publicstorage 0.2.0 → 1.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5f5319690995d8d7f216e02fa0ead21f9c3bfc136e2e9ed8d46333f840909474
-  data.tar.gz: f0da92039ca2c9dd5dbf1256a5e3df2cffa304763ee766229e0175bfcf817047
+  metadata.gz: 385e04642b7995999e422333ae7c251c4e3556ae9958b0df0f35a387ca35297a
+  data.tar.gz: 0c99f850bedd80f482187364625aa583518768ad6fc7f4b749f5dafdfe0871d8
 SHA512:
-  metadata.gz: 723e933f0da23b57b731e9123a7a412eb9984e7f54897a77304d1f0753cc274f1ffbf68711b59d25e4f9283f38ca4e3ee0aa8ffc9e606cdf26f0a92d4581e634
-  data.tar.gz: 5149248cc1fa272c78cc9ac8666bfe52501b23228a2dba66533c94c25ca5294fc9e31717fae635ea46361253428197c0007f1f577fca0e34cde15f49021bb4bf
+  metadata.gz: 25af8483c84aded81acac087be7e7bf7c42a404200d8d1091b7f32184de6a61fd6abccdff9cd4d6b2efcf582f3619c37250adeb21440232d3afa90c685ea2894
+  data.tar.gz: d623ad587cb3c5b3d33aad2221002a411211f1cc5d39741b86e1ffa9f64732bc1ba9220a72e7e05bec40e852589a8ce09097a89ac22db489bda30df7776d3322

data/README.md

@@ -1,4 +1,4 @@
-# PublicStorage
+# Public Storage
 
 [![LICENSE](https://img.shields.io/badge/license-MIT-blue.svg)](https://github.com/ksylvest/publicstorage/blob/main/LICENSE)
 [![RubyGems](https://img.shields.io/gem/v/publicstorage)](https://rubygems.org/gems/publicstorage)
@@ -6,12 +6,25 @@
 [![Yard](https://img.shields.io/badge/docs-site-blue.svg)](https://publicstorage.ksylvest.com)
 [![CircleCI](https://img.shields.io/circleci/build/github/ksylvest/publicstorage)](https://circleci.com/gh/ksylvest/publicstorage)
 
+A Ruby library offering both a CLI and API for scraping [Public Storage](https://www.publicstorage.com/) self-storage facilities and prices.
+
 ## Installation
 
 ```bash
 gem install publicstorage
 ```
 
