national-rail 0.1.0

data/lib/national-rail/journey_planner.rb

@@ -0,0 +1,157 @@
+require 'mechanize'
+require 'hpricot'
+require 'active_support'
+
+module NationalRail
+
+  class JourneyPlanner
+
+    class HpricotParser < Mechanize::Page
+      attr_reader :doc
+      def initialize(uri = nil, response = nil, body = nil, code = nil)
+        @doc = Hpricot(body)
+        super(uri, response, body, code)
+      end
+    end
+
+    class TimeParser
+      def initialize(date)
+        @date = date
+        @last_time = nil
+      end
+      def time(hours_and_minutes)
+        time = parse(hours_and_minutes)
+        if @last_time && (time < @last_time)
+          @date += 1
+          time = parse(hours_and_minutes)
+        end
+        @last_time = time
+        time
+      end
+      def parse(hours_and_minutes)
+        Time.zone.parse("#{hours_and_minutes} #{@date}")
+      end
+    end
+
+    class SummaryRow
+
+      attr_reader :departure_time, :number_of_changes
+
+      def initialize(agent, date, departure_time, number_of_changes, link)
+        @agent, @date = agent, date
+        @departure_time, @number_of_changes, @link = departure_time, number_of_changes, link
+      end
+
+      def details
+        @agent.transact do
+          parse_details(@link.click, @date)
+        end
+      end
+
+      def parse_details(page, date)
+        details = {}
+        description = (page.doc/"table#journeyLegDetails tbody tr.lastRow td[@colspan=6] div").first.inner_text.gsub(/\s+/, " ").strip
+        origins, destinations = (/from (.*) to (.*)/.match(description)[1,2]).map{ |s| s.split(",").map(&:strip) }
+        details[:origins], details[:destinations] = origins, destinations
+        parser = TimeParser.new(date)
+
+        origin_code = (page.doc/"td.origin abbr").inner_html.strip
+        departs_at = (page.doc/"td.leaving").inner_html.strip
+        details[:initial_stop] = {
+          :station_code => origin_code,
+          :departs_at => parser.time(departs_at)
+        }
+
+        details[:stops] = []
+        (page.doc/".callingpoints table > tbody > tr").each do |tr|
+          if (tr/".calling-points").length > 0
+            station_code = (tr/".calling-points > a > abbr").inner_html.strip
+            arrives_at = (tr/".arrives").inner_html.strip
+            departs_at = (tr/".departs").inner_html.strip
+            departs_at = arrives_at if arrives_at.present? && departs_at.blank?
+            arrives_at = departs_at if arrives_at.blank? && departs_at.present?
+            details[:stops] << {
+              :station_code => station_code,
+              :arrives_at => parser.time(arrives_at),
+              :departs_at => parser.time(departs_at)
+            }
+          end
+        end
+
+        destination_code = (page.doc/"td.destination abbr").inner_html.strip
+        arrives_at = (page.doc/"td.arriving").inner_html.strip
+        details[:final_stop] = {
+          :station_code => destination_code,
+          :arrives_at => parser.time(arrives_at)
+        }
