uhaul 0.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: 70a37ead64b21ccff91595e73dfedf11b378c2e5ea14acdd0d18e9d45735e3a5
+  data.tar.gz: d355639b46eb03ce4ce4be447520f05551ecfc8f23f7ebacc8260a6bb5d71203
+SHA512:
+  metadata.gz: b73ef2ae65031b627177dd1b8c5fa647ce6273b3f09efb8975a42182d6be90e05d6f6e27c5c50dfae3c836f81cd48e1211fd57bf2918081cc6664999b3daef24
+  data.tar.gz: bec905b2070dd9a60d0d32b1f3f3de1fbe184d27f495152524911f90d9eff9c1b3c2b6b833b1e3f3cb26960de25a606396d3ca0b9dcf9292f42a1498bed1fc98

data/Gemfile

@@ -0,0 +1,15 @@
+# frozen_string_literal: true
+
+source 'https://rubygems.org'
+
+gemspec
+
+gem 'rake'
+gem 'rspec'
+gem 'rspec_junit_formatter'
+gem 'rubocop'
+gem 'rubocop-rake'
+gem 'rubocop-rspec'
+gem 'vcr'
+gem 'webmock'
+gem 'yard'

data/README.md

@@ -0,0 +1,51 @@
+# UHaul
+
+[![LICENSE](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/ksylvest/uhaul/blob/main/LICENSE)
+[![RubyGems](https://img.shields.io/gem/v/uhaul)](https://rubygems.org/gems/uhaul)
+[![GitHub](https://img.shields.io/badge/github-repo-blue.svg)](https://github.com/ksylvest/uhaul)
+[![Yard](https://img.shields.io/badge/docs-site-blue.svg)](https://uhaul.ksylvest.com)
+[![CircleCI](https://img.shields.io/circleci/build/github/ksylvest/uhaul)](https://circleci.com/gh/ksylvest/uhaul)
+
+## Installation
+
+```bash
+gem install uhaul
+```
+
+## Configuration
+
+```ruby
+require 'uhaul'
+
+UHaul.configure do |config|
+  config.user_agent = '../..' # ENV['NSASTORAGE_USER_AGENT']
+  config.timeout = 30 # ENV['NSASTORAGE_TIMEOUT']
+  config.proxy_url = 'http://user:pass@superproxy.zenrows.com:1337' # ENV['NSASTORAGE_PROXY_URL']
+end
+```
+
+## Usage
+
+```ruby
+require 'uhaul'
+
+sitemap = UHaul::Facility.sitemap
+sitemap.links.each do |link|
+  url = link.loc
+  facility = UHaul::Facility.fetch(url:)
+
+  puts facility.text
+
+  facility.prices.each do |price|
+    puts price.text
+  end
+
+  puts
+end
+```
+
+## CLI
+
+```bash
+uhaul crawl
+```

data/bin/console

@@ -0,0 +1,11 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'bundler/setup'
+require 'uhaul'
+
+# You can add fixtures and/or initialization code here to make experimenting
+# with your gem easier. You can also use a different console, if you like.
+
+require 'irb'
+IRB.start(__FILE__)

data/bin/setup

@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+set -euo pipefail
+IFS=$'\n\t'
+set -vx
+
+bundle install
+
+# Do any other automated setup that you need to do here

data/exe/uhaul

@@ -0,0 +1,7 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'uhaul'
+
+cli = UHaul::CLI.new
+cli.parse

data/lib/uhaul/address.rb

@@ -0,0 +1,63 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The address (street + city + state + zip) of a facility.
+  class Address
+    ADDRESS_SELECTOR = '.item-des-box .text-box .part_title_1'
+    ADDRESS_REGEX = /(?<street>.+),\s+(?<city>.+),\s+(?<state>.+)\s+(?<zip>\d{5})/
+    # @attribute [rw] street
+    #   @return [String]
+    attr_accessor :street
+
+    # @attribute [rw] city
+    #   @return [String]
+    attr_accessor :city
+
+    # @attribute [rw] state
+    #   @return [String]
+    attr_accessor :state
+
+    # @attribute [rw] zip
+    #   @return [String]
+    attr_accessor :zip
+
+    # @param street [String]
+    # @param city [String]
+    # @param state [String]
+    # @param zip [String]
+    def initialize(street:, city:, state:, zip:)
+      @street = street
+      @city = city
+      @state = state
+      @zip = zip
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "street=#{@street.inspect}",
+        "city=#{@city.inspect}",
+        "state=#{@state.inspect}",
+        "zip=#{@zip.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String]
+    def text
+      "#{street}, #{city}, #{state} #{zip}"
+    end
+
+    # @param data [Hash]
+    #
+    # @return [Address]
+    def self.parse(data:)
+      new(
+        street: data['streetAddress'],
+        city: data['addressLocality'],
+        state: data['addressRegion'],
+        zip: data['postalCode']
+      )
+    end
+  end
+end