+## Configuration
+
+```ruby
+require 'publicstorage'
+
+PublicStorage.configure do |config|
+  config.user_agent = '../..' # ENV['PUBLICSTORAGE_USER_AGENT']
+  config.timeout = 30 # ENV['PUBLICSTORAGE_TIMEOUT']
+end
+```
+
 ## Usage
 
 ```ruby
@@ -20,7 +33,7 @@ require 'publicstorage'
 sitemap = PublicStorage::Facility.sitemap
 sitemap.links.each do |link|
   url = link.loc
-  facility = ExtraSpace::Facility.fetch(url:)
+  facility = PublicStorage::Facility.fetch(url:)
 
   puts facility.text
 
@@ -37,3 +50,7 @@ end
 ```bash
 publicstorage crawl
 ```
+
+```bash
+publicstorage crawl "https://www.publicstorage.com/self-storage-ca-venice/120.html"
+```

data/lib/publicstorage/cli.rb

@@ -21,7 +21,7 @@ module PublicStorage
       command = argv.shift
 
       case command
-      when 'crawl' then crawl
+      when 'crawl' then crawl(*argv)
       else
         warn("unsupported command=#{command.inspect}")
         exit(Code::ERROR)
@@ -30,8 +30,9 @@ module PublicStorage
 
     private
 
-    def crawl
-      PublicStorage::Facility.crawl
+    # @url [String] optional
+    def crawl(url = nil)
+      Crawl.run(url: url)
       exit(Code::OK)
     end
 
@@ -52,6 +53,11 @@ module PublicStorage
 
         options.on('-h', '--help', 'help') { help(options) }
         options.on('-v', '--version', 'version') { version }
+
+        options.separator <<~COMMANDS
+          commands:
+            crawl [url]
+        COMMANDS
       end
     end
   end

data/lib/publicstorage/config.rb

@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+module PublicStorage
+  # The core configuration.
+  class Config
+    # @attribute [rw] user_agent
+    #   @return [String]
+    attr_accessor :user_agent
+
+    # @attribute [rw] timeout
+    #   @return [Integer]
+    attr_accessor :timeout
+
+    def initialize
+      @user_agent = ENV.fetch('PUBLICSTORAGE_USER_AGENT', "publicstorage.rb/#{VERSION}")
+      @timeout = Integer(ENV.fetch('PUBLICSTORAGE_TIMEOUT', 60))
+    end
+  end
+end

data/lib/publicstorage/crawl.rb

@@ -0,0 +1,40 @@
+# frozen_string_literal: true
+
+module PublicStorage
+  # Handles the crawl command via CLI.
+  class Crawl
+    def self.run(...)
+      new(...).run
+    end
+
+    # @param stdout [IO] optional
+    # @param stderr [IO] optional
+    # @param url [String] optional
+    def initialize(stdout: $stdout, stderr: $stderr, url: nil)
+      @stdout = stdout
+      @stderr = stderr
+      @url = url
+    end
+
+    def run
+      if @url
+        process(url: @url)
+      else
+        sitemap = Facility.sitemap
+        @stdout.puts("count=#{sitemap.links.count}")
+        @stdout.puts
+        sitemap.links.each { |link| process(url: link.loc) }
+      end
+    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/publicstorage/crawler.rb

@@ -3,14 +3,7 @@
 module PublicStorage
   # Used to fetch and parse either HTML or XML via a URL.
   class Crawler
-    # Raised for unexpected HTTP responses.
-    class FetchError < StandardError
-      # @param url [String]
-      # @param response [HTTP::Response]
-      def initialize(url:, response:)
-        super("url=#{url} status=#{response.status.inspect} body=#{response.body.inspect}")
-      end
-    end
+    HOST = 'https://www.publicstorage.com'
 
     # @param url [String]
     # @raise [FetchError]
@@ -26,10 +19,22 @@ module PublicStorage
       new.xml(url:)
     end
 
+    # @return [HTTP::Client]
+    def connection
+      @connection ||= begin
+        config = PublicStorage.config
+
+        connection = HTTP.persistent(HOST)
+        connection = connection.headers('User-Agent' => config.user_agent) if config.user_agent
+        connection = connection.timeout(config.timeout) if config.timeout
+        connection
+      end
+    end
+
     # @param url [String]
     # @return [HTTP::Response]
     def fetch(url:)
-      response = HTTP.get(url)
+      response = connection.get(url)
       raise FetchError.new(url:, response: response.flush) unless response.status.ok?
 
       response

data/lib/publicstorage/dimensions.rb

@@ -41,11 +41,11 @@ module PublicStorage
       "#{format('%g', @width)}' × #{format('%g', @depth)}' (#{@sqft} sqft)"
     end
 
-    # @param element [Nokogiri::XML::Element]
+    # @param data [Hash]
     #
     # @return [Dimensions]
-    def self.parse(element:)
-      match = element.at(SELECTOR).text.match(/(?<depth>[\d\.]+)'x(?<width>[\d\.]+)'/)
+    def self.parse(data:)
+      match = data['dimension'].match(/(?<depth>[\d\.]+)'x(?<width>[\d\.]+)'/)
       depth = Float(match[:depth])
       width = Float(match[:width])
       sqft = Integer(depth * width)

data/lib/publicstorage/fetch_error.rb

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

data/lib/publicstorage/price.rb

@@ -3,6 +3,8 @@
 module PublicStorage
   # The price (id + dimensions + rate) for a facility
   class Price
+    GTM_SELECTOR = 'button[data-gtmdata]'
+
     # @attribute [rw] id
     #   @return [String]
     attr_accessor :id
@@ -43,8 +45,10 @@ module PublicStorage
     #
     # @return [Price]
     def self.parse(element:)
-      rates = Rates.parse(element:)
-      dimensions = Dimensions.parse(element:)
+      data = JSON.parse(element.at(GTM_SELECTOR).attribute('data-gtmdata'))
+
+      rates = Rates.parse(data:)
+      dimensions = Dimensions.parse(data:)
 
       new(
         id: element.attr('data-unitid'),

data/lib/publicstorage/rates.rb

@@ -3,9 +3,6 @@
 module PublicStorage
   # The rates (street + web) for a facility
   class Rates
-    STREET_SELECTOR = '.unit-prices .unit-pricing .unit-strike-through-price'
-    WEB_SELECTOR = '.unit-prices .unit-pricing .unit-price'
-
     # @attribute [rw] street
     #   @return [Integer]
     attr_accessor :street
@@ -35,12 +32,12 @@ module PublicStorage
       "$#{@street} (street) | $#{@web} (web)"
     end
 
-    # @param element [Nokogiri::XML::Element]
+    # @param data [Hash]
     #
     # @return [Rates]
-    def self.parse(element:)
-      street = Integer(element.at(STREET_SELECTOR).text.match(/(?<value>\d+)/)[:value])
-      web = Integer(element.at(WEB_SELECTOR).text.match(/(?<value>\d+)/)[:value])
+    def self.parse(data:)
+      street = data['listprice']
+      web = data['saleprice']
       new(street:, web:)
     end
   end

data/lib/publicstorage/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module PublicStorage
-  VERSION = '0.2.0'
+  VERSION = '1.0.0'
 end

data/lib/publicstorage.rb

@@ -9,6 +9,18 @@ loader.inflector.inflect 'publicstorage' => 'PublicStorage'
 loader.inflector.inflect 'cli' => 'CLI'
 loader.setup
 
+# An interface for PublicStorage.
 module PublicStorage
   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

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: publicstorage
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 1.0.0
 platform: ruby
 authors:
 - Kevin Sylvestre
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2024-11-27 00:00:00.000000000 Z
+date: 2024-12-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: http
@@ -82,9 +82,12 @@ files:
 - lib/publicstorage.rb
 - lib/publicstorage/address.rb
 - lib/publicstorage/cli.rb
+- lib/publicstorage/config.rb
+- lib/publicstorage/crawl.rb
 - lib/publicstorage/crawler.rb
 - lib/publicstorage/dimensions.rb
 - lib/publicstorage/facility.rb
+- lib/publicstorage/fetch_error.rb
 - lib/publicstorage/geocode.rb
 - lib/publicstorage/link.rb
 - lib/publicstorage/price.rb
@@ -97,8 +100,9 @@ licenses:
 metadata:
   rubygems_mfa_required: 'true'
   homepage_uri: https://github.com/ksylvest/publicstorage
-  source_code_uri: https://github.com/ksylvest/publicstorage
-  changelog_uri: https://github.com/ksylvest/publicstorage
+  source_code_uri: https://github.com/ksylvest/publicstorage/tree/v1.0.0
+  changelog_uri: https://github.com/ksylvest/publicstorage/releases/tag/v1.0.0
+  documentation_uri: https://publicstorage.ksylvest.com/
 post_install_message: 
 rdoc_options: []
 require_paths: