csv import: prevent datatype error

Gemfile

@@ -45,9 +45,9 @@ group :development do
   # Access an interactive console on exception pages or by calling 'console' anywhere in the code.
   gem 'web-console', '>= 3.3.0'
   gem 'listen', '>= 3.0.5', '< 3.2'
+  gem 'dotenv-rails'
   # Spring speeds up development by keeping your application running in the background. Read more: https://github.com/rails/spring
   gem 'spring'
-  gem 'dotenv-rails'
   gem 'spring-watcher-listen', '~> 2.0.0'
 end
 

lib/tasks/crawler.rake

 require 'open-uri'
 
-@logger ||= Logger.new("#{Rails.root}/log/crawler.log")
+@logger ||= Logger.new("./log/import_data.log")
 
 namespace :crawl do
   desc 'crawl industries locations jobs'
@@ -31,7 +31,7 @@ def get_job_links(page, link)
 end
 
 def crawl_company(company_link)
-  uri = URI.parse(CGI.escape(company_link)) # fix error: uri must be ascii only
+  uri = URI.parse(URI.escape(company_link)) # fix error: uri must be ascii only
   document = Nokogiri::HTML(URI.open(uri))
   company_name = document.css('.content .name').text
   return if company_name.empty?
@@ -48,7 +48,7 @@ rescue StandardError => e
 end
 
 def crawl_job(job_link)
-  uri = URI.parse(CGI.escape(job_link)) # fix error: uri must be ascii only
+  uri = URI.parse(URI.escape(job_link)) # fix error: uri must be ascii only
   document = Nokogiri::HTML(URI.open(uri))
   job_title = document.at_css('.job-desc p.title').text
   return if job_title.empty?

lib/tasks/ftp_import.rake → lib/tasks/csv_import.rake

@@ -2,8 +2,10 @@ require 'csv'
 require 'zip'
 require_relative '../common/ftp'
 
-namespace :ftp_import do
-  desc 'FTP import csv file'
+@logger ||= Logger.new("./log/import_data.log")
+
+namespace :csv_import do
+  desc 'Download csv file from FTP and import'
   task csv: :environment do
     destination_dir = './lib/data'
     ftp = Ftp.new('192.168.1.156', 'training', 'training')
@@ -26,10 +28,10 @@ def extract_zip(file, destination)
 end
 
 def import_job(direction)
-  # i = 0
+  index = 0
   CSV.foreach("#{direction}/jobs.csv", headers: true) do |row|
-    # i+=1
-    next if row['name'].blank? || !row['category'].is_a?(String) || row['company name'].blank?
+    index += 1
+    next if integer?(row['category'])
 
     title = row['name'].strip
     company = Company.find_or_create_by(name: row['company name']) do |c|
@@ -41,8 +43,7 @@ def import_job(direction)
     industry = Industry.find_or_create_by(name: row['category'].strip)
     level = row['level'].try(:strip)
     salary = row['salary'].try(:strip)
-    locations_name = row['work place'].is_a?(Array) ? row['work place'] : row['work place'].split(',')
-    # byebug if i == 61
+    locations_name = row['work place'].tr('"[]', '').split(',')
     locations = Location.where(city: locations_name)
     locations = locations_name.map { |city| Location.create(oversea: false, city: city) } if locations.empty?
     description = "Benefits:\n#{row['benefit']}\n"\
@@ -58,5 +59,9 @@ def import_job(direction)
   end
 rescue StandardError => e
   puts e
-  @logger.error e.message
+  @logger.error "Job #{index}: #{e.message}"
+end
+
+def integer?(str)
+  str.to_i.to_s == str
 end