data/lib/uhaul/cli.rb

@@ -0,0 +1,58 @@
+# frozen_string_literal: true
+
+require 'optparse'
+
+module UHaul
+  # Used when interacting with the library from the command line interface (CLI).
+  #
+  # Usage:
+  #
+  #   cli = UHaul::CLI.new
+  #   cli.parse
+  class CLI
+    module Code
+      OK = 0
+      ERROR = 1
+    end
+
+    # @param argv [Array<String>]
+    def parse(argv = ARGV)
+      parser.parse!(argv)
+      command = argv.shift
+
+      case command
+      when 'crawl' then crawl
+      else
+        warn("unsupported command=#{command.inspect}")
+        exit(Code::ERROR)
+      end
+    end
+
+    private
+
+    def crawl
+      Crawl.run
+      exit(Code::OK)
+    end
+
+    def help(options)
+      puts(options)
+      exit(Code::OK)
+    end
+
+    def version
+      puts(VERSION)
+      exit(Code::OK)
+    end
+
+    # @return [OptionParser]
+    def parser
+      OptionParser.new do |options|
+        options.banner = 'usage: uhaul [options] <command> [<args>]'
+
+        options.on('-h', '--help', 'help') { help(options) }
+        options.on('-v', '--version', 'version') { version }
+      end
+    end
+  end
+end

data/lib/uhaul/config.rb

@@ -0,0 +1,58 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The core configuration.
+  class Config
+    # @attribute [rw] accept_language
+    #   @return [String]
+    attr_accessor :accept_language
+
+    # @attribute [rw] user_agent
+    #   @return [String]
+    attr_accessor :user_agent
+
+    # @attribute [rw] timeout
+    #   @return [Integer]
+    attr_accessor :timeout
+
+    # @attribute [rw] proxy_url
+    #   @return [String]
+    attr_accessor :proxy_url
+
+    def initialize
+      @accept_language = ENV.fetch('NSASTORAGE_ACCEPT_LANGUAGE', 'en-US,en;q=0.9')
+      @user_agent = ENV.fetch('NSASTORAGE_USER_AGENT', "uhaul.rb/#{VERSION}")
+      @timeout = Integer(ENV.fetch('NSASTORAGE_TIMEOUT', 60))
+      @proxy_url = ENV.fetch('NSASTORAGE_PROXY_URL', nil)
+    end
+
+    # @return [Boolean]
+    def headers?
+      !@user_agent.nil?
+    end
+
+    # @return [Boolean]
+    def timeout?
+      !@timeout.zero?
+    end
+
+    # @return [Boolean]
+    def proxy?
+      !@proxy_url.nil?
+    end
+
+    # @return [Hash<String, String>] e.g { 'User-Agent' => 'uhaul.rb/1.0.0' }
+    def headers
+      {
+        'Accept-Language' => @accept_language,
+        'User-Agent' => @user_agent
+      }
+    end
+
+    # @return [Array] e.g. ['proxy.example.com', 8080, 'user', 'pass']
+    def via
+      proxy_uri = URI.parse(@proxy_url)
+      [proxy_uri.host, proxy_uri.port, proxy_uri.user, proxy_uri.password]
+    end
+  end
+end