+        details
+      rescue => e
+        filename = File.join(File.dirname(__FILE__), "..", "..", "details-error.html")
+        File.open(filename, "w") { |f| f.write(page.parser.to_html) }
+        raise e
+      end
+    end
+
+    def initialize
+      Time.zone ||= 'London'
+      @agent = Mechanize.new
+      @agent.pluggable_parser.html = HpricotParser
+      @agent.user_agent_alias = "Mac FireFox"
+    end
+
+    def plan(options = {})
+      summary_rows = []
+      @agent.get("http://www.nationalrail.co.uk/") do |home_page|
+        times_page = home_page.form_with(:action => "http://ojp.nationalrail.co.uk/en/s/planjourney/plan") do |form|
+          form["jpState"] = "single"
+          form["commandName"] = "journeyPlannerCommand"
+          form["from.searchTerm"] = options[:from]
+          form["to.searchTerm"] = options[:to]
+          form["timeOfOutwardJourney.arrivalOrDeparture"] = "DEPART"
+          form["timeOfOutwardJourney.monthDay"] = options[:time].strftime("%d/%m/%Y")
+          form["timeOfOutwardJourney.hour"] = options[:time].strftime("%H")
+          form["timeOfOutwardJourney.minute"] = options[:time].strftime("%M")
+          # form["timeOfReturnJourney.arrivalOrDeparture"] = "DEPART"
+          # form["timeOfReturnJourney.monthDay"] = "Today"
+          # form["timeOfReturnJourney.hour"] = "15"
+          # form["timeOfReturnJourney.minute"] = "0"
+          form["viaMode"] = "VIA"
+          form["via.searchTerm"] = ""
+          form["offSetOption"] = "0"
+          form["_reduceTransfers"] = "on"
+          form["operatorMode"] = "SHOW"
+          form["operator.code"] = ""
+          form["_lookForSleeper"] = "on"
+          form["_directTrains"] = "on"
+          form["_includeOvertakenTrains"] = "on"
+        end.click_button
+
+        if (times_page.doc/"error-message").any?
+          raise (times_page.doc/"error-message").first.inner_text.gsub(/\s+/, " ").strip
+        end
+
+        date = Date.parse((times_page.doc/".journey-details span:nth-child(0)").first.inner_text.gsub(/\s+/, " ").gsub(/\+ 1 day/, '').strip)
+
+        (times_page.doc/"table#outboundJourneyTable > tbody > tr:not(.status):not(.changes)").each do |tr|
+
+          if (tr.attributes["class"] == "day-heading")
+            date = Date.parse((tr/"th > p > span").first.inner_text.strip)
+            next
+          end
+
+          departure_time = TimeParser.new(date).time((tr/"td.leaving").inner_text.strip)
+          number_of_changes = (tr/"td:nth-child(6)").inner_text.strip
+          anchor = (tr/"a[@id^=journeyOption]").first
+          link = times_page.links.detect { |l| l.attributes["id"] == anchor["id"] }
+
+          summary_rows << SummaryRow.new(@agent, date, departure_time, number_of_changes, link)
+        end
+      end
+      summary_rows
+    rescue => e
+      filename = File.join(File.dirname(__FILE__), "..", "..", "summary-error.html")
+      File.open(filename, "w") { |f| f.write(@agent.current_page.parser.to_html) }
+      raise e
+    end
+  end
+end

