slaw 0.4.1 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 017be61dfca724d0875193ce8043557300bb1a45
-  data.tar.gz: b42b32a4e9eb80761b856e343d0c2c278c934adb
+  metadata.gz: 5d293949a3ac2383cd254f0efd2268ef24286f1d
+  data.tar.gz: 2439db865559b2f9b494b7755004d3f7a7d9b4ba
 SHA512:
-  metadata.gz: b2654b2c2242caff29794f601bfcccd8b6898ed1879e0ccf35bbdebfc3ed511bf68cbf2d5ad63eb4c6b3c898cbe9674686c1e4233198aa4e8ae85192ac6c63d1
-  data.tar.gz: ffd5e6fa31da9d4fb9c3766fcd850994418a108fb364eac9eab873b394ea0c9ba50c039deeffd5abd94eeeee7c5068c4399faf8503006d4795aa3a5990e33528
+  metadata.gz: 2a7af42dc109723ea0908e8bb3ba9cbf4269f923dbd8a68e6cabd87f53e98dec59884cb7d4daf0df8de98b3d209302a9ba214e45d04fadffd372fabf91c9db26
+  data.tar.gz: 41bac579c874b9ef29aaf5a87e243fd9ba268efe46c60e94cdc728e0e4d0f01b37b60b40fb0fa80a82889a43ac44174c6d8385386fb72a98c37dd0c5868c3b97

data/README.md

@@ -40,7 +40,7 @@ installed by default on most systems (including Mac). On Ubuntu you can use:
 
 The simplest way to use Slaw is via the commandline:
 
-    $ slaw convert myfile.pdf
+    $ slaw parse myfile.pdf
 
 ## Overview
 
@@ -79,7 +79,7 @@ extractor = Slaw::Extract::Extractor.new
 text = extractor.extract_from_pdf('/path/to/file.pdf')
 
 # parse the text into a XML and
-generator = Slaw::ZA::BylawGenerator.new
+generator = Slaw::ActGenerator.new
 bylaw = generator.generate_from_text(text)
 puts bylaw.to_xml(indent: 2)
 

data/bin/slaw

@@ -4,13 +4,14 @@ require 'thor'
 require 'slaw'
 
 class SlawCLI < Thor
-  desc "convert FILE", "convert FILE into Akoma Ntoso XML"
-  option :input, enum: ['text', 'pdf'], desc: "Type of input, determined by file extension by default."
-  option :output, enum: ['text', 'xml'], default: 'xml', desc: "Type of output required."
-
   # TODO: support different grammars and locales
 
-  def convert(name)
+  desc "parse FILE", "parse FILE into Akoma Ntoso XML"
+  option :input, enum: ['text', 'pdf'], desc: "Type of input if it can't be determined automatically"
+  option :pdftotext, desc: "Location of the pdftotext binary if not in PATH"
+  option :definitions, type: :boolean, desc: "Find and link definitions (this can be slow). Default: false"
+  def parse(name)
+    Slaw::Extract::Extractor.pdftotext_path = options[:pdftotext] if options[:pdftotext]
     extractor = Slaw::Extract::Extractor.new
 
     case options[:input]
@@ -22,14 +23,13 @@ class SlawCLI < Thor
       text = extractor.extract_from_file(name)
     end
 
-    case options[:output]
-    when 'text'
-      puts text
-    when 'xml'
-      generator = Slaw::ZA::BylawGenerator.new
-      bylaw = generator.generate_from_text(text)
-      puts bylaw.to_xml(indent: 2)
-    end
+    generator = Slaw::ActGenerator.new
+    act = generator.generate_from_text(text)
+
+    # definitions?
+    generator.builder.link_definitions(act.doc) if options[:definitions]
+
+    puts act.to_xml(indent: 2)
   end
 end
 

data/lib/slaw.rb

@@ -17,7 +17,7 @@ require 'slaw/parse/builder'
 require 'slaw/parse/cleanser'
 require 'slaw/parse/error'
 
