picturehouse_uk 4.0.0 → 5.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: cb14c9688f95477661a9dae62733b969a842bd60
-  data.tar.gz: 68b02405cb818334fa91d1a85355ea342517c08b
+SHA256:
+  metadata.gz: 5c4c71062363f2efdabda5696692051f79c7915f250367ff8c16536e058c821b
+  data.tar.gz: a76bf492b267b7ae89cbbe829b4269dea9ac0207add6b41fe47685d1df76fbb9
 SHA512:
-  metadata.gz: 6d127befd1075646556a86f7579a4c3f00d84f45dc84902dedec6728881aaa4daafb0d1a3164706ae1417b2bd1f90b582cecf6e2263c4e3446d5dacfb4ff82f4
-  data.tar.gz: c9030d13d6db43f2ec2b1a8aeded059bd54d81f7678887a60b3ba9241baa3cf35f39f6294224b2e36af6ab0fcb9de711369a35d37ddb913e39d764a59a6ab719
+  metadata.gz: b1722850d1b2f161725a961f420b8122c404161050c794b02d93a8f735c5d7053745c2eb35de8d0200e5cc20cd2525024ed94a4cc63357fe2157fc9e149fdc11
+  data.tar.gz: d5225d4815b31cd31d072820f55e257e72cc415e619c8cbb3eaa6b5c7609d81b624107554d367036164f07da58d83bc5656d5e0ad42df154d955af02411d06fd

data/.github/workflows/ci.yml

@@ -0,0 +1,26 @@
+name: CI
+
+on:
+  push:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Set up Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: '3.4'
+        bundler-cache: true
+
+    - name: Run tests
+      run: bundle exec rake
+
+    - name: Run live tests
+      run: bundle exec rake live

data/.github/workflows/release.yml

@@ -0,0 +1,26 @@
+name: Release Gem
+
+on:
+  workflow_dispatch:
+
+jobs:
+  release:
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write
+      contents: write
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Set up Ruby
+      uses: ruby/setup-ruby@v1
+      with:
+        ruby-version: '3.4'
+        bundler-cache: true
+
+    - name: Run tests
+      run: bundle exec rake test
+
+    - name: Release gem
+      uses: rubygems/release-gem@v1

data/.gitignore

@@ -15,3 +15,4 @@ spec/reports
 test/tmp
 test/version_tmp
 tmp
+vendor/*

data/.ruby-version

@@ -0,0 +1 @@
+3.4.7

data/CHANGELOG.md

@@ -2,6 +2,31 @@
 All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
+## [5.0.0] - 2025-10-31
+
+### Added
+- JSON API client for fetching performance data from `/api/get-movies-ajax` endpoint
+- `PicturehouseUk::Internal::Api` for making API requests
+- `PicturehouseUk::Internal::Parser::Screenings` for parsing JSON responses (replaced old HTML parser)
+- Comprehensive tests for JSON API functionality
+- Sample JSON fixtures for testing
+
+### Changed
+- `Performance.at()` now uses JSON API instead of HTML parsing
+- Improved variant detection (arts, baby, kids, senior, imax)
+- Booking URLs now use Vista format: `https://web.picturehouses.com/order/showtimes/{cinema_id}-{session_id}/seats`
+- Enhanced 3D film detection from both title and attributes
+- Increased test coverage to 98.65%
+
+### Removed
+- Old HTML parser `Parser::Screenings` (completely replaced with JSON API version)
+- Test file for old HTML parser
+
+### Technical Details
+- The Picturehouse website now dynamically populates screening data via AJAX calls to their JSON API
+- The `#show_all_date_list` div is populated client-side with data from the API endpoint
+- HTML parsing approach replaced with direct API consumption for better reliability
+
 ## 4.0.0 - 2016-02-10
 
 The cinebase standardisation release.

data/Rakefile

@@ -37,8 +37,8 @@ task :fixtures do
   require_relative 'rake/fixture_creator'
 
   FixtureCreator.new.home
-  %w(Duke_Of_Yorks Dukes_At_Komedia Phoenix_Picturehouse
-     National_Media_Museum).each do |cinema_id|
+  FixtureCreator.new.cinemas
+  %w(duke-of-york-s-picturehouse duke-s-at-komedia phoenix-picturehouse).each do |cinema_id|
     FixtureCreator.new.cinema(cinema_id)
   end
 end

data/lib/picturehouse_uk/cinema.rb

@@ -2,9 +2,9 @@ module PicturehouseUk
   # The object representing a cinema on the Picturehouse UK website
   class Cinema < Cinebase::Cinema
     # address css
-    ADDRESS_CSS = '.static-content #contact-us + p:first'.freeze
-    # cinema link css
-    CINEMA_LINKS_CSS = '.footer .col-sm-3 option + option'.freeze
+    ADDRESS_CSS = '.cinemaAdrass:not(.openingTime)'.freeze
+    # cinema link css on the /cinema page
+    CINEMA_LINKS_CSS = 'a[href*="/cinema/"]'.freeze
 
     # @!attribute [r] id
     #   @return [Integer] the numeric id of the cinema on the Cineworld website
@@ -147,15 +147,15 @@ module PicturehouseUk
     private
 
     def self.cinema_links
-      home_doc.css(CINEMA_LINKS_CSS)
+      cinemas_doc.css(CINEMA_LINKS_CSS)
     end
     private_class_method :cinema_links
 
-    def self.home_doc
-      @home_doc ||=
-        Nokogiri::HTML(PicturehouseUk::Internal::Website.new.home)
+    def self.cinemas_doc
+      @cinemas_doc ||=
+        Nokogiri::HTML(PicturehouseUk::Internal::Website.new.cinemas)
     end
-    private_class_method :home_doc
+    private_class_method :cinemas_doc
 
     def address_node
       @address_node ||= info_doc.css(ADDRESS_CSS)
@@ -163,11 +163,11 @@ module PicturehouseUk
 
     def info_doc
       @info_doc ||=
-        Nokogiri::HTML(PicturehouseUk::Internal::Website.new.info(id))
+        Nokogiri::HTML(PicturehouseUk::Internal::Website.new.information(id))
     end
 
     # @api private
-    # Utility class to parse the links spat out from the options
+    # Utility class to parse the links from the cinemas page
     class ListParser
       def initialize(nodes)
         @nodes = nodes
@@ -175,22 +175,38 @@ module PicturehouseUk
 
       def to_hash
         @nodes.each_with_object({}) do |node, result|
-          result[id(node)] = { name: name(node), url: url(node) }
+          cinema_id = id(node)
+          cinema_name = name(node)
+          # Skip links that don't have valid cinema data
+          next unless cinema_id && cinema_name
+          # Skip duplicate entries
+          next if result.key?(cinema_id)
+
+          result[cinema_id] = { name: cinema_name, url: url(node) }
         end
       end
 
       private
 
       def id(node)
-        url(node).match(%r{/cinema/(.+)$})[1]
+        href = url(node)
+        return nil unless href
+        match = href.match(%r{/cinema/([^/?#]+)})
+        match ? match[1] : nil
       end
 
       def name(node)
-        node.children.first.to_s.split(' — ')[1]
+        # Cinema name is in a <p> tag within the link
+        p_tag = node.css('p').first
+        return nil unless p_tag
+        # Get text and remove the location span
+        p_tag.children.find { |child| child.text? }&.text&.strip
       end
 
       def url(node)
-        node.get_attribute('data-href')
+        href = node.get_attribute('href')
+        # Convert relative URLs to absolute if needed
+        href&.start_with?('http') ? href : "https://www.picturehouses.com#{href}"
       end
     end
   end

data/lib/picturehouse_uk/internal/api.rb

@@ -0,0 +1,42 @@
+require 'net/http'
+require 'json'
+require 'openssl'
+
+module PicturehouseUk
+  # @api private
+  module Internal
+    # Fetches JSON data from the Picturehouse API
+    class Api
+      API_BASE = 'https://www.picturehouses.com'.freeze
+      API_ENDPOINT = '/api/get-movies-ajax'.freeze
+
+      # Fetch movie data for a cinema
+      # @param cinema_id [String] the cinema ID
+      # @param date [String] the date in YYYY-MM-DD format, or 'show_all_dates'
+      # @return [Hash] parsed JSON response
+      def get_movies(cinema_id, date = 'show_all_dates')
+        uri = URI("#{API_BASE}#{API_ENDPOINT}")
+
+        http = Net::HTTP.new(uri.host, uri.port)
+        http.use_ssl = true
+        http.verify_mode = OpenSSL::SSL::VERIFY_NONE
+
+        request = Net::HTTP::Post.new(uri.path)
+        request.set_form_data(
+          'start_date' => date,
+          'cinema_id' => cinema_id,
+          'filters' => []
+        )
+
+        response = http.request(request)
+
+        return {} unless response.is_a?(Net::HTTPSuccess)
+
+        JSON.parse(response.body)
+      rescue StandardError => e
+        warn "Failed to fetch movies for cinema #{cinema_id}: #{e.message}"
+        {}
+      end
+    end
+  end
+end