data/lib/uhaul/crawl.rb

@@ -0,0 +1,37 @@
+# frozen_string_literal: true
+
+module UHaul
+  # Handles the crawl command via CLI.
+  class Crawl
+    def self.run(...)
+      new(...).run
+    end
+
+    # @param stdout [IO] optional
+    # @param stderr [IO] optional
+    # @param options [Hash] optional
+    def initialize(stdout: $stdout, stderr: $stderr, options: {})
+      @stdout = stdout
+      @stderr = stderr
+      @options = options
+    end
+
+    def run
+      sitemap = Facility.sitemap
+      @stdout.puts("count=#{sitemap.links.count}")
+      @stdout.puts
+
+      sitemap.links.each { |link| process(url: link.loc) }
+    end
+
+    def process(url:)
+      @stdout.puts(url)
+      facility = Facility.fetch(url: url)
+      @stdout.puts(facility.text)
+      facility.prices.each { |price| @stdout.puts(price.text) }
+      @stdout.puts
+    rescue FetchError => e
+      @stderr.puts("url=#{url} error=#{e.message}")
+    end
+  end
+end

data/lib/uhaul/crawler.rb

@@ -0,0 +1,73 @@
+# frozen_string_literal: true
+
+module UHaul
+  # Used to fetch and parse either HTML or XML via a URL.
+  class Crawler
+    HOST = 'https://www.uhaul.com'
+
+    # @attribute url [String]
+    # @raise [FetchError]
+    # @return [Hash]
+    def self.json(url:)
+      new.json(url:)
+    end
+
+    # @param url [String]
+    # @raise [FetchError]
+    # @return [Nokogiri::HTML::Document]
+    def self.html(url:)
+      new.html(url:)
+    end
+
+    # @param url [String]
+    # @raise [FetchError]
+    # @return [Nokogiri::XML::Document]
+    def self.xml(url:)
+      new.xml(url:)
+    end
+
+    # @return [HTTP::Client]
+    def connection
+      @connection ||= begin
+        config = UHaul.config
+
+        connection = HTTP.use(:auto_deflate).use(:auto_inflate).persistent(HOST)
+        connection = connection.headers(config.headers) if config.headers?
+        connection = connection.timeout(config.timeout) if config.timeout?
+        connection = connection.via(*config.via) if config.proxy?
+
+        connection
+      end
+    end
+
+    # @param url [String]
+    # @return [HTTP::Response]
+    def fetch(url:)
+      response = connection.get(url)
+      raise FetchError.new(url:, response: response.flush) unless response.status.ok?
+
+      response
+    end
+
+    # @param url [String]
+    # @raise [FetchError]
+    # @return [Hash]
+    def json(url:)
+      JSON.parse(String(fetch(url:).body))
+    end
+
+    # @param url [String]
+    # @raise [FetchError]
+    # @return [Nokogiri::XML::Document]
+    def html(url:)
+      Nokogiri::HTML(String(fetch(url:).body))
+    end
+
+    # @param url [String]
+    # @raise [FetchError]
+    # @return [Nokogiri::XML::Document]
+    def xml(url:)
+      Nokogiri::XML(String(fetch(url:).body))
+    end
+  end
+end

data/lib/uhaul/dimensions.rb

@@ -0,0 +1,75 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The dimensions (width + depth + sqft) of a price.
+  class Dimensions
+    DEFAULT_WIDTH = 5.0 # feet
+    DEFAULT_DEPTH = 5.0 # feet
+    DEFAULT_HEIGHT = 8.0 # feet
+
+    DIMENSIONS_REGEX = /(?<width>[\d\.]+) x (?<depth>[\d\.]+)/
+
+    # @attribute [rw] depth
+    #   @return [Float]
+    attr_accessor :depth
+
+    # @attribute [rw] width
+    #  @return [Float]
+    attr_accessor :width
+
+    # @attribute [rw] height
+    #   @return [Float]
+    attr_accessor :height
+
+    # @param depth [Float]
+    # @param width [Float]
+    # @param height [Float]
+    def initialize(depth:, width:, height: DEFAULT_HEIGHT)
+      @depth = depth
+      @width = width
+      @height = height
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "depth=#{@depth.inspect}",
+        "width=#{@width.inspect}",
+        "height=#{@height.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String] e.g. "5×5"
+    def id
+      "#{format('%g', @width)}×#{format('%g', @depth)}"
+    end
+
+    # @return [Integer]
+    def sqft
+      Integer(@width * @depth)
+    end
+
+    # @return [Integer]
+    def cuft
+      Integer(@width * @depth * @height)
+    end
+
+    # @return [String] e.g. "10' × 10' (100 sqft)"
+    def text
+      "#{format('%g', @width)}' × #{format('%g', @depth)}' (#{sqft} sqft)"
+    end
+
+    # @param element [Nokogiri::XML::Element]
+    #
+    # @return [Dimensions]
+    def self.parse(element:)
+      text = element.at_css('.unit-select-item-detail').text
+      match = DIMENSIONS_REGEX.match(text)
+
+      width = match ? Float(match[:width]) : DEFAULT_WIDTH
+      depth = match ? Float(match[:depth]) : DEFAULT_DEPTH
+      new(depth:, width:, height: DEFAULT_HEIGHT)
+    end
+  end
+end