-require 'slaw/za/bylaw_generator'
+require 'slaw/generator'
 require 'slaw/extract/extractor'
 
 module Slaw

data/lib/slaw/extract/extractor.rb

@@ -1,5 +1,6 @@
 require 'open3'
 require 'tempfile'
+require 'mimemagic'
 
 module Slaw
   module Extract
@@ -31,13 +32,19 @@ module Slaw
       def extract_from_file(filename)
         ext = filename[-4..-1].downcase
 
-        case ext
-        when '.pdf'
+        mimetype = get_mimetype(filename)
+
+        case mimetype && mimetype.type
+        when 'application/pdf'
           extract_from_pdf(filename)
-        when '.txt'
+        when 'text/plain', nil
           extract_from_text(filename)
         else
-          raise ArgumentError.new("Unsupported file type #{ext}")
+          if mimetype.text?
+            extract_from_text(filename)
+          else
+            raise ArgumentError.new("Unsupported file type #{ext} (#{mimetype || unknown})")
+          end
         end
       end
 
@@ -103,6 +110,11 @@ module Slaw
         end
       end
 
+      def get_mimetype(filename)
+        File.open(filename) { |f| MimeMagic.by_magic(f) } \
+          || MimeMagic.by_path(filename)
+      end
+
       # Get location of the pdftotext executable for all instances.
       def self.pdftotext_path
         @@pdftotext_path

data/lib/slaw/generator.rb

@@ -0,0 +1,36 @@
+module Slaw
+  # Base class for generating Act documents
+  class ActGenerator
+    Treetop.load(File.dirname(__FILE__) + "/za/act.treetop")
+
+    # [Treetop::Runtime::CompiledParser] compiled parser
+    attr_accessor :parser
+
+    # [Slaw::Parse::Builder] builder used by the generator
+    attr_accessor :builder
+
+    def initialize
+      @parser = Slaw::ZA::ActParser.new
+      @builder = Slaw::Parse::Builder.new(parser: @parser)
+      @cleanser = Slaw::Parse::Cleanser.new
+      @document_class = Slaw::Act
+    end
+
+    # Generate a Slaw::Act instance from plain text.
+    #
+    # @param text [String] plain text
+    #
+    # @return [Slaw::Act] the resulting act
+    def generate_from_text(text)
+      act = @document_class.new
+      act.doc = @builder.parse_and_process_text(cleanup(text))
+      act
+    end
+
+    def cleanup(text)
+      text = @cleanser.cleanup(text)
+      text = @cleanser.reformat(text)
+      text
+    end
+  end
+end

data/lib/slaw/parse/builder.rb

@@ -139,7 +139,6 @@ module Slaw
       def postprocess(doc)
         normalise_headings(doc)
         find_short_title(doc)
-        link_definitions(doc)
         nest_blocklists(doc)
 
         doc

data/lib/slaw/version.rb

@@ -1,3 +1,3 @@
 module Slaw
-  VERSION = "0.4.1"
+  VERSION = "0.5.0"
 end

data/lib/slaw/za/{bylaw.treetop → act.treetop}

@@ -1,19 +1,20 @@
 # encoding: UTF-8
 
 require 'slaw/parse/grammar_helpers'
+require 'slaw/za/act_nodes'
 
 module Slaw
   module ZA
-    grammar Bylaw
+    grammar Act
       include Slaw::Parse::GrammarHelpers
 