data/lib/picturehouse_uk/internal/parser/address.rb

@@ -15,7 +15,7 @@ module PicturehouseUk
         # @note Uses the address naming from http://microformats.org/wiki/adr
         def address
           {
-            street_address:   array[1],
+            street_address:   address_lines[0],
             extended_address: extended_address,
             locality:         town,
             region:           region,
@@ -27,23 +27,38 @@ module PicturehouseUk
         private
 
         def array
-          @array ||= Array(@html.gsub(/\<.?p.?\>/, '').split('<br>'))
+          @array ||= begin
+            # Split on <br> tags first, then strip all HTML tags from each part
+            @html.split(/<br\s*\/?>/).map do |part|
+              # Strip all HTML tags and clean up whitespace
+              Nokogiri::HTML::DocumentFragment.parse(part).text.strip
+            end.reject(&:empty?)
+          end
+        end
+
+        # Skip the first element (cinema name) and return address lines
+        def address_lines
+          @address_lines ||= array[1..-1] || []
         end
 
         def extended_address
-          array.length > 5 ? array[2] : nil
+          address_lines.length > 4 ? address_lines[1] : nil
         end
 
         def postal_code
-          array[-1]
+          address_lines[-1]
         end
 
         def region
-          array[-2] == town ? nil : array[-2]
+          # If there are 4+ address lines (street, town, region, postcode),
+          # the region is at -2, otherwise nil
+          address_lines.length >= 4 ? address_lines[-2] : nil
         end
 
         def town
-          @town ||= array[0].to_s.split(', ')[-1]
+          # Town is always the second-to-last item (before postcode)
+          # unless there's a region, then it's third-to-last
+          @town ||= address_lines.length >= 4 ? address_lines[-3] : address_lines[-2]
         end
       end
     end

data/lib/picturehouse_uk/internal/parser/screenings.rb

@@ -3,170 +3,122 @@ module PicturehouseUk
   module Internal
     # @api private
     module Parser
-      # Parses screenings page into an array of hashes for an individual cinema
+      # Parses JSON API response into an array of screening hashes
       class Screenings
-        # css for a day of films & screenings
-        LISTINGS = '.listings li:not(.dark)'.freeze
-        DATE = '.nav-collapse.collapse'.freeze
-
         def initialize(cinema_id)
           @cinema_id = cinema_id
         end
 
-        # parse the cinema page into an array of screenings attributes
+        # Parse the JSON response into an array of screening attributes
         # @return [Array<Hash>]
         def to_a
-          doc.css(LISTINGS).flat_map do |node|
-            FilmWithShowtimes.new(node,
-                                  date_from_html(node.css(DATE).to_s)).to_a
-          end
+          return [] unless json_data['response'] == 'success'
+          return [] unless json_data['movies']
+
+          json_data['movies'].flat_map do |movie|
+            parse_movie(movie)
+          end.compact
         end
 
         private
 
-        def date_from_html(html)
-          if html =~ /listings-further-ahead-today/
-            Date.now
-          else
-            html.match(/listings-further-ahead-(\d{4})(\d{2})(\d{2})/) do |m|
-              Date.new(m[1].to_i, m[2].to_i, m[3].to_i)
-            end
-          end
-        end
+        def parse_movie(movie)
+          return [] if movie['show_times'].nil? || movie['show_times'].empty?
 
