earth 0.3.13 → 0.3.14

data/.document

@@ -0,0 +1,5 @@
+README.rdoc
+lib/**/*.rb
+bin/*
+features/**/*.feature
+LICENSE

data/.gitignore

@@ -0,0 +1,23 @@
+## MAC OS
+.DS_Store
+
+## TEXTMATE
+*.tmproj
+tmtags
+
+## EMACS
+*~
+\#*
+.\#*
+
+## VIM
+*.swp
+
+## PROJECT::GENERAL
+coverage
+rdoc
+pkg
+doc
+
+## PROJECT::SPECIFIC
+Gemfile.lock

data/Gemfile

@@ -1,9 +1,10 @@
-gem 'data_miner', :path => ENV['LOCAL_DATA_MINER'] if ENV['LOCAL_DATA_MINER']
-gem 'sniff', :path => ENV['LOCAL_SNIFF'] if ENV['LOCAL_SNIFF']
-
 source :rubygems
 
-gemspec :path => '.'
+# Specify your gem's dependencies in earth.gemspec
+gemspec
+
+gem 'data_miner', :path => ENV['LOCAL_DATA_MINER'] if ENV['LOCAL_DATA_MINER']
+gem 'sniff', :path => ENV['LOCAL_SNIFF'] if ENV['LOCAL_SNIFF']
 
 if RUBY_VERSION =~ /^1\.8/
   gem 'fastercsv'

data/Rakefile

@@ -0,0 +1,27 @@
+require 'bundler'
+Bundler.setup
+Bundler::GemHelper.install_tasks
+
+require 'bueller'
+Bueller::Tasks.new
+
+begin
+  require 'rake/rdoctask'
+  Rake::RDocTask.new do |rdoc|
+    rdoc.rdoc_dir = 'rdoc'
+    rdoc.title = 'emitter'
+    rdoc.options << '--line-numbers' << '--inline-source'
+    rdoc.rdoc_files.include('README*')
+    rdoc.rdoc_files.include('lib/**/*.rb')
+  end
+rescue LoadError
+   puts "Rdoc is not available"
+end
+
+require 'sniff'
+require 'sniff/rake_tasks'
+Sniff::RakeTasks.define_tasks do |t|
+  t.earth_domains = :all
+  t.cucumber = true
+  t.rspec = true
+end

data/earth.gemspec

@@ -0,0 +1,41 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path("../lib", __FILE__)
+require "earth/version"
+
+Gem::Specification.new do |s|
+  s.name        = "earth"
+  s.version     = Earth::VERSION
+  s.date = "2011-01-27"
+  s.platform    = Gem::Platform::RUBY
+  s.authors = ["Seamus Abshere", "Derek Kastner", "Andy Rossmeissl"]
+  s.email = %q{andy@rossmeissl.net}
+  s.homepage    = "https://github.com/brighterplanet/earth"
+  s.summary     = %Q{Land, sky, and sea}
+  s.description = %Q{An earth-simulation environment with ActiveRecord models and data}
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.extra_rdoc_files = [
+    "LICENSE",
+     "LICENSE-PREAMBLE",
+     "README.markdown"
+  ]
+  s.require_paths = ["lib"]
+  
+  s.rdoc_options = ["--charset=UTF-8"]
+  s.require_paths = ["lib"]
+  s.rubygems_version = %q{1.3.7}
+
+  s.add_runtime_dependency 'activerecord', '~> 3'
+  s.add_runtime_dependency 'data_miner', '~> 1.1.1'
+  s.add_runtime_dependency 'errata'
+  s.add_runtime_dependency 'falls_back_on', '>= 0.0.3'
+  s.add_runtime_dependency 'geokit', '>= 1.5.0'
+  s.add_runtime_dependency 'cohort_scope', '>= 0.0.7'
+  s.add_runtime_dependency 'conversions', '>= 1.4.5'
+  s.add_runtime_dependency 'weighted_average', '>= 0.0.4'
+  s.add_runtime_dependency 'loose_tight_dictionary', '>= 0.0.8'
+  s.add_development_dependency 'sniff', '~> 0.5.3'
+  s.add_development_dependency 'bueller'
+end

data/features/support/env.rb

@@ -3,5 +3,8 @@ require 'bundler'
 