       ########
       # major containers
 
-      rule bylaw
+      rule act
         preamble
         chapters:chapter*
-        schedules:schedules <Bylaw>
+        schedules:schedules <Act>
       end
 
       rule preamble

data/lib/slaw/za/{bylaw_nodes.rb → act_nodes.rb}

@@ -1,58 +1,77 @@
 module Slaw
   module ZA
-    module Bylaw
-      class Bylaw < Treetop::Runtime::SyntaxNode
+    module Act
+      class Act < Treetop::Runtime::SyntaxNode
+        FRBR_URI = '/za/act/1980/01'
+
         def to_xml(b)
           b.act(contains: "originalVersion") { |b|
-            b.meta { |b|
-              b.identification(source: "#openbylaws") { |b|
-                # TODO: correct values
-                b.FRBRWork { |b|
-                  b.FRBRthis(value: '/za/by-law/locale/1980/name/main')
-                  b.FRBRuri(value: '/za/by-law/locale/1980/name')
-                  b.FRBRalias(value: 'By-Law Short Title')
-                  b.FRBRdate(date: '1980-01-01', name: 'Generation')
-                  b.FRBRauthor(href: '#council', as: '#author')
-                  b.FRBRcountry(value: 'za')
-                }
-                b.FRBRExpression { |b|
-                  b.FRBRthis(value: '/za/by-law/locale/1980/name/main/eng@')
-                  b.FRBRuri(value: '/za/by-law/locale/1980/name/eng@')
-                  b.FRBRdate(date: '1980-01-01', name: 'Generation')
-                  b.FRBRauthor(href: '#council', as: '#author')
-                  b.FRBRlanguage(language: 'eng')
-                }
-                b.FRBRManifestation { |b|
-                  b.FRBRthis(value: '/za/by-law/locale/1980/name/main/eng@')
-                  b.FRBRuri(value: '/za/by-law/locale/1980/name/eng@')
-                  b.FRBRdate(date: Time.now.strftime('%Y-%m-%d'), name: 'Generation')
-                  b.FRBRauthor(href: '#openbylaws', as: '#author')
-                }
-              }
+            write_meta(b)
+            write_preamble(b)
+            write_body(b)
+            write_schedules(b)
+          }
+        end
 
-              b.publication(date: '1980-01-01',
-                            name: 'Province of Western Cape: Provincial Gazette',
-                            number: 'XXXX',
-                            showAs: 'Province of Western Cape: Provincial Gazette')
+        def write_meta(b)
+          b.meta { |b|
+            write_identification(b)
 
-              b.references(source: "#this") {
-                b.TLCOrganization(id: 'openbylaws', href: 'http://openbylaws.org.za', showAs: "openbylaws.org.za")
-                b.TLCOrganization(id: 'council', href: '/ontology/organization/za/council.cape-town', showAs: "Cape Town City Council")
-                b.TLCRole(id: 'author', href: '/ontology/role/author', showAs: 'Author')
-              }
+            b.publication(date: '1980-01-01',
+                          name: 'Publication Name',
+                          number: 'XXXX',
+                          showAs: 'Publication Name')
+
+            b.references(source: "#this") {
+              b.TLCOrganization(id: 'slaw', href: 'https://github.com/longhotsummer/slaw', showAs: "Slaw")
+              b.TLCOrganization(id: 'council', href: '/ontology/organization/za/council', showAs: "Council")
+              b.TLCRole(id: 'author', href: '/ontology/role/author', showAs: 'Author')
             }
+          }
+        end
 
-            if preamble.text_value != ""
-              b.preamble { |b|
-                preamble.to_xml(b)
-              }
-            end
+        def write_identification(b)
+          b.identification(source: "#slaw") { |b|
+            # use stub values so that we can generate a validating document
+            b.FRBRWork { |b|
+              b.FRBRthis(value: "#{FRBR_URI}/main")
+              b.FRBRuri(value: '/za/act/locale/1980/name')
+              b.FRBRalias(value: 'Short Title')
+              b.FRBRdate(date: '1980-01-01', name: 'Generation')
+              b.FRBRauthor(href: '#council', as: '#author')
+              b.FRBRcountry(value: 'za')
+            }
+            b.FRBRExpression { |b|
+              b.FRBRthis(value: '/za/act/locale/1980/name/main/eng@')
+              b.FRBRuri(value: '/za/act/locale/1980/name/eng@')
+              b.FRBRdate(date: '1980-01-01', name: 'Generation')
+              b.FRBRauthor(href: '#council', as: '#author')
+              b.FRBRlanguage(language: 'eng')
+            }
+            b.FRBRManifestation { |b|
+              b.FRBRthis(value: '/za/act/locale/1980/name/main/eng@')
+              b.FRBRuri(value: '/za/act/locale/1980/name/eng@')
+              b.FRBRdate(date: Time.now.strftime('%Y-%m-%d'), name: 'Generation')
+              b.FRBRauthor(href: '#slaw', as: '#author')
+            }
+          }
+        end
 