-        def doc
-          @doc ||= Nokogiri::HTML(page)
+          movie['show_times'].map do |timing|
+            parse_timing(movie, timing)
+          end.compact
         end
 
-        def page
-          @page ||= PicturehouseUk::Internal::Website.new.whats_on(@cinema_id)
-        end
-      end
-    end
-
-    # @api private
-    # collection of timings for a specific film
-    class FilmWithShowtimes
-      # film name css
-      NAME = '.top-mg-sm a'.freeze
-      # variants css
-      VARIANTS = '.film-times .col-xs-10'.freeze
-
-      def initialize(node, date)
-        @node = node
-        @date = date
-      end
-
-      # The film name
-      # @return [String]
-      def name
-        TitleSanitizer.new(raw_name).sanitized
-      end
+        def parse_timing(movie, timing)
+          # Skip sold out or unavailable screenings
+          return nil if timing['SoldoutStatus'] == 2
+          return nil unless timing['date_f'] && timing['time']
 
-      # Showings hashes
-      # @return [Array<Hash>]
-      def to_a
-        Array(@node.css(VARIANTS)).flat_map do |variant|
-          Variant.new(variant, @date).to_a.map do |hash|
-            {
-              film_name: name,
-              dimension: dimension
-            }.merge(hash)
+          # Skip advance booking restrictions if not advertised
+          if movie['ABgtToday'] == true && movie['AdvertiseAdvanceBookingDate'] == false
+            return nil
           end
-        end
-      end
-
-      private
 
-      def dimension
-        raw_name =~ /3d/i ? '3d' : '2d'
-      end
-
-      def raw_name
-        @raw_name ||= @node.css(NAME).children.first.to_s
-      end
-    end
-
-    # @api private
-    # variants can have multiple screenings
-    class Variant
-      SHOWTIMES  = '.btn'.freeze
-      VARIANT    = '.film-type-desc'.freeze
-      TRANSLATOR = {
-        'Big Scream'    => 'baby',
-        'IMAX'          => 'imax',
-        "Kids' Club"    => 'kids',
-        'NT Live'       => 'arts',
-        'Screen Arts'   => 'arts',
-        'Silver Screen' => 'senior'
-      }.freeze
-
-      def initialize(node, date)
-        @node = node
-        @date = date
-      end
+          {
+            film_name: sanitize_title(movie['Title']),
+            dimension: determine_dimension(movie['Title'], timing),
+            variant: extract_variants(timing),
+            booking_url: booking_url(timing),
+            starting_at: parse_datetime(timing['date_f'], timing['time'])
+          }
+        end
 
-      # Variant arrays
-      # @return [Array<Hash>]
-      def to_a
-        @node.css(SHOWTIMES).map do |node|
-          { variant: variant }.merge(Showtime.new(@node, @date).to_hash)
+        def sanitize_title(title)
+          TitleSanitizer.new(title).sanitized
         end
-      end
 
-      private
+        def determine_dimension(title, timing)
+          # Check title for 3D indicator
+          return '3d' if title =~ /3d/i
 
-      def variant
-        @variant ||= TRANSLATOR.select do |k, _|
-          variant_text.include?(k)
-        end.values.uniq
-      end
+          # Check timing attributes for 3D
+          if timing['SessionAttributesNames']
+            return '3d' if timing['SessionAttributesNames'].any? { |attr| attr =~ /3d/i }
+          end
 
-      def variant_text
-        @variant_text ||= @node.css(VARIANT).to_s
-      end
-    end
+          '2d'
+        end
 
-    # @api private
-    # parse an individual screening node
-    class Showtime
-      def initialize(node, date)
-        @node = node
-        @date = date
-      end
+        def extract_variants(timing)
+          return [] unless timing['SessionAttributesNames']
 
-      def to_hash
-        {
-          booking_url: booking_url,
-          starting_at: starting_at
-        }
-      end
+          variants = []
 
-      private
+          timing['SessionAttributesNames'].each do |attribute|
+            # Map known attributes to variant types
+            variants << 'baby' if attribute =~ /big scream/i
+            variants << 'imax' if attribute =~ /imax/i
+            variants << 'kids' if attribute =~ /kids|toddler/i
+            variants << 'arts' if attribute =~ /nt live|screen arts|rbo|roh|met opera/i
+            variants << 'senior' if attribute =~ /silver screen/i
+          end
 
-      def booking_url
-        return if href.nil? || href.empty?
-        "https://picturehouses.com#{href}"
-      end
+          variants.uniq.sort
+        end
 