data/lib/national-rail/live_departure_boards.rb

@@ -0,0 +1,107 @@
+require 'mechanize'
+require 'hpricot'
+
+module NationalRail
+
+  class LiveDepartureBoards
+
+    class HpricotParser < Mechanize::Page
+      attr_reader :doc
+      def initialize(uri = nil, response = nil, body = nil, code = nil)
+        @doc = Hpricot(body)
+        super(uri, response, body, code)
+      end
+    end
+    
+    class SummaryRow
+      attr_reader :attributes
+      def initialize(agent, attributes)
+        @agent, @attributes = agent, attributes
+      end
+      def [](key)
+        @attributes[key]
+      end
+      def details
+        page = nil
+        @agent.transact do
+          page = @attributes[:details][:link].click
+          station = nil
+          (page.doc/"table.arrivaltable tbody tr").map do |tr|
+            tds = (tr/"td")
+            tds.unshift(station) if tds.length < 4
+            station = tds[0]
+            station_url = (station/"a").first.attributes["href"]
+            station_code = /station_query\=(.*)$/.match(station_url)[1]
+            deparr = (tds[1]/".deparr").remove
+            direction = (deparr.inner_text.strip == "Dep") ? "departure" : "arrival"
+            {
+              :station => {
+                :name => (station/"a").inner_text.strip,
+                :url => station_url,
+                :code => station_code
+              },
+              :timetabled => {
+                :time => tds[1].inner_text.strip,
+                :direction => direction
+              },
+              :expected => {
+                :time => tds[2].inner_text.strip
+              },
+              :actual => {
+                :time => tds[3].inner_text.strip
+              }
+            }
+          end
+        end
+      rescue => e
+        File.open("error.html", "w") { |f| f.write(page.parser.to_html) } if page
+        raise e
+      end
+    end
+
+    def initialize
+      @agent = Mechanize.new
+      @agent.pluggable_parser.html = HpricotParser
+    end
+
+    def summary(options = {})
+      summary_url = "http://realtime.nationalrail.co.uk/ldb/sumdep.aspx"
+      page = @agent.get(summary_url, "T" => options[:from], "S" => options[:to])
+      (page.doc/"table.arrivaltable tbody tr").map do |tr|
+        destination = (tr/"td:nth-child(1)")
+        destination_url = (destination/"a").first.attributes["href"]
+        destination_code = /station_query\=(.*)$/.match(destination_url)[1]
+        operator = (tr/"td:nth-child(5)")
+        operator_url = (operator/"a").first.attributes["href"]
+        operator_code = /atocCode\=(.*)$/.match(operator_url)[1]
+        details = (tr/"td:nth-child(6)")
+        details_url = (details/"a").first.attributes["href"]
+        details_link = page.links.detect { |l| l.attributes["href"] == details_url }
+        SummaryRow.new(@agent,
+          :destination => {
+            :name => destination.inner_text,
+            :url => destination_url,
+            :code => destination_code
+          },
+          :platform => (tr/"td:nth-child(2)").inner_text,
+          :timetabled_at => (tr/"td:nth-child(3)").inner_text,
+          :expected_at => (tr/"td:nth-child(4)").inner_text,
+          :operator => {
+            :name => operator.inner_text,
+            :url => operator_url,
+            :code => operator_code
+          },
+          :details => {
+            :url => details_url,
+            :link => details_link
+          }
+        )
+      end
+    rescue => e
+      File.open("error.html", "w") { |f| f.write(@agent.current_page.parser.to_html) }
+      raise e
+    end
+
+  end
+
+end

data/lib/national-rail/station_list.rb

@@ -0,0 +1,32 @@
+require 'mechanize'
+require 'hpricot'
+
+module NationalRail
+
+  class StationList
+
+    class HpricotParser < Mechanize::Page
+      attr_reader :doc
+      def initialize(uri = nil, response = nil, body = nil, code = nil)
+        @doc = Hpricot(body)
+        super(uri, response, body, code)
+      end
+    end
+
+    def initialize
+      @agent = Mechanize.new
+      @agent.pluggable_parser.html = HpricotParser
+    end
+
+    def each
+      @agent.get("http://www.nationalrail.co.uk/stations/codes/") do |stations_page|
+        (stations_page.doc/"table:nth-child(0) tbody tr").each do |tr|
+          name, code = (tr/"td").map { |td| td.inner_text }
+          yield(name, code) unless name == "DUMMY TEST"
+        end
+      end
+    end
+
+  end
+
+end

metadata

@@ -0,0 +1,107 @@
+--- !ruby/object:Gem::Specification 
+name: national-rail
+version: !ruby/object:Gem::Version 
+  prerelease: false
+  segments: 
+  - 0
+  - 1
+  - 0
+  version: 0.1.0
+platform: ruby
+authors: 
+- James Mead
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2010-05-15 00:00:00 +01:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: mechanize
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 1
+        - 0
+        - 0
+        version: 1.0.0
+  type: :runtime
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
+  name: hpricot
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 0
+        - 8
+        - 2
+        version: 0.8.2
+  type: :runtime
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: activesupport
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 2
+        - 3
+        - 5
+        version: 2.3.5
+  type: :runtime
+  version_requirements: *id003
+description: 
+email: james@floehopper.org
+executables: []
+
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- lib/national-rail/index.html
+- lib/national-rail/journey_planner.rb
+- lib/national-rail/live_departure_boards.rb
+- lib/national-rail/station_list.rb
+- lib/national-rail.rb
+has_rdoc: true
+homepage: 
+licenses: []
+
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.3.6
+signing_key: 
+specification_version: 3
+summary: A Ruby API for the National Rail website
+test_files: []
+