-            b.body { |b|
-              chapters.elements.each { |e| e.to_xml(b) }
+        def write_preamble(b)
+          if preamble.text_value != ""
+            b.preamble { |b|
+              preamble.to_xml(b)
             }
+          end
+        end
+
+        def write_body(b)
+          b.body { |b|
+            chapters.elements.each { |e| e.to_xml(b) }
           }
+        end
 
+        def write_schedules(b)
           schedules.to_xml(b)
         end
       end
@@ -179,7 +198,7 @@ module Slaw
         # a section title of the form:
         #
         # Definitions
-        # 1. In this by-law...
+        # 1. In this act...
 
         def num
           section_title_prefix.number_letter.text_value
@@ -194,7 +213,7 @@ module Slaw
         # a section title of the form:
         #
         # 1. Definitions
-        # In this by-law...
+        # In this act...
         #
         # In this format, the title is optional and the section content may
         # start where we think the title is.
@@ -299,26 +318,26 @@ module Slaw
             b.component(id: 'component-0') { |b|
               b.doc(name: 'schedules') { |b|
                 b.meta { |b| 
-                  b.identification(source: "#openbylaws") { |b|
+                  b.identification(source: "#slaw") { |b|
                     b.FRBRWork { |b|
-                      b.FRBRthis(value: '/za/by-law/locale/1980/name/main/schedules')
-                      b.FRBRuri(value: '/za/by-law/locale/1980/name/schedules')
+                      b.FRBRthis(value: '/za/act/locale/1980/name/main/schedules')
+                      b.FRBRuri(value: '/za/act/locale/1980/name/schedules')
                       b.FRBRdate(date: '1980-01-01', name: 'Generation')
                       b.FRBRauthor(href: '#council', as: '#author')
                       b.FRBRcountry(value: 'za')
                     }
                     b.FRBRExpression { |b|
-                      b.FRBRthis(value: '/za/by-law/locale/1980/name/main//schedules/eng@')
-                      b.FRBRuri(value: '/za/by-law/locale/1980/name/schedules/eng@')
+                      b.FRBRthis(value: '/za/act/locale/1980/name/main//schedules/eng@')
+                      b.FRBRuri(value: '/za/act/locale/1980/name/schedules/eng@')
                       b.FRBRdate(date: '1980-01-01', name: 'Generation')
                       b.FRBRauthor(href: '#council', as: '#author')
                       b.FRBRlanguage(language: 'eng')
                     }
                     b.FRBRManifestation { |b|
-                      b.FRBRthis(value: '/za/by-law/locale/1980/name/main/schedules/eng@')
-                      b.FRBRuri(value: '/za/by-law/locale/1980/name/schedules/eng@')
+                      b.FRBRthis(value: '/za/act/locale/1980/name/main/schedules/eng@')
+                      b.FRBRuri(value: '/za/act/locale/1980/name/schedules/eng@')
                       b.FRBRdate(date: Time.now.strftime('%Y-%m-%d'), name: 'Generation')
-                      b.FRBRauthor(href: '#openbylaws', as: '#author')
+                      b.FRBRauthor(href: '#slaw', as: '#author')
                     }
                   }
                 }