data/lib/uhaul/facility.rb

@@ -0,0 +1,196 @@
+# frozen_string_literal: true
+
+module UHaul
+  # A facility (address + geocode + prices) on uhaul.com.
+  #
+  # e.g. https://www.uhaul.com/Locations/Self-Storage-near-Inglewood-CA-90301/712030/
+  class Facility
+    class ParseError < StandardError; end
+
+    SITEMAP_URLS = %w[
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-AL.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-AK.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-AZ.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-AR.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-CA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-CO.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-CT.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-DC.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-DE.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-FL.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-GA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-HI.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-ID.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-IL.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-IN.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-IA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-KS.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-KY.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-LA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-ME.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MD.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MI.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MN.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MS.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MO.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-MT.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NE.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NV.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NH.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NJ.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NM.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NY.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-NC.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-ND.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-OH.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-OK.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-OR.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-PA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-RI.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-SC.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-SD.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-TN.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-TX.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-UT.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-VT.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-VA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-WA.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-WV.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-WI.ashx
+      https://www.uhaul.com/Locations/Sitemaps/Sitemap-for-Storage-in-WY.ashx
+    ].freeze
+
+    DEFAULT_EMAIL = 'service@uhaul.com'
+    DEFAULT_PHONE = '+1-800-468-4285
+'
+
+    # @attribute [rw] id
+    #   @return [String]
+    attr_accessor :id
+
+    # @attribute [rw] url
+    #   @return [String]
+    attr_accessor :url
+
+    # @attribute [rw] name
+    #   @return [String]
+    attr_accessor :name
+
+    # @attribute [rw] phone
+    #   @return [String]
+    attr_accessor :phone
+
+    # @attribute [rw] email
+    #   @return [String]
+    attr_accessor :email
+
+    # @attribute [rw] address
+    #   @return [Address]
+    attr_accessor :address
+
+    # @attribute [rw] geocode
+    #   @return [Geocode, nil]
+    attr_accessor :geocode
+
+    # @attribute [rw] prices
+    #   @return [Array<Price>]
+    attr_accessor :prices
+
+    # @return [Sitemap]
+    def self.sitemap
+      links = sitemaps.map(&:links).reduce(&:+)
+      Sitemap.new(links:)
+    end
+
+    # @return [Array<Sitemap>]
+    def self.sitemaps
+      SITEMAP_URLS.map do |url|
+        Sitemap.fetch(url:)
+      end
+    end
+
+    # @param url [String]
+    #
+    # @return [Facility]
+    def self.fetch(url:)
+      document = Crawler.html(url:)
+      parse(url:, document:)
+    end
+
+    # @param url [String]
+    # @param document [Nokogiri::HTML::Document]
+    #
+    # @return [Facility]
+    def self.parse(url:, document:)
+      data = parse_ld_json_script(document:)
+
+      id = data['@id'].match(%r{(?<id>\d+)/#schema$})[:id]
+      name = data['name']
+
+      geocode = Geocode.parse(data: data['geo'])
+      address = Address.parse(data: data['address'])
+      prices = [] # TODO
+
+      new(id:, url:, name:, address:, geocode:, prices:)
+    end
+
+    # @param document [Nokogiri::HTML::Document]
+    #
+    # @raise [ParseError]
+    #
+    # @return [Hash]
+    def self.parse_ld_json_script(document:)
+      parse_ld_json_scripts(document:).find do |data|
+        data['@type'] == 'SelfStorage'
+      end || raise(ParseError, 'missing ld+json')
+    end
+
+    # @param document [Nokogiri::HTML::Document]
+    #
+    # @return [Array<Hash>]
+    def self.parse_ld_json_scripts(document:)
+      elements = document.xpath('//script[@type="application/ld+json"]')
+
+      elements.map { |element| element.text.empty? ? {} : JSON.parse(element.text) }
+    end
+
+    # @param id [String]
+    # @param url [String]
+    # @param name [String]
+    # @param address [Address]
+    # @param geocode [Geocode]
+    # @param phone [String]
+    # @param email [String]
+    # @param prices [Array<Price>]
+    def initialize(id:, url:, name:, address:, geocode:, phone: DEFAULT_PHONE, email: DEFAULT_EMAIL, prices: [])
+      @id = id
+      @url = url
+      @name = name
+      @address = address
+      @geocode = geocode
+      @phone = phone
+      @email = email
+      @prices = prices
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "id=#{@id.inspect}",
+        "url=#{@url.inspect}",
+        "address=#{@address.inspect}",
+        "geocode=#{@geocode.inspect}",
+        "phone=#{@phone.inspect}",
+        "email=#{@email.inspect}",
+        "prices=#{@prices.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String]
+    def text
+      "#{@id} | #{@name} | #{@phone} | #{@email} | #{@address.text} | #{@geocode ? @geocode.text : 'N/A'}"
+    end
+  end
+end

