edifact_parser 0.9.0

data/lib/edifact_parser.rb

@@ -0,0 +1,20 @@
+$:.unshift(File.dirname(__FILE__) + '/../lib')
+
+require 'edifact_parser/parser'
+require 'edifact_parser/tokenizer'
+require 'stringio'
+
+module EdifactParser
+
+  def self.load_io(input)
+    tok     = EdifactParser::Tokenizer.new input
+    parser  = EdifactParser::Parser.new tok
+    handler = parser.parse
+    handler.result
+  end
+
+  def self.load(edi)
+    load_io(StringIO.new(edi))
+  end
+
+end

data/lib/edifact_parser/handler.rb

@@ -0,0 +1,71 @@
+module EdifactParser
+  class Handler
+    attr_reader :stack
+
+    def initialize
+      @stack = [[:array]]
+      @element_started = false
+      @empty = true
+    end
+
+    def start_segment
+      push [:array]
+    end
+
+    def start_element
+      push [:array]
+      @element_started = true
+    end
+
+    def scalar(s)
+      @stack.last << [:scalar, s]
+      @empty = false
+    end
+
+    def colon
+      if @empty
+        @stack.last << [:nil]
+      end
+      @empty = true
+    end
+
+    def end_segment
+      @stack.pop
+    end
+
+    def end_element
+      if @element_started
+        @stack.pop
+        @element_started = false
+      end
+    end
+
+    def qualifier(q)
+      @stack.last << [:scalar, q]
+    end
+
+    def result
+      root = @stack.first
+      process(root.first, root.drop(1))
+    end
+
+    private
+
+      def process(type, rest)
+        case type
+        when :array
+          rest.map { |x| process(x.first, x.drop(1)) }
+        when :scalar
+          rest.first
+        when :nil
+          nil
+        end
+      end
+
+      def push(o)
+        @stack.last << o
+        @stack << o
+      end
+
+  end
+end