data/slaw.gemspec

@@ -27,4 +27,5 @@ Gem::Specification.new do |spec|
   spec.add_runtime_dependency "builder", "~> 3.2.2"
   spec.add_runtime_dependency "log4r", "~> 1.1.10"
   spec.add_runtime_dependency "thor", "~> 0.19.1"
+  spec.add_runtime_dependency "mimemagic", "~> 0.2.1"
 end

data/spec/act_spec.rb

@@ -0,0 +1,56 @@
+# encoding: UTF-8
+
+require 'spec_helper'
+require 'slaw'
+
+describe Slaw::Act do
+  let(:filename) { File.dirname(__FILE__) + "/fixtures/community-fire-safety.xml" }
+  subject { Slaw::Act.new(filename) }
+
+  it 'should have correct basic properties' do
+    subject.title.should == 'Community Fire Safety By-law'
+    subject.amended?.should be_true
+  end
+
+  it 'should set the title correctly' do
+    subject.title = 'foo'
+    subject.meta.at_xpath('./a:identification/a:FRBRWork/a:FRBRalias', a: Slaw::NS)['value'].should == 'foo'
+  end
+
+  it 'should set the title if it doesnt exist' do
+    subject.meta.at_xpath('./a:identification/a:FRBRWork/a:FRBRalias', a: Slaw::NS).remove
+    subject.title = 'bar'
+    subject.title.should == 'bar'
+  end
+
+  it 'should set the publication details' do
+    subject.meta.at_xpath('./a:publication', a: Slaw::NS).remove
+
+    subject.published!(name: 'foo', number: '1234', date: '2014-01-01')
+    subject.publication['name'].should == 'foo'
+    subject.publication['showAs'].should == 'foo'
+    subject.publication['number'].should == '1234'
+  end
+
+  it 'should get/set the work date' do
+    subject.date.should == '2002-02-28'
+
+    subject.date = '2014-01-01'
+    subject.date.should == '2014-01-01'
+    subject.meta.at_xpath('./a:identification/a:FRBRWork/a:FRBRdate[@name="Generation"]', a: Slaw::NS)['date'].should == '2014-01-01'
+    subject.meta.at_xpath('./a:identification/a:FRBRExpression/a:FRBRdate[@name="Generation"]', a: Slaw::NS)['date'].should == '2014-01-01'
+
+    subject.id_uri.should == '/za/by-law/2014/2002'
+  end
+
+  it 'should update the uri when the year changes' do
+    subject.id_uri.should == '/za/by-law/cape-town/2002/community-fire-safety'
+    subject.year = '1980'
+    subject.id_uri.should == '/za/by-law/1980/2002'
+  end
+
+  it 'should validate' do
+    subject.validate.should == []
+    subject.validates?.should be_true
+  end
+end

data/spec/bylaw_spec.rb

@@ -12,9 +12,16 @@ describe Slaw::ByLaw do
     subject.amended?.should be_true
   end
 
-  it 'should set the title correctly' do
-    subject.title = 'foo'
-    subject.meta.at_xpath('./a:identification/a:FRBRWork/a:FRBRalias', a: Slaw::NS)['value'].should == 'foo'
+  it 'should update the uri when the region changes' do
+    subject.id_uri.should == '/za/by-law/cape-town/2002/community-fire-safety'
+    subject.region = 'foo-bar'
+    subject.id_uri.should == '/za/by-law/foo-bar/2002/community-fire-safety'
+  end
+
+  it 'should update the uri when the name changes' do
+    subject.id_uri.should == '/za/by-law/cape-town/2002/community-fire-safety'
+    subject.name = 'foo-bar'
+    subject.id_uri.should == '/za/by-law/cape-town/2002/foo-bar'
   end
 
   it 'should set the title if it doesnt exist' do
@@ -23,15 +30,6 @@ describe Slaw::ByLaw do
     subject.title.should == 'bar as amended'
   end
 