data/lib/uhaul/features.rb

@@ -0,0 +1,77 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The features (e.g. climate-controlled, inside-drive-up-access, outside-drive-up-access, etc) of a price.
+  class Features
+    # @param element [Nokogiri::XML::Element]
+    #
+    # @return [Features]
+    def self.parse(element:)
+      text = element.text
+
+      new(
+        climate_controlled: text.include?('Heated and Cooled'),
+        drive_up_access: text.include?('Drive Up Access'),
+        first_floor_access: text.include?('1st Floor')
+      )
+    end
+
+    # @param climate_controlled [Boolean]
+    # @param drive_up_access [Boolean]
+    # @param first_floor_access [Boolean]
+    def initialize(climate_controlled:, drive_up_access:, first_floor_access:)
+      @climate_controlled = climate_controlled
+      @drive_up_access = drive_up_access
+      @first_floor_access = first_floor_access
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "climate_controlled=#{@climate_controlled}",
+        "drive_up_access=#{@drive_up_access}",
+        "first_floor_access=#{@first_floor_access}"
+      ]
+
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String] e.g. ""
+    def id
+      [].tap do |ids|
+        ids << 'cc' if climate_controlled?
+        ids << 'dua' if drive_up_access?
+        ids << 'ffa' if first_floor_access?
+      end.join('-')
+    end
+
+    # @return [String] e.g. "Climate Controlled + First Floor Access"
+    def text
+      amenities.join(' + ')
+    end
+
+    # @return [Array<String>]
+    def amenities
+      [].tap do |amenities|
+        amenities << 'Climate Controlled' if climate_controlled?
+        amenities << 'Drive-Up Access' if drive_up_access?
+        amenities << 'First Floor Access' if first_floor_access?
+      end
+    end
+
+    # @return [Boolean]
+    def climate_controlled?
+      @climate_controlled
+    end
+
+    # @return [Boolean]
+    def drive_up_access?
+      @drive_up_access
+    end
+
+    # @return [Boolean]
+    def first_floor_access?
+      @first_floor_access
+    end
+  end
+end

data/lib/uhaul/fetch_error.rb

@@ -0,0 +1,12 @@
+# frozen_string_literal: true
+
+module UHaul
+  # Raised for unexpected HTTP responses.
+  class FetchError < Error
+    # @param url [String]
+    # @param response [HTTP::Response]
+    def initialize(url:, response:)
+      super("url=#{url} status=#{response.status.inspect} body=#{String(response.body).inspect}")
+    end
+  end
+end

data/lib/uhaul/geocode.rb

@@ -0,0 +1,48 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The geocode (latitude + longitude) of a facility.
+  class Geocode
+    LATITUDE_REGEX = /\\u0022lat\\u0022:(?<latitude>[\+\-\d\.]+)/
+    LONGITUDE_REGEX = /\\u0022long\\u0022:(?<longitude>[\+\-\d\.]+)/
+
+    # @attribute [rw] latitude
+    #   @return [Float]
+    attr_accessor :latitude
+
+    # @attribute [rw] longitude
+    #   @return [Float]
+    attr_accessor :longitude
+
+    # @param data [Hash]
+    #
+    # @return [Geocode]
+    def self.parse(data:)
+      latitude = data[:latitude]
+      longitude = data[:longitude]
+
+      new(latitude:, longitude:)
+    end
+
+    # @param latitude [Float]
+    # @param longitude [Float]
+    def initialize(latitude:, longitude:)
+      @latitude = latitude
+      @longitude = longitude
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "latitude=#{@latitude.inspect}",
+        "longitude=#{@longitude.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String]
+    def text
+      "#{@latitude},#{@longitude}"
+    end
+  end
+end