data/lib/edifact_parser/parser.rb

@@ -0,0 +1,226 @@
+#
+# DO NOT MODIFY!!!!
+# This file is automatically generated by Racc 1.4.9
+# from Racc grammer file "".
+#
+
+require 'racc/parser.rb'
+module EdifactParser
+  class Parser < Racc::Parser
+
+
+  require_relative 'handler'
+
+  attr_reader :handler
+
+  def initialize(tokenizer, handler = Handler.new)
+    @tokenizer = tokenizer
+    @handler   = handler
+    super()
+  end
+
+  def next_token
+    @tokenizer.next_token
+  end
+
+  def parse
+    do_parse
+    handler
+  end
+##### State transition tables begin ###
+
+racc_action_table = [
+     7,    27,    26,     4,    22,    17,    18,    17,    18,    27,
+    26,     7,     7,     7,     8,    19 ]
+
+racc_action_check = [
+     0,    15,    15,     0,    11,    11,    11,     6,     6,    16,
+    16,     3,    10,     2,     1,     8 ]
+
+racc_action_pointer = [
+    -2,    14,    11,     9,   nil,   nil,     0,   nil,    15,   nil,
+    10,    -2,   nil,   nil,   nil,    -2,     6,   nil,   nil,   nil,
+   nil,   nil,   nil,   nil,   nil,   nil,   nil,   nil,   nil ]
+
+racc_action_default = [
+   -23,   -23,    -1,   -23,    -3,    -5,   -23,   -13,   -23,    -4,
+    -2,   -23,    -8,    -9,   -10,   -11,   -12,   -20,   -21,    29,
+    -6,    -7,   -22,   -14,   -16,   -17,   -18,   -19,   -15 ]
+
+racc_goto_table = [
+     9,    12,     2,    23,    28,    10,    21,    11,     9,    20,
+     3,     1 ]
+
+racc_goto_check = [
+     4,     8,     2,    13,    13,     2,     8,     6,     4,     7,
+     3,     1 ]
+
+racc_goto_pointer = [
+   nil,    11,     2,    10,    -2,   nil,     1,    -2,    -5,   nil,
+   nil,   nil,   nil,   -12,   nil,   nil ]
+
+racc_goto_default = [
+   nil,   nil,   nil,   nil,     5,     6,   nil,   nil,   nil,    13,
+    14,    15,    16,   nil,    24,    25 ]
+
+racc_reduce_table = [
+  0, 0, :racc_error,
+  1, 10, :_reduce_none,
+  2, 10, :_reduce_none,
+  1, 12, :_reduce_none,
+  2, 11, :_reduce_none,
+  1, 11, :_reduce_none,
+  3, 13, :_reduce_none,
+  2, 15, :_reduce_none,
+  1, 15, :_reduce_none,
+  1, 17, :_reduce_none,
+  1, 17, :_reduce_none,
+  1, 17, :_reduce_none,
+  1, 17, :_reduce_none,
+  1, 14, :_reduce_13,
+  2, 18, :_reduce_none,
+  2, 19, :_reduce_none,
+  1, 22, :_reduce_none,
+  1, 22, :_reduce_none,
+  1, 24, :_reduce_18,
+  1, 23, :_reduce_19,
+  1, 20, :_reduce_20,
+  1, 21, :_reduce_21,
+  1, 16, :_reduce_22 ]
+
+racc_reduce_n = 23
+
+racc_shift_n = 29
+
+racc_token_table = {
+  false => 0,
+  :error => 1,
+  :QUALIFIER => 2,
+  :STRING => 3,
+  :NUMBER => 4,
+  :OPTIONAL_BEGIN => 5,
+  :SEGMENT_END => 6,
+  :PLUS => 7,
+  :COLON => 8 }
+
+racc_nt_base = 9
+
+racc_use_result_var = true
+
+Racc_arg = [
+  racc_action_table,
+  racc_action_check,
+  racc_action_default,
+  racc_action_pointer,
+  racc_goto_table,
+  racc_goto_check,
+  racc_goto_default,
+  racc_goto_pointer,
+  racc_nt_base,
+  racc_reduce_table,
+  racc_token_table,
+  racc_shift_n,
+  racc_reduce_n,
+  racc_use_result_var ]
+
+Racc_token_to_s_table = [
+  "$end",
+  "error",
+  "QUALIFIER",
+  "STRING",
+  "NUMBER",
+  "OPTIONAL_BEGIN",
+  "SEGMENT_END",
+  "PLUS",
+  "COLON",
+  "$start",
+  "document",
+  "segments",
+  "beginning",
+  "segment",
+  "qual",
+  "values",
+  "segment_end",
+  "value",
+  "p_scalar",
+  "c_scalar",
+  "plus",
+  "col",
+  "scalar",
+  "string",
+  "number" ]
+
+Racc_debug_parser = true
+
+##### State transition tables end #####
+
+# reduce 0 omitted
+
+# reduce 1 omitted
+
+# reduce 2 omitted
+
+# reduce 3 omitted
+
+# reduce 4 omitted
+
+# reduce 5 omitted
+
+# reduce 6 omitted
+
+# reduce 7 omitted
+
+# reduce 8 omitted
+
+# reduce 9 omitted
+
+# reduce 10 omitted
+
+# reduce 11 omitted
+
+# reduce 12 omitted
+
+def _reduce_13(val, _values, result)
+ @handler.start_segment; @handler.qualifier val[0] 
+    result
+end
+
+# reduce 14 omitted
+
+# reduce 15 omitted
+
+# reduce 16 omitted
+
+# reduce 17 omitted
+
+def _reduce_18(val, _values, result)
+ @handler.scalar val[0] 
+    result
+end
+
+def _reduce_19(val, _values, result)
+ @handler.scalar val[0].gsub("?", "") 
+    result
+end
+
+def _reduce_20(val, _values, result)
+ @handler.end_element; @handler.start_element 
+    result
+end
+
+def _reduce_21(val, _values, result)
+ @handler.colon 
+    result
+end
+
+def _reduce_22(val, _values, result)
+ @handler.end_element; @handler.end_segment 
+    result
+end
+
+def _reduce_none(val, _values, result)
+  val[0]
+end
+
+  end   # class Parser
+  end   # module EdifactParser