-  it 'should set the publication details' do
-    subject.meta.at_xpath('./a:publication', a: Slaw::NS).remove
-
-    subject.published!(name: 'foo', number: '1234', date: '2014-01-01')
-    subject.publication['name'].should == 'foo'
-    subject.publication['showAs'].should == 'foo'
-    subject.publication['number'].should == '1234'
-  end
-
   it 'should get/set the work date' do
     subject.date.should == '2002-02-28'
 
@@ -48,21 +46,4 @@ describe Slaw::ByLaw do
     subject.year = '1980'
     subject.id_uri.should == '/za/by-law/cape-town/1980/community-fire-safety'
   end
-
-  it 'should update the uri when the region changes' do
-    subject.id_uri.should == '/za/by-law/cape-town/2002/community-fire-safety'
-    subject.region = 'foo-bar'
-    subject.id_uri.should == '/za/by-law/foo-bar/2002/community-fire-safety'
-  end
-
-  it 'should update the uri when the name changes' do
-    subject.id_uri.should == '/za/by-law/cape-town/2002/community-fire-safety'
-    subject.name = 'foo-bar'
-    subject.id_uri.should == '/za/by-law/cape-town/2002/foo-bar'
-  end
-
-  it 'should validate' do
-    subject.validate.should == []
-    subject.validates?.should be_true
-  end
 end

data/spec/za/{bylaw_spec.rb → act_spec.rb}

@@ -3,9 +3,8 @@
 require 'builder'
 
 require 'slaw'
-require 'slaw/za/bylaw_generator'
 
-describe Slaw::ZA::BylawGenerator do
+describe Slaw::ActGenerator do
   def parse(rule, s)
     subject.builder.text_to_syntax_tree(s, {root: rule})
   end
@@ -64,7 +63,7 @@ EOS
 
     it 'should handle parts and odd section numbers' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 PART 1
 PREVENTION AND SUPPRESSION OF HEALTH NUISANCES
 1.
@@ -189,7 +188,7 @@ EOS
 
   context 'preamble' do
     it 'should consider any text at the start to be preamble' do
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 foo
 bar
 (1) stuff
@@ -208,7 +207,7 @@ baz
     end
 
     it 'should support an optional preamble' do
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 PREAMBLE
 foo
 1. Section
@@ -219,7 +218,7 @@ EOS
     end
 
     it 'should support no preamble' do
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 1. Section
 bar
 EOS
@@ -235,7 +234,7 @@ EOS
   context 'sections' do
     it 'should handle section numbers after title' do
       subject.parser.options = {section_number_after_title: true}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 Section
 1. (1) hello
 EOS
@@ -247,7 +246,7 @@ EOS
 
     it 'should handle section numbers before title' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 1. Section
 (1) hello
 EOS
@@ -259,7 +258,7 @@ EOS
 
     it 'should handle section numbers without a dot' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 1 A section
 (1) hello
 2 Another section
@@ -277,7 +276,7 @@ EOS
 
     it 'should handle sections without titles' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 1. No owner or occupier of any shop or business premises or vacant land, blah blah
 2. Notwithstanding the provision of any other By-law or legislation no person shall—
 EOS
@@ -294,7 +293,7 @@ EOS
 
     it 'should handle sections without titles and with subsections' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 10. (1) Transporters must remove medical waste.
 (2) Without limiting generality, stuff.
 EOS
@@ -308,7 +307,7 @@ EOS
 
     it 'should realise complex section titles are actually section content' do
       subject.parser.options = {section_number_after_title: false}
-      node = parse :bylaw, <<EOS
+      node = parse :act, <<EOS
 10. The owner of any premises which is let or sublet to more than one tenant, shall maintain at all times in a clean and sanitary condition every part of such premises as may be used in common by more than one tenant.
 11. No person shall keep, cause or suffer to be kept any factory or trade premises so as to cause or give rise to smells or effluvia that constitute a health nuisance.
 EOS

