cran.rb

module PackageManager
  class CRAN < Base
    HAS_VERSIONS = true
    HAS_DEPENDENCIES = true
    BIBLIOTHECARY_SUPPORT = true
    URL = 'https://cran.r-project.org/'
    COLOR = '#198CE7'

    def self.package_link(project, _version = nil)
      "https://cran.r-project.org/package=#{project.name}"
    end

    def self.download_url(name, version = nil)
      "https://cran.r-project.org/src/contrib/#{name}_#{version}.tar.gz"
    end

    def self.documentation_url(name, _version = nil)
      "http://cran.r-project.org/web/packages/#{name}/#{name}.pdf"
    end

    def self.check_status_url(project)
      "http://cran.r-project.org/web/packages/#{project.name}/index.html"
    end

    def self.project_names
      html = get_html("https://cran.r-project.org/web/packages/available_packages_by_date.html")
      html.css('tr')[1..-1].map{|tr| tr.css('td')[1].text.strip}
    end

    def self.recent_names
      project_names[0..15].uniq
    end

    def self.project(name)
      html = get_html("https://cran.r-project.org/web/packages/#{name}/index.html")
      info = {}
      table = html.css('table')[0]
      table.css('tr').each do |tr|
        tds = tr.css('td').map(&:text)
        info[tds[0]] = tds[1]
      end

      {name: name, html: html, info: info}
    end

    def self.mapping(project)
      {
        :name => project[:name],
        :homepage => project[:info].fetch('URL:', '').split(',').first,
        :description => project[:html].css('h2').text.split(':')[1..-1].join(':').strip,
        :licenses => project[:info]['License:'],
        :repository_url => repo_fallback('', (project[:info].fetch('URL:', '').split(',').first.presence || project[:info]['BugReports:']))
      }
    end

    def self.versions(project)
      [{
        :number => project[:info]['Version:'],
        :published_at => project[:info]['Published:']
      }] + find_old_versions(project)
    end

    def self.find_old_versions(project)
      archive_page = get_html("https://cran.r-project.org/src/contrib/Archive/#{project[:name]}/")
      archive_page.css('table').css('tr').select do |tr|
        tds = tr.css('td')
        tds[1] && tds[1].text.match(/tar\.gz$/)
      end.map do |tr|
        tds = tr.css('td')
        {
          :number => tds[1].text.strip.split('_').last.gsub('.tar.gz', ''),
          :published_at => tds[2].text.strip
        }
      end
    end

    def self.dependencies(name, version, project)
      find_and_map_dependencies(name, version, project)
    end

    def self.find_dependencies(name, version)
      begin
        begin
          url = "https://cran.rstudio.com/src/contrib/#{name}_#{version}.tar.gz"
          head_response = Typhoeus.head(url)
          raise if head_response.code != 200
        rescue
          url = "https://cran.rstudio.com/src/contrib/Archive/#{name}/#{name}_#{version}.tar.gz"
        end

        folder_name = "#{name}_#{version}"
        tarball_name = "#{folder_name}.tar.gz"
        downloaded_file = File.open "/tmp/#{tarball_name}", 'wb'
        request = Typhoeus::Request.new(url)
        request.on_headers do |response|
          if response.code != 200
            return []
          end
        end
        request.on_body {|chunk| downloaded_file.write(chunk) }
        request.on_complete { downloaded_file.close }
        request.run

        `mkdir /tmp/#{folder_name} && tar xvzf /tmp/#{tarball_name} -C /tmp/#{folder_name}  --strip-components 1`

        contents = `cat /tmp/#{folder_name}/DESCRIPTION`

        `rm -rf /tmp/#{folder_name} /tmp/#{tarball_name}`

        Bibliothecary.analyse_file('DESCRIPTION', contents).first.fetch(:dependencies)
      ensure
        `rm -rf /tmp/#{folder_name} /tmp/#{tarball_name}`
        []
      end
    end
  end
end