-      def hour
-        split[0]
-      end
+        def booking_url(timing)
+          return nil unless timing['SessionId']
 
-      def href
-        @href ||= @node['href']
-      end
+          # Vista booking URL format
+          "https://web.picturehouses.com/order/showtimes/#{@cinema_id}-#{timing['SessionId']}/seats"
+        end
 
-      def min
-        split[1]
-      end
+        def parse_datetime(date_str, time_str)
+          # date_str format: "2024-10-30"
+          # time_str format: "19:30"
+          return nil unless date_str && time_str
+
+          date_parts = date_str.split('-').map(&:to_i)
+          time_parts = time_str.split(':').map(&:to_i)
+
+          # Create Time object in local timezone, then convert to UTC
+          Time.new(
+            date_parts[0], # year
+            date_parts[1], # month
+            date_parts[2], # day
+            time_parts[0], # hour
+            time_parts[1], # minute
+            0,             # second
+            '+00:00'       # UTC timezone
+          )
+        rescue StandardError => e
+          warn "Failed to parse datetime from #{date_str} #{time_str}: #{e.message}"
+          nil
+        end
 
-      def split
-        @split ||= @node.text.split('.').map(&:to_i)
-      end
+        def json_data
+          @json_data ||= api_client.get_movies(@cinema_id)
+        end
 
-      def starting_at
-        @starting_at ||= @date.to_time + (hour * 60 + min) * 60
+        def api_client
+          @api_client ||= Api.new
+        end
       end
     end
   end

data/lib/picturehouse_uk/internal/website.rb

@@ -12,18 +12,10 @@ module PicturehouseUk
         get("cinema/#{id}")
       end
 
-      # get the cinema screenings page for passed id
-      # @return [String]
-      def whats_on(id)
-        get("cinema/#{id}/Whats_On")
-      rescue OpenURI::HTTPError
-        ''
-      end
-
       # get the cinema contact information page for passed id
       # @return [String]
-      def info(id)
-        get("cinema/info/#{id}")
+      def information(id)
+        get("cinema/#{id}/information")
       rescue OpenURI::HTTPError
         ''
       end
@@ -34,12 +26,18 @@ module PicturehouseUk
         get(nil)
       end
 
+      # get the cinemas listing page
+      # @return [String]
+      def cinemas
+        get("cinema")
+      end
+
       private
 
       def get(path)
         # SSL verification doesn't work on picturehouses.com
-        open("https://www.picturehouses.com/#{path}",
-             ssl_verify_mode: OpenSSL::SSL::VERIFY_NONE).read
+        URI.open("https://www.picturehouses.com/#{path}",
+                 ssl_verify_mode: OpenSSL::SSL::VERIFY_NONE).read
       end
     end
   end

data/lib/picturehouse_uk/version.rb

@@ -1,6 +1,6 @@
 # Ruby interface for http://www.picturehouses.co.uk
-# @version 4.0.0
+# @version 5.0.0
 module PicturehouseUk
   # Gem version
-  VERSION = '4.0.0'.freeze
+  VERSION = '5.0.0'.freeze
 end

data/lib/picturehouse_uk.rb

@@ -3,6 +3,7 @@ require 'nokogiri'
 
 require_relative './picturehouse_uk/version'
 
+require_relative './picturehouse_uk/internal/api'
 require_relative './picturehouse_uk/internal/parser/address'
 require_relative './picturehouse_uk/internal/parser/screenings'
 require_relative './picturehouse_uk/internal/title_sanitizer'

data/picturehouse_uk.gemspec

@@ -12,15 +12,14 @@ Gem::Specification.new do |spec|
   spec.homepage      = ''
   spec.licenses      = %w(AGPL MIT)
 
-  spec.files         = `git ls-files`.split($RS)
+  spec.files         = `git ls-files`.split("\n")
   spec.executables   = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
   spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
   spec.require_paths = ['lib']
 
-  spec.add_development_dependency 'bundler', '~> 1.3'
-  spec.add_development_dependency 'codeclimate-test-reporter'
   spec.add_development_dependency 'minitest-reporters'
   spec.add_development_dependency 'rake'
+  spec.add_development_dependency 'simplecov'
   spec.add_development_dependency 'webmock'
 
   spec.add_runtime_dependency 'cinebase', '~> 3.0'