 Bundler.setup
 
+require 'data_miner'
+DataMiner.logger = Logger.new(nil)
+
 require 'sniff'
 Sniff.init File.expand_path('../..', File.dirname(__FILE__)), :earth => :all, :cucumber => true

data/lib/earth.rb

@@ -8,6 +8,8 @@ require 'falls_back_on'
 require 'falls_back_on/active_record_ext'
 require 'weighted_average'
 require 'loose_tight_dictionary'
+require 'slither'
+require 'errata'
 
 # The earth module is an interface for establishing a taps server (used to fetch 
 # data) and for loading data models from various domains.
@@ -15,7 +17,7 @@ module Earth
   extend self
 
   def taps_server
-    @taps_server || 'http://carbon:neutral@data.brighterplanet.com:5001'
+    @taps_server || 'http://carbon:neutral@data.brighterplanet.com:5000'
   end
 
   # taps_server is a URL. See the data_miner gem docs

data/lib/earth/air/aircraft/data_miner.rb

@@ -126,8 +126,8 @@ Aircraft.class_eval do
     # ('A'..'Z').each do |letter|
     #   import( "ICAO manufacturers and names for aircraft with an ICAO code starting with the letter #{letter} from the FAA",
     #           :url => "http://www.faa.gov/air_traffic/publications/atpubs/CNT/5-2-#{letter}.htm",
-    #           :errata => Errata.new(:url => 'http://spreadsheets.google.com/pub?key=tObVAGyqOkCBtGid0tJUZrw',
-    #                                 :responder => Aircraft::Guru.new),
+    #           :errata => { :url => 'http://spreadsheets.google.com/pub?key=tObVAGyqOkCBtGid0tJUZrw',
+    #                        :responder => Aircraft::Guru.new },
     #           :encoding => 'windows-1252',
     #           :row_xpath => '//table/tr[2]/td/table/tr',
     #           :column_xpath => 'td' ) do
@@ -146,7 +146,7 @@ Aircraft.class_eval do
     
     # import "aircraft BTS names",
     #        :url => 'http://www.transtats.bts.gov/Download_Lookup.asp?Lookup=L_AIRCRAFT_TYPE',
-    #        :errata => Errata.new(:url => 'https://spreadsheets.google.com/pub?key=0AoQJbWqPrREqdEZ2d3JQMzV5T1o1T3JmVlFyNUZxdEE&hl=en&single=true&gid=0&output=csv') do
+    #        :errata => { :url => 'https://spreadsheets.google.com/pub?key=0AoQJbWqPrREqdEZ2d3JQMzV5T1o1T3JmVlFyNUZxdEE&hl=en&single=true&gid=0&output=csv'} do
     #   key 'bts_code',   :field_name => 'Code'
     #   store 'bts_name', :field_name => 'Description'
     # end

data/lib/earth/air/airline/data_miner.rb

@@ -1,3 +1,5 @@
+require 'errata'
+
 Airline.class_eval do
   class Airline::Guru
     # needed by errata
@@ -27,8 +29,7 @@ Airline.class_eval do
     
     import "the T100 AIRLINE_ID lookup table, which also includes IATA codes",
            :url => 'http://www.transtats.bts.gov/Download_Lookup.asp?Lookup=L_AIRLINE_ID',
-           :errata => Errata.new(:url => 'http://static.brighterplanet.com/science/data/transport/air/airlines/errata.csv',
-                                 :responder => Airline::Guru.new) do
+           :errata => { :url => 'http://static.brighterplanet.com/science/data/transport/air/airlines/errata.csv', :responder => Airline::Guru.new } do
       key 'iata_code', :field_name => 'Description', :split => { :pattern => /:/, :keep => 1 }
       store 'dot_airline_id_code', :field_name => 'Code'
       store 'name', :field_name => 'Description', :split => { :pattern => /:/, :keep => 0 }

data/lib/earth/automobile/automobile_fuel_type/data_miner.rb

@@ -48,7 +48,7 @@ AutomobileFuelType.class_eval do
       AutomobileFuelType.all.each do |fuel_type|
         current_name = fuel_type.fuel_name_for_efs
         