data/lib/uhaul/link.rb

@@ -0,0 +1,21 @@
+# frozen_string_literal: true
+
+module UHaul
+  # A link in a sitemap.
+  class Link
+    # @attribute [rw] loc
+    #   @return [String]
+    attr_accessor :loc
+
+    # @param loc [String]
+    # @param lastmod [String, nil]
+    def initialize(loc:)
+      @loc = loc
+    end
+
+    # @return [String]
+    def inspect
+      "#<#{self.class.name} loc=#{@loc.inspect}>"
+    end
+  end
+end

data/lib/uhaul/price.rb

@@ -0,0 +1,78 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The price (id + dimensions + rate) for a facility.
+  class Price
+    ID_REGEX = %r{(?<id>\d+)/(?:rent|reserve)/}
+    PRICE_SELECTOR = '[data-unit-size="small"],[data-unit-size="medium"],[data-unit-size="large"]'
+
+    # @attribute [rw] id
+    #   @return [String]
+    attr_accessor :id
+
+    # @attribute [rw] dimensions
+    #   @return [Dimensions]
+    attr_accessor :dimensions
+
+    # @attribute [rw] features
+    #   @return [Features]
+    attr_accessor :features
+
+    # @attribute [rw] rates
+    #   @return [Rates]
+    attr_accessor :rates
+
+    # @param facility_id [Integer]
+    #
+    # @return [Array<Price>]
+    def self.fetch(facility_id:)
+      url = "https://www.uhaul.com/facility-units/#{facility_id}"
+      data = Crawler.json(url:)['data']
+      return [] if data['error']
+
+      html = data['html']['units']
+      Nokogiri::HTML(html).css(PRICE_SELECTOR).map { |element| parse(element:) }
+    end
+
+    # @param id [String]
+    # @param dimensions [Dimensions]
+    # @param features [Features]
+    # @param rates [Rates]
+    def initialize(id:, dimensions:, features:, rates:)
+      @id = id
+      @dimensions = dimensions
+      @features = features
+      @rates = rates
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "id=#{@id.inspect}",
+        "dimensions=#{@dimensions.inspect}",
+        "features=#{@features.inspect}",
+        "rates=#{@rates.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String] e.g. "123 | 5' × 5' (25 sqft) | $100 (street) / $90 (web)"
+    def text
+      "#{@id} | #{@dimensions.text} | #{@rates.text} | #{@features.text}"
+    end
+
+    # @param element [Nokogiri::XML::Element]
+    #
+    # @return [Price]
+    def self.parse(element:)
+      link = element.at_xpath(".//a[contains(text(), 'Rent')]|//a[contains(text(), 'Reserve')]")
+      dimensions = Dimensions.parse(element:)
+      features = Features.parse(element:)
+      rates = Rates.parse(element:)
+
+      id = link ? ID_REGEX.match(link['href'])[:id] : "#{dimensions.id}-#{features.id}"
+
+      new(id:, dimensions:, features:, rates:)
+    end
+  end
+end

data/lib/uhaul/rates.rb

@@ -0,0 +1,59 @@
+# frozen_string_literal: true
+
+module UHaul
+  # The rates (street + web) for a facility
+  class Rates
+    STREET_SELECTOR = '.part_item_old_price'
+    WEB_SELECTOR = '.part_item_price'
+    VALUE_REGEX = /(?<value>[\d\.]+)/
+
+    # @attribute [rw] street
+    #   @return [Integer]
+    attr_accessor :street
+
+    # @attribute [rw] web
+    #   @return [Integer]
+    attr_accessor :web
+
+    # @param element [Nokogiri::XML::Element]
+    #
+    # @return [Rates]
+    def self.parse(element:)
+      street = parse_value(element: element.at_css(STREET_SELECTOR))
+      web = parse_value(element: element.at_css(WEB_SELECTOR))
+
+      new(street: street || web, web: web || street)
+    end
+
+    # @param element [Nokogiri::XML::Element]
+    #
+    # @return [Float, nil]
+    def self.parse_value(element:)
+      return if element.nil?
+
+      match = VALUE_REGEX.match(element.text)
+      Float(match[:value]) if match
+    end
+
+    # @param street [Integer]
+    # @param web [Integer]
+    def initialize(street:, web:)
+      @street = street
+      @web = web
+    end
+
+    # @return [String]
+    def inspect
+      props = [
+        "street=#{@street.inspect}",
+        "web=#{@web.inspect}"
+      ]
+      "#<#{self.class.name} #{props.join(' ')}>"
+    end
+
+    # @return [String] e.g. "$80 (street) | $60 (web)"
+    def text
+      "$#{@street} (street) | $#{@web} (web)"
+    end
+  end
+end