data/lib/edifact_parser/parser.tab.rb

@@ -0,0 +1,180 @@
+#
+# DO NOT MODIFY!!!!
+# This file is automatically generated by Racc 1.4.9
+# from Racc grammer file "".
+#
+
+require 'racc/parser.rb'
+module EdifactParser
+  class Parser < Racc::Parser
+
+module_eval(<<'...end parser.y/module_eval...', 'parser.y', 35)
+
+  require_relative 'handler'
+
+  attr_reader :handler
+
+  def initialize(tokenizer, handler = Handler.new)
+    @tokenizer = tokenizer
+    @handler   = handler
+    super()
+  end
+
+  def next_token
+    @tokenizer.next_token
+  end
+
+  def parse
+    do_parse
+    handler
+  end
+...end parser.y/module_eval...
+##### State transition tables begin ###
+
+racc_action_table = [
+     6,     7,    14,    10,    11,    10,    11,     6,     4,     3 ]
+
+racc_action_check = [
+     8,     4,     8,    13,    13,     5,     5,     2,     1,     0 ]
+
+racc_action_pointer = [
+     7,     8,     1,   nil,     1,     2,   nil,   nil,    -6,   nil,
+   nil,   nil,   nil,     0,   nil,   nil ]
+
+racc_action_default = [
+   -10,   -10,   -10,    -2,   -10,   -10,    -7,    16,   -10,    -4,
+    -5,    -6,    -1,   -10,    -9,    -3 ]
+
+racc_goto_table = [
+     9,     5,     8,     2,    12,     1,   nil,    13,    15 ]
+
+racc_goto_check = [
+     6,     3,     4,     2,     5,     1,   nil,     3,     6 ]
+
+racc_goto_pointer = [
+   nil,     5,     3,    -1,    -3,    -4,    -5,   nil ]
+
+racc_goto_default = [
+   nil,   nil,   nil,   nil,   nil,   nil,   nil,   nil ]
+
+racc_reduce_table = [
+  0, 0, :racc_error,
+  4, 10, :_reduce_none,
+  1, 11, :_reduce_2,
+  3, 13, :_reduce_none,
+  1, 13, :_reduce_none,
+  1, 15, :_reduce_none,
+  1, 15, :_reduce_6,
+  1, 12, :_reduce_7,
+  1, 16, :_reduce_8,
+  1, 14, :_reduce_9 ]
+
+racc_reduce_n = 10
+
+racc_shift_n = 16
+
+racc_token_table = {
+  false => 0,
+  :error => 1,
+  :QUALIFIER => 2,
+  :STRING => 3,
+  :NUMBER => 4,
+  :SPACE => 5,
+  "+" => 6,
+  ":" => 7,
+  "'" => 8 }
+
+racc_nt_base = 9
+
+racc_use_result_var = true
+
+Racc_arg = [
+  racc_action_table,
+  racc_action_check,
+  racc_action_default,
+  racc_action_pointer,
+  racc_goto_table,
+  racc_goto_check,
+  racc_goto_default,
+  racc_goto_pointer,
+  racc_nt_base,
+  racc_reduce_table,
+  racc_token_table,
+  racc_shift_n,
+  racc_reduce_n,
+  racc_use_result_var ]
+
+Racc_token_to_s_table = [
+  "$end",
+  "error",
+  "QUALIFIER",
+  "STRING",
+  "NUMBER",
+  "SPACE",
+  "\"+\"",
+  "\":\"",
+  "\"'\"",
+  "$start",
+  "segment",
+  "qual",
+  "plus",
+  "values",
+  "segment_end",
+  "value",
+  "colon" ]
+
+Racc_debug_parser = false
+
+##### State transition tables end #####
+
+# reduce 0 omitted
+
+# reduce 1 omitted
+
+module_eval(<<'.,.,', 'parser.y', 8)
+  def _reduce_2(val, _values, result)
+     raise "test" 
+    result
+  end
+.,.,
+
+# reduce 3 omitted
+
+# reduce 4 omitted
+
+# reduce 5 omitted
+
+module_eval(<<'.,.,', 'parser.y', 17)
+  def _reduce_6(val, _values, result)
+     @handler.scalar val[0] 
+    result
+  end
+.,.,
+
+module_eval(<<'.,.,', 'parser.y', 21)
+  def _reduce_7(val, _values, result)
+     @handler.end_element; @handler.start_element; @handler.end_component; @handler.start_component 
+    result
+  end
+.,.,
+
+module_eval(<<'.,.,', 'parser.y', 25)
+  def _reduce_8(val, _values, result)
+     @handler.end_component; @handler.start_component 
+    result
+  end
+.,.,
+
+module_eval(<<'.,.,', 'parser.y', 29)
+  def _reduce_9(val, _values, result)
+     @handler.end_segment 
+    result
+  end
+.,.,
+
+def _reduce_none(val, _values, result)
+  val[0]
+end
+
+  end   # class Parser
+  end   # module EdifactParser