-        unless current_name.nil?
+        if current_name.present?
           co2_factor_sql = fuel_years.project(fuel_years[:co2_emission_factor]).where(fuel_years[:fuel_common_name].eq(fuel_types[:fuel_name_for_efs]).and(fuel_years[:year].eq(latest_year))).to_sql
           ch4_factor_sql = AutomobileTypeFuelYear.weighted_average_relation(:ch4_emission_factor, :weighted_by => :total_travel).where(type_fuel_years[:fuel_common_name].eq(fuel_types[:fuel_name_for_efs]).and(type_fuel_years[:year].eq(latest_year))).to_sql
           n2o_factor_sql = AutomobileTypeFuelYear.weighted_average_relation(:n2o_emission_factor, :weighted_by => :total_travel).where(type_fuel_years[:fuel_common_name].eq(fuel_types[:fuel_name_for_efs]).and(type_fuel_years[:year].eq(latest_year))).to_sql

data/lib/earth/automobile/automobile_make_fleet_year/data_miner.rb

@@ -13,7 +13,7 @@ AutomobileMakeFleetYear.class_eval do
     
     import "annual corporate average fuel economy data for domestic and imported vehicle fleets from the NHTSA",
            :url => 'https://spreadsheets.google.com/pub?key=0AoQJbWqPrREqdEdXWXB6dkVLWkowLXhYSFVUT01sS2c&hl=en&single=true&gid=0&output=csv',
-           :errata => 'http://static.brighterplanet.com/science/data/transport/automobiles/make_fleet_years/errata.csv',
+           :errata => { 'url' => 'http://static.brighterplanet.com/science/data/transport/automobiles/make_fleet_years/errata.csv' },
            :select => lambda { |row| row['volume'].to_i > 0 } do
       key   'name', :synthesize => lambda { |row| [ row['manufacturer_name'], row['year_content'], row['fleet'][2,2] ].join ' ' }
       store 'make_year_name', :synthesize => lambda { |row| [ row['manufacturer_name'], row['year_content'] ].join ' ' }

