:

$ rails g impex:install
  create  config/initializers/impex.rb
  create  db/migrate/20160831080017_create_impex_histories.rb

The impex.rb initializer permits to manage the path of the CSV files and the whitelisting for the history.

Impex.configure({
  file_loader: { loader: :file_system, path: "#{Rails.root}public/" },
  history_whitelisting: {
    buildings: [:manager_name],
    people:    [:email, :phone_number, :cell_phone, :address]
  }
})

The rake task impex:all iterates through each file available in the path defined in config.

The project is based on the principle of "composition over inheritance". There is 3 main entities in the project: the core, the FileLoader system and the HistoryManager system.

The FileLoader is in charge of fetching and converting CSV files into a format that the impex can consume. The project provides one file loader by default: Impex::FileLoader::FileSystem.

If you want to add a new FileLoader (e.g FileLoader::FileLoader::S3), you can create a new class that inherits from Impex::FileLoader::Base and implements the instance method #load that returns an array of Impex::File.

In order to ease the convertion of the CSV file to a format that is consumable by the impex, the project provides an Impex::FileFormatter.build(INSTANCE_OF_FILE_CLASS) method that takes an instance of the ruby File class anf returns an Impex::File instance.

The HistoryManager is in charge of filter csv columns by maintaining an history of the value used in past for each column. The project provides one history manager by default: Impex::HistoryManager::ActiveRecord.

If you want to add a new HistoryManager (e.g FileLoader::HistoryManager::Redis), you can create a new class that inherits from Impex::HistoryManager::Base and implements the instance methods #filter_data_with_history(row) and #update_history(row).

Once a new FileLoader or HistoryManager is added to the project then you can choose it by providing a config (configs name have to be in underscore syntax)

file_loader: { loader: :s3, path: "#{Rails.root}public/" },
history_manager: { manager: :redis, table: "impex_histories" }

The entry point of the project is the method Impex::Engine.run. This method accepts a block. The block (that takes a Impex::Row instance as parameter) is called just before history checks and insertion of each row.

So you can re-arrange the mapping of the row (recall: each header value has to correspond to a table column name)

Impex::Engine.run do |row|
  if row.table == "users"
    row.columns[:email] = row.columns.delete('Company email')
    # ...
  end
end

by overriding the method Impex::Engine#insert_row(record, row) you can manage the way to update the record according to the filtered row.

Thanks for reading.

Twitter
GitHub
LinkedIn