require 'zip' require 'stringio' class FileProcessorJob include Sidekiq::Job def perform(inn, output_path) Rails.logger.info "=== Начало загрузки для ИНН: #{inn} ===" download_inn(inn, output_path) Rails.logger.info "=== Завершено для ИНН: #{inn} ===" rescue => e Rails.logger.error "=== ОШИБКА: #{e.message} ===" Rails.logger.error e.backtrace.first(5).join("\n") raise end private def get_content(url, json_need = false) headers = { 'User-Agent' => "Mozilla/5.0 (Windows NT 10.0; Win64; x64)", 'Accept-Language' => 'en-US,en;q=0.9', 'Accept' => 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8' } content = HTTParty.get(url, headers: headers).body json_need ? JSON.parse(content) : content end def download_inn(inn, output) url = "https://bo.nalog.gov.ru/advanced-search/organizations/search?query=#{inn}&page=0&size=20" content = get_content(url, true) id = content['content'][0]['id'] url = "https://bo.nalog.gov.ru/nbo/organizations/#{id}/bfo/" content = get_content(url, true) periods_and_details_ids = content.map { |item| [item['period'], item['typeCorrections'][0]['correction']['id']] } periods_and_details_ids.each do |period, details_id| url = "https://bo.nalog.gov.ru/download/bfo/#{id}?auditReport=false&balance=true"\ "&capitalChange=true&clarification=false&targetedFundsUsing=false&detailsId=#{details_id}"\ "&financialResult=true&fundsMovement=true&type=XLS&period=#{period}" content = get_content(url) zip_file = Zip::File.open_buffer(StringIO.new(content)) zip_file.each do |entry| output_file = File.join(output, entry.name) FileUtils.mkdir_p(File.dirname(output_file)) zip_file.extract(entry, output_file) { true } # перезаписывать если существует end end end end