data/lib/earth/automobile/automobile_make_model_year_variant/data_miner.rb

@@ -106,11 +106,58 @@ AutomobileMakeModelYearVariant.class_eval do
   }
   
   class AutomobileMakeModelYearVariant::ParserB
-    attr_accessor :year
+    ::Slither.define :fuel_economy_guide_b do |d|
+      d.rows do |row|
+        row.trap { true } # there's only one section
+        row.column 'active_year'      , 4,    :type => :integer  #   ACTIVE YEAR
+        row.column 'state_code'       , 1,    :type => :string  #   STATE CODE:  F=49-STATE,C=CALIFORNIA
+        row.column 'carline_clss'     , 2,    :type => :integer  #   CARLINE CLASS CODE
+        row.column 'carline_mfr_code' , 3,    :type => :integer  #   CARLINE MANUFACTURER CODE
+        row.column 'carline_name'     , 28,   :type => :string  #   CARLINE NAME
+        row.column 'disp_cub_in'      , 4,    :type => :integer   #  DISP CUBIC INCHES
+        row.column 'fuel_system'      , 2,    :type => :string   #  FUEL SYSTEM: 'FI' FOR FUEL INJECTION, 2-DIGIT INTEGER VALUE FOR #OF VENTURIES IF CARBURETOR SYSTEM.
+        row.column 'model_trans'      , 6,    :type => :string   #  TRANSMISSION TYPE
+        row.column 'no_cyc'           , 2,    :type => :integer   #  NUMBER OF ENGINE CYLINDERS
+        row.column 'date_time'        , 12,   :type => :string  #   DATE AND TIME RECORD ENTERED -YYMMDDHHMMSS (YEAR, MONTH, DAY, HOUR, MINUTE, SECOND)
+        row.column 'release_date'     , 6,    :type => :string   #  RELEASE DATE - YYMMDD (YEAR, MONTH, DAY)
+        row.column 'vi_mfr_code'      , 3,    :type => :integer   #  VI MANUFACTURER CODE
+        row.column 'carline_code'     , 5,    :type => :integer   #  CARLINE CODE
+        row.column 'basic_eng_id'     , 5,    :type => :integer   #  BASIC ENGINE INDEX
+        row.column 'carline_mfr_name' , 32,   :type => :string  #   CARLINE MANUFACTURER NAME
+        row.column 'suppress_code'    , 1,    :type => :integer    # SUPPRESSION CODE (NO SUPPRESSED RECORD IF FOR PUBLIC ACCESS)
+        row.column 'est_city_mpg'     , 3,    :type => :integer    # ESTIMATED (CITY) MILES PER GALLON - 90% OF UNADJUSTED VALUE
+        row.spacer 2
+        row.column 'highway_mpg'      , 3,    :type => :integer    # ESTIMATED (HWY) MILES PER GALLON - 78% OF UNADJUSTED VALUE
+        row.spacer 2
+        row.column 'combined_mpg'     , 3,    :type => :integer    # COMBINED MILES PER GALLON
+        row.spacer 2
+        row.column 'unadj_city_mpg'   , 3,    :type => :integer    # UNADJUSTED  CITY MILES PER GALLON
+        row.spacer 2
+        row.column 'unadj_hwy_mpg'    , 3,    :type => :integer    # UNADJUSTED HIGHWAY MILES PER GALLON
+        row.spacer 2
+        row.column 'unadj_comb_mpg'   , 3,    :type => :integer    # UNADJUSTED COMBINED MILES PER GALLON
+        row.spacer 2
+        row.column 'ave_anl_fuel'     , 6,    :type => :integer    # "$" in col 147, Annual Fuel Cost starting col 148 in I5
+        row.column 'opt_disp'         , 8,    :type => :string    # OPTIONAL DISPLACEMENT
+        row.column 'engine_desc1'     , 10,   :type => :string   #  ENGINE DESCRIPTION 1
+        row.column 'engine_desc2'     , 10,   :type => :string   #  ENGINE DESCRIPTION 2
+        row.column 'engine_desc3'     , 10,   :type => :string   #  ENGINE DESCRIPTION 3
+        row.column 'body_type_2d'     , 10,   :type => :string   #  BODY TYPE 2 DOOR - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM '2DR-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
+        row.column 'body_type_4d'     , 10,   :type => :string   #  BODY TYPE 4 DOOR - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM '4DR-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
+        row.column 'body_type_hbk'    , 10,   :type => :string   #  BODY TYPE HBK    - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM 'HBK-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
+        row.column 'puerto_rico'      , 1,    :type => :string    # '*' IF FOR PUERTO RICO SALES ONLY
+        row.column 'overdrive'        , 4,    :type => :string    # OVERDRIVE:  ' OD ' FOR OVERDRIVE, 'EOD ' FOR ELECTRICALLY OPERATED OVERDRIVE AND 'AEOD' FOR AUTOMATIC OVERDRIVE
+        row.column 'drive_system'     , 3,    :type => :string    # FWD=FRONT WHEEL DRIVE, RWD=REAR,  4WD=4-WHEEL
+        row.column 'filler'           , 1,    :type => :string    # NOT USED
+        row.column 'fuel_type'        , 1,    :type => :string    # R=REGULAR(UNLEADED), P=PREMIUM,  D=DIESEL
+        row.column 'trans_desc'       , 15,   :type => :string   #  TRANSMISSION DESCRIPTORS
+      end
+    end
+    attr_reader :year
     def initialize(options = {})
-      @year = options[:year]
+      options = options.stringify_keys
+      @year = options['year']
     end
-  
     def apply(row)
       row.merge!({
         'make'           => row['carline_mfr_name'], # make it line up with the errata