data/lib/edifact_parser/tokenizer.rb

@@ -0,0 +1,42 @@
+require 'strscan'
+
+module EdifactParser
+  class Tokenizer
+    OPTIONAL_BEGIN = /^UNA:\+\.\?\s'/
+    QUALIFIER = /(^|(?<='))(UNB|UNH|BGM|DTM|PAI|ALI|IMD|FTX|LOC|GIS|DGS|GIR|RFF|MEA|QTY|MOA|RTE|NAD|DOC|CTA|COM|TAX|PCD|CUX|TDT|TSR|PII|TOD|PAC|EQD|SEL|ALC|RNG|INP|LIN|PRI|UNS|UNT|UNZ|CNT)/
+    STRING = /[A-Za-z0-9\s\.]*(\?')*[A-Za-z0-9\s\.]*(\?\+)*[A-Za-z0-9\s\.]*(\?:)*[A-Za-z0-9\s\.]*/
+    SPACE = /\s+/
+    NUMBER = /[0-9]+(?=[\+:'])/
+    LINE_BREAK = /\n/
+    SEGMENT_END = /(?<!\?)'/
+    PLUS = /(?<!\?)\+/
+    COLON = /(?<!\?)\:/
+
+    def initialize(io)
+      @ss = StringScanner.new io.read
+    end
+
+    def next_token
+      return if @ss.eos?
+
+      # ignore spaces and line breaks
+      @ss.scan(SPACE)
+      @ss.scan(LINE_BREAK)
+
+      return if @ss.eos?
+
+      case
+      when text = @ss.scan(OPTIONAL_BEGIN) then [:OPTIONAL_BEGIN, text]
+      when text = @ss.scan(QUALIFIER) then [:QUALIFIER, text]
+      when text = @ss.scan(SEGMENT_END) then [:SEGMENT_END, text]
+      when text = @ss.scan(PLUS) then [:PLUS, text]
+      when text = @ss.scan(COLON) then [:COLON, text]
+      when text = @ss.scan(NUMBER) then [:NUMBER, text.to_i]
+      when text = @ss.scan(STRING) then [:STRING, text]
+      else
+        x = @ss.getch
+        [x, x]
+      end
+    end
+  end
+end

metadata

@@ -0,0 +1,67 @@
+--- !ruby/object:Gem::Specification
+name: edifact_parser
+version: !ruby/object:Gem::Version
+  version: 0.9.0
+  prerelease: 
+platform: ruby
+authors:
+- Paul Van de Vreede
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2013-01-26 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: racc
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '1.4'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '1.4'
+description: EdifactParser is a simple parser that parses the EDIFACT structure into
+  a Ruy structure so that you can validate and convert the EDIFACT into any structure
+  you like.
+email: paul@vdvreede.net
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/edifact_parser/handler.rb
+- lib/edifact_parser/parser.rb
+- lib/edifact_parser/parser.tab.rb
+- lib/edifact_parser/tokenizer.rb
+- lib/edifact_parser.rb
+homepage: https://github.com/pvdvreede/edifact_parser
+licenses: []
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  none: false
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 1.8.23
+signing_key: 
+specification_version: 3
+summary: Parser for converting EDIFACT documents into a Ruby array of hashes.
+test_files: []