data/lib/uhaul/sitemap.rb

@@ -0,0 +1,40 @@
+# frozen_string_literal: true
+
+module UHaul
+  # A sitemap on uhaul.com.
+  class Sitemap
+    # @attribute [rw] links
+    #   @return [Array<Link>]
+    attr_accessor :links
+
+    # @param document [NokoGiri::XML::Document]
+    #
+    # @return [Sitemap]
+    def self.parse(document:)
+      links = document.xpath('//xmlns:url').map do |url|
+        loc = url.at_xpath('xmlns:loc').text
+        Link.new(loc:)
+      end
+
+      new(links: links.filter { |link| link.loc.match(%r{/Self-Storage/.*/\d+}) })
+    end
+
+    # @param url [String]
+    #
+    # @return [Sitemap]
+    def self.fetch(url:)
+      document = Crawler.xml(url:)
+      parse(document:)
+    end
+
+    # @param links [Array<Link>]
+    def initialize(links:)
+      @links = links
+    end
+
+    # @return [String]
+    def inspect
+      "#<#{self.class.name} links=#{@links.inspect}>"
+    end
+  end
+end

data/lib/uhaul/version.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module UHaul
+  VERSION = '0.1.0'
+end

data/lib/uhaul.rb

@@ -0,0 +1,26 @@
+# frozen_string_literal: true
+
+require 'http'
+require 'nokogiri'
+require 'zeitwerk'
+
+loader = Zeitwerk::Loader.for_gem
+loader.inflector.inflect 'uhaul' => 'UHaul'
+loader.inflector.inflect 'cli' => 'CLI'
+loader.setup
+
+# An interface for UHaul.
+module UHaul
+  class Error < StandardError; end
+
+  # @return [Config]
+  def self.config
+    @config ||= Config.new
+  end
+
+  # @yield [config]
+  # @yieldparam config [Config]
+  def self.configure
+    yield config
+  end
+end

metadata

@@ -0,0 +1,139 @@
+--- !ruby/object:Gem::Specification
+name: uhaul
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- Kevin Sylvestre
+autorequire:
+bindir: exe
+cert_chain: []
+date: 2024-12-17 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: http
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: json
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: optparse
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: zeitwerk
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+description: Uses HTTP.rb to scrape uhaul.com.
+email:
+- kevin@ksylvest.com
+executables:
+- uhaul
+extensions: []
+extra_rdoc_files: []
+files:
+- Gemfile
+- README.md
+- bin/console
+- bin/setup
+- exe/uhaul
+- lib/uhaul.rb
+- lib/uhaul/address.rb
+- lib/uhaul/cli.rb
+- lib/uhaul/config.rb
+- lib/uhaul/crawl.rb
+- lib/uhaul/crawler.rb
+- lib/uhaul/dimensions.rb
+- lib/uhaul/facility.rb
+- lib/uhaul/features.rb
+- lib/uhaul/fetch_error.rb
+- lib/uhaul/geocode.rb
+- lib/uhaul/link.rb
+- lib/uhaul/price.rb
+- lib/uhaul/rates.rb
+- lib/uhaul/sitemap.rb
+- lib/uhaul/version.rb
+homepage: https://github.com/ksylvest/uhaul
+licenses:
+- MIT
+metadata:
+  rubygems_mfa_required: 'true'
+  homepage_uri: https://github.com/ksylvest/uhaul
+  source_code_uri: https://github.com/ksylvest/uhaul
+  changelog_uri: https://github.com/ksylvest/uhaul
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: 3.2.0
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubygems_version: 3.5.22
+signing_key:
+specification_version: 4
+summary: A crawler for UHaul.
+test_files: []