@@ -125,7 +172,6 @@ AutomobileMakeModelYearVariant.class_eval do
       })
       row
     end
-  
     def _displacement(str)
       str = str.gsub(/[\(\)]/, '').strip
       if str =~ /^(.+)L$/
@@ -134,73 +180,14 @@ AutomobileMakeModelYearVariant.class_eval do
         $1.to_f / 1000
       end
     end
-  
-    def add_hints!(bus)
-      bus[:format] = :fixed_width
-      bus[:cut] = '13-' if year == 1995
-      bus[:schema_name] = :fuel_economy_guide_b
-      bus[:select] = lambda do |row|
-        (row['suppress_code'].blank? or row['suppress_code'].to_f == 0) and row['state_code'] == 'F'
-      end
-      Slither.define :fuel_economy_guide_b do |d|
-        d.rows do |row|
-          row.trap { true } # there's only one section
-          row.column 'active_year'      , 4,    :type => :integer  #   ACTIVE YEAR
-          row.column 'state_code'       , 1,    :type => :string  #   STATE CODE:  F=49-STATE,C=CALIFORNIA
-          row.column 'carline_clss'     , 2,    :type => :integer  #   CARLINE CLASS CODE
-          row.column 'carline_mfr_code' , 3,    :type => :integer  #   CARLINE MANUFACTURER CODE
-          row.column 'carline_name'     , 28,   :type => :string  #   CARLINE NAME
-          row.column 'disp_cub_in'      , 4,    :type => :integer   #  DISP CUBIC INCHES
-          row.column 'fuel_system'      , 2,    :type => :string   #  FUEL SYSTEM: 'FI' FOR FUEL INJECTION, 2-DIGIT INTEGER VALUE FOR #OF VENTURIES IF CARBURETOR SYSTEM.
-          row.column 'model_trans'      , 6,    :type => :string   #  TRANSMISSION TYPE
-          row.column 'no_cyc'           , 2,    :type => :integer   #  NUMBER OF ENGINE CYLINDERS
-          row.column 'date_time'        , 12,   :type => :string  #   DATE AND TIME RECORD ENTERED -YYMMDDHHMMSS (YEAR, MONTH, DAY, HOUR, MINUTE, SECOND)
-          row.column 'release_date'     , 6,    :type => :string   #  RELEASE DATE - YYMMDD (YEAR, MONTH, DAY)
-          row.column 'vi_mfr_code'      , 3,    :type => :integer   #  VI MANUFACTURER CODE
-          row.column 'carline_code'     , 5,    :type => :integer   #  CARLINE CODE
-          row.column 'basic_eng_id'     , 5,    :type => :integer   #  BASIC ENGINE INDEX
-          row.column 'carline_mfr_name' , 32,   :type => :string  #   CARLINE MANUFACTURER NAME
-          row.column 'suppress_code'    , 1,    :type => :integer    # SUPPRESSION CODE (NO SUPPRESSED RECORD IF FOR PUBLIC ACCESS)
-          row.column 'est_city_mpg'     , 3,    :type => :integer    # ESTIMATED (CITY) MILES PER GALLON - 90% OF UNADJUSTED VALUE
-          row.spacer 2
-          row.column 'highway_mpg'      , 3,    :type => :integer    # ESTIMATED (HWY) MILES PER GALLON - 78% OF UNADJUSTED VALUE
-          row.spacer 2
-          row.column 'combined_mpg'     , 3,    :type => :integer    # COMBINED MILES PER GALLON
-          row.spacer 2
-          row.column 'unadj_city_mpg'   , 3,    :type => :integer    # UNADJUSTED  CITY MILES PER GALLON
-          row.spacer 2
-          row.column 'unadj_hwy_mpg'    , 3,    :type => :integer    # UNADJUSTED HIGHWAY MILES PER GALLON
-          row.spacer 2
-          row.column 'unadj_comb_mpg'   , 3,    :type => :integer    # UNADJUSTED COMBINED MILES PER GALLON
-          row.spacer 2
-          row.column 'ave_anl_fuel'     , 6,    :type => :integer    # "$" in col 147, Annual Fuel Cost starting col 148 in I5
-          row.column 'opt_disp'         , 8,    :type => :string    # OPTIONAL DISPLACEMENT
-          row.column 'engine_desc1'     , 10,   :type => :string   #  ENGINE DESCRIPTION 1
-          row.column 'engine_desc2'     , 10,   :type => :string   #  ENGINE DESCRIPTION 2
-          row.column 'engine_desc3'     , 10,   :type => :string   #  ENGINE DESCRIPTION 3
-          row.column 'body_type_2d'     , 10,   :type => :string   #  BODY TYPE 2 DOOR - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM '2DR-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
-          row.column 'body_type_4d'     , 10,   :type => :string   #  BODY TYPE 4 DOOR - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM '4DR-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
-          row.column 'body_type_hbk'    , 10,   :type => :string   #  BODY TYPE HBK    - IF THE BODY TYPE APPLIES IT WILL TAKE THE FORM 'HBK-PPP/LL' WHERE PPP=PASSENGER INTERIOR VOLUME AND LL=LUGGAGE INTERIOR VOLUME.
-          row.column 'puerto_rico'      , 1,    :type => :string    # '*' IF FOR PUERTO RICO SALES ONLY
-          row.column 'overdrive'        , 4,    :type => :string    # OVERDRIVE:  ' OD ' FOR OVERDRIVE, 'EOD ' FOR ELECTRICALLY OPERATED OVERDRIVE AND 'AEOD' FOR AUTOMATIC OVERDRIVE
-          row.column 'drive_system'     , 3,    :type => :string    # FWD=FRONT WHEEL DRIVE, RWD=REAR,  4WD=4-WHEEL
-          row.column 'filler'           , 1,    :type => :string    # NOT USED
-          row.column 'fuel_type'        , 1,    :type => :string    # R=REGULAR(UNLEADED), P=PREMIUM,  D=DIESEL
