The idea of this gist is to generate the number/proportion of companies that would have to get removed from the new banking rollout if it had gone out on 1st Jan 2017 until 1st Nov 2017. This means that we care about companies that were created before Jan 1st and were still users and any companies created since then.

new_banking_rollout.rb

def check_banking(company,first_compare_date)

    # look at vat status on first_compare_date, currently and for the year so far
    previous_vat_registration = company.is_vat_registered_on?(first_compare_date)
    current_vat_registration = company.is_vat_registered?
    vat_registered_this_year = company.is_vat_deregistered? && company.vat_deregistration_effective_date > first_compare_date

    # mfa bank feeds, check currently enabled and currently disabled separately
    still_enabled_previous_bank_feeds = company.bank_feeds.to_a.
        select { |feed| feed.state == 'enabled' }.
        select { |feed| feed.bank_service&.mfa? }.
        select { |feed| feed.enabled_at < first_compare_date }.any?
    
    disabled_previous_bank_feeds = company.bank_feeds.to_a.
        select { |feed| feed.state == 'disabled' }.
        select { |feed| feed.enabled_at.present? }.
        select { |feed| feed.bank_service&.mfa? }.
        select { |feed| feed.disabled_at > first_compare_date }.
        select { |feed| feed.enabled_at < first_compare_date }.any?
    
    previous_bank_feeds = still_enabled_previous_bank_feeds || disabled_previous_bank_feeds
      
    current_bank_feeds = company.bank_feeds.to_a.
        select { |feed| feed.state == 'enabled' }.
        select { |feed| feed.bank_service.mfa? }.any?
    bank_feeds_this_year = company.bank_feeds.to_a.
        select { |feed| feed.state == 'disabled' }.
        select { |feed| feed.enabled_at.present? }.
        select { |feed| feed.bank_service&.mfa? }.
        select { |feed| feed.enabled_at > first_compare_date }.any?

    # ruth thinks that people having stock, and then deleting it with no transactions
    # for anything other than 'testing/playing around' wil be rare

    # check for stock items that existed as of Jan 1st
    previous_stock_items = company.stock_items.any? { |stock| stock.created_at.to_date < first_compare_date }
    # check for any added this year
    stock_items_this_year = company.stock_items.any? { |stock| stock.created_at.to_date > first_compare_date }

    starts_in_new_banking = ![previous_vat_registration, previous_vat_registration, previous_stock_items].any?
    ends_in_new_banking = ![current_vat_registration,
        vat_registered_this_year,
        current_bank_feeds,
        bank_feeds_this_year,
        stock_items_this_year].any?
    moves_out_of_new_banking = starts_in_new_banking && !ends_in_new_banking
      
    company_data = {
      "company_id": company.id.to_i,
      "channel": company.subscription.channel.to_s,
      "previous_vat": previous_vat_registration,
      "previous_bank_feeds": previous_bank_feeds,
      "previous_stock": previous_stock_items,
      "current_vat": current_vat_registration || vat_registered_this_year,
      "current_bank_feeds": current_bank_feeds || bank_feeds_this_year,
      "current_stock": stock_items_this_year,
      "starts_in_new_banking": starts_in_new_banking,
      "moves_out_of_new_banking": moves_out_of_new_banking,
    }
  
    return company_data
end

switch_to_slave_db

models_to_include = [
  :subscription,
  :subscription_payments,
  :bank_feeds,
  { bank_feeds: :bank_service },
  :stock_items,
  :setup_stage,
]

jan_first = Date.new(2017,1,1)

batch = 1

lower_limit = (batch-1)*50000
upper_limit = batch*50000

file = "/data/results/new_banking_rollout_#{batch}.csv"

CSV.open(file, "w") do |csv|
  Company.includes(models_to_include).
    where('type': 'UkLimitedCompany').
    where('companies.id >= ?',lower_limit).where('companies.id < ?',upper_limit).
    where('bank_services.yodlee_mfa_type is not null').
    where('bank_feeds.enabled_at is not null').
    where.not(:subscriptions => { 'status': ['Expired Free Trial','Live Free Trial'] }).
    where('subscriptions.cancelled_at is null or subscriptions.cancelled_at > ?', jan_first).
    find_each(batch_size: 10) do |company|
    
    # look for first payment date to get subscription date
    first_compare_date = company.subscription_payments.to_a.sort_by { |payment| payment&.dated_on }.first&.dated_on
        
    # if the subscription date does not exist, grab the setup complete date for non Direct
    if !first_compare_date && company.setup_stage.stage_type == 'SetupDone'
      if company.subscription.channel != 'Direct'
        first_compare_date = company.setup_stage.created_at.to_date 
        # also get this for a Complimentary direct company if they have never paid us
      elsif company.subscription.status == 'Complimentary'
        first_compare_date = company.setup_stage.created_at.to_date
      end
    end

    # if no date found then skip this company since they never got to a point
    # that they would be eligible for new banking
    next unless first_compare_date
    
    #finally set first compare date to be jan first if it is before that
    first_compare_date = jan_first if first_compare_date < jan_first   
      
    data = check_banking(company,first_compare_date)

    csv << data.keys if csv.lineno.zero?
    csv << data.values
    csv.flush
  end
end

For companies created since Jan 1st we care about their status on their subscription date or set-up complete date when subscription date is not appropriate.