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: slaw
 version: !ruby/object:Gem::Version
-  version: 0.4.1
+  version: 0.5.0
 platform: ruby
 authors:
 - Greg Kempe
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-03-04 00:00:00.000000000 Z
+date: 2015-03-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -122,6 +122,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: 0.19.1
+- !ruby/object:Gem::Dependency
+  name: mimemagic
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.2.1
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 0.2.1
 description: Slaw is a lightweight library for rendering and generating Akoma Ntoso
   acts from plain text and PDF documents.
 email:
@@ -143,6 +157,7 @@ files:
 - lib/slaw/bylaw.rb
 - lib/slaw/collection.rb
 - lib/slaw/extract/extractor.rb
+- lib/slaw/generator.rb
 - lib/slaw/lifecycle_event.rb
 - lib/slaw/logging.rb
 - lib/slaw/namespace.rb
@@ -159,10 +174,10 @@ files:
 - lib/slaw/schemas/xml.xsd
 - lib/slaw/version.rb
 - lib/slaw/xml_support.rb
-- lib/slaw/za/bylaw.treetop
-- lib/slaw/za/bylaw_generator.rb
-- lib/slaw/za/bylaw_nodes.rb
+- lib/slaw/za/act.treetop
+- lib/slaw/za/act_nodes.rb
 - slaw.gemspec
+- spec/act_spec.rb
 - spec/bylaw_spec.rb
 - spec/extract/extractor_spec.rb
 - spec/fixtures/community-fire-safety.xml
@@ -170,7 +185,7 @@ files:
 - spec/parse/cleanser_spec.rb
 - spec/spec_helper.rb
 - spec/xml_helpers.rb
-- spec/za/bylaw_spec.rb
+- spec/za/act_spec.rb
 homepage: ''
 licenses:
 - MIT
@@ -196,6 +211,7 @@ signing_key:
 specification_version: 4
 summary: A lightweight library for using Akoma Ntoso acts in Ruby.
 test_files:
+- spec/act_spec.rb
 - spec/bylaw_spec.rb
 - spec/extract/extractor_spec.rb
 - spec/fixtures/community-fire-safety.xml
@@ -203,4 +219,4 @@ test_files:
 - spec/parse/cleanser_spec.rb
 - spec/spec_helper.rb
 - spec/xml_helpers.rb
-- spec/za/bylaw_spec.rb
+- spec/za/act_spec.rb

data/lib/slaw/za/bylaw_generator.rb

@@ -1,41 +0,0 @@
-require 'slaw/za/bylaw_nodes'
-
-module Slaw
-  # Support specifically for South Africa
-  module ZA
-
-    # Support class for generating South African bylaws
-    class BylawGenerator
-      Treetop.load(File.dirname(__FILE__) + "/bylaw.treetop")
-
-      # [Treetop::Runtime::CompiledParser] compiled bylaw parser
-      attr_accessor :parser
-
-      # [Slaw::Parse::Builder] builder used by the generator
-      attr_accessor :builder
-
-      def initialize
-        @parser = Slaw::ZA::BylawParser.new
-        @builder = Slaw::Parse::Builder.new(parser: @parser)
-        @cleanser = Slaw::Parse::Cleanser.new
-      end
-
-      # Generate a Slaw::Bylaw instance from plain text.
-      #
-      # @param text [String] plain text
-      #
-      # @return [Slaw::ByLaw] the resulting bylaw
-      def generate_from_text(text)
-        bylaw = Slaw::ByLaw.new
-        bylaw.doc = @builder.parse_and_process_text(cleanup(text))
-        bylaw
-      end
-
-      def cleanup(text)
-        text = @cleanser.cleanup(text)
-        text = @cleanser.reformat(text)
-        text
-      end
-    end
-  end
-end