-          row.column 'trans_desc'       , 15,   :type => :string   #  TRANSMISSION DESCRIPTORS
-        end
-      end
-    end
   end
+  
   class AutomobileMakeModelYearVariant::ParserC
-    attr_accessor :year
+    attr_reader :year
     def initialize(options = {})
-      @year = options[:year]
-    end
-    
-    def add_hints!(bus)
-      # File will decide format based on filename
+      options = options.stringify_keys
+      @year = options['year']
     end
-    
     def apply(row)
       row.merge!({
         'make'           => row['Manufacturer'], # make it line up with the errata
@@ -216,15 +203,13 @@ AutomobileMakeModelYearVariant.class_eval do
       row
     end
   end
+  
   class AutomobileMakeModelYearVariant::ParserD
-    attr_accessor :year
+    attr_reader :year
     def initialize(options = {})
-      @year = options[:year]
+      options = options.stringify_keys
+      @year = options['year']
     end
-    
-    def add_hints!(bus)
-    end
-    
     def apply(row)
       row.merge!({
         'make'           => row['MFR'],          # make it line up with the errata
@@ -240,6 +225,7 @@ AutomobileMakeModelYearVariant.class_eval do
       row
     end
   end
+  
   class AutomobileMakeModelYearVariant::ParserE
     OLD_FUEL_CODES = {
       'CNG' => 'C',
@@ -248,15 +234,11 @@ AutomobileMakeModelYearVariant.class_eval do
       'GP' => 'P',
       'GPR' => 'P'
     }
-    
-    attr_accessor :year
+    attr_reader :year
     def initialize(options = {})
-      @year = options[:year]
+      options = options.stringify_keys
+      @year = options['year']
     end
-    
-    def add_hints!(bus)
-    end
-    
     def apply(row)
       row.merge!({
         'make'           => row['Division'],          # make it line up with the errata
@@ -317,12 +299,17 @@ AutomobileMakeModelYearVariant.class_eval do
     # FIXME TODO one Jaguar in the 1990 FEG has no model name
     # FIXME TODO 14 records in the 1995 FEG are missing fuel efficiencies
     (85..97).each do |yy|
+    # [85, 95, 96].each do |yy|
       filename = (yy == 96) ? "#{yy}MFGUI.ASC" : "#{yy}MFGUI.DAT"
       import("19#{ yy } Fuel Economy Guide",
              :url => "http://www.fueleconomy.gov/FEG/epadata/#{yy}mfgui.zip",
+             :format => :fixed_width,
+             :cut => ((yy == 95) ? '13-' : nil),
+             :schema_name => :fuel_economy_guide_b,
+             :select => lambda { |row| (row['suppress_code'].blank? or row['suppress_code'].to_f == 0) and row['state_code'] == 'F' },
              :filename => filename,
              :transform => { :class => AutomobileMakeModelYearVariant::ParserB, :year => "19#{yy}".to_i },
-             :errata => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv') do
+             :errata => { 'url' => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv' }) do
         key   'row_hash'
         store 'name', :field_name => 'model'
         store 'make_name', :field_name => 'make'
@@ -360,7 +347,7 @@ AutomobileMakeModelYearVariant.class_eval do
     }.sort { |a, b| a.first <=> b.first }.each do |year, options|
       import "#{ year } Fuel Economy Guide",
              options.merge(:transform => { :class => AutomobileMakeModelYearVariant::ParserC, :year => year },
-                           :errata => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv') do
+                           :errata => { 'url' => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv' }) do
         key   'row_hash'
         store 'name', :field_name => 'model'
         store 'make_name', :field_name => 'make'
@@ -393,7 +380,7 @@ AutomobileMakeModelYearVariant.class_eval do
     }.sort { |a, b| a.first <=> b.first }.each do |year, options|
       import "#{ year } Fuel Economy Guide",
              options.merge(:transform => { :class => AutomobileMakeModelYearVariant::ParserD, :year => year },
-                           :errata => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv') do
+                           :errata => { 'url' => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv' }) do
         key   'row_hash'
         store 'name', :field_name => 'model'
         store 'make_name', :field_name => 'make'
@@ -422,7 +409,7 @@ AutomobileMakeModelYearVariant.class_eval do
     }.sort { |a, b| a.first <=> b.first }.each do |year, options|
       import "#{ year } Fuel Economy Guide",
              options.merge(:transform => { :class => AutomobileMakeModelYearVariant::ParserE, :year => year },
-                           :errata => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv') do
+                           :errata => { 'url' => 'http://static.brighterplanet.com/science/data/transport/automobiles/fuel_economy_guide/errata.csv' }) do
         key   'row_hash'
         store 'name', :field_name => 'model'
         store 'make_name', :field_name => 'make'

data/lib/earth/pet/species.rb

@@ -4,7 +4,7 @@ class Species < ActiveRecord::Base
   scope :thoroughly_researched, :conditions => 'marginal_dietary_requirement IS NOT NULL'
   
   falls_back_on :diet_emission_intensity => lambda { weighted_average :diet_emission_intensity, :weighted_by => :population }, # kg CO2 / joule
-                :marginal_dietary_requirement => lambda { (thoroughly_researched.map(&:weighted_diet_size).sum / thoroughly_researched.sum(:population) ) / thoroughly_researched.weighted_average(:weight, :weighted_by => :population)}, # joules
+                :marginal_dietary_requirement => lambda { Species.marginal_dietary_requirement_fallback },
                 :fixed_dietary_requirement => 0, # force a zero intercept to be respectful of our tiny tiny animal friends
                 :weight => lambda { weighted_average :weight, :weighted_by => :population } # kg
 
@@ -16,6 +16,15 @@ class Species < ActiveRecord::Base
     def [](name)
       find_by_name name.to_s
     end
+
+    def marginal_dietary_requirement_fallback
+      total_diet_size = thoroughly_researched.map(&:weighted_diet_size).sum
+      total_population = thoroughly_researched.sum(:population)
+      return 0.0 unless total_population > 0.0
+      average_weight = thoroughly_researched.weighted_average(:weight, :weighted_by => :population)
+      return 0.0 unless average_weight > 0.0
+      (total_diet_size / total_population) / average_weight
+    end
   end
   
   def diet_size