regex_generator 0.3.4 → 0.3.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: aba286d39cd71e85e8393e0aff3f9f4d1872650b4dad12063a7ce84d6d8b2966
-  data.tar.gz: e291cbac8deb7bcaca1889ed16740a424b711e9e150b2e9136401da2bd3ff608
+  metadata.gz: 1ef2c0d2c4fb493acc32581875db3cfae61249d5e447d5f2d0a146bb4ef5de2f
+  data.tar.gz: 312b3c3762c88421675312ea32a663532892bb2e52352ace680242ac90c9bec7
 SHA512:
-  metadata.gz: 11f42b7a3e9043d23e99069f2d6cecf3611b1748f77acfa88c037ab621e4c319f73a958b084e87b3174eaa2732897bdf901e4c5cca096069bd09cf71eff219ee
-  data.tar.gz: 379e3baa608f1ae95c600114944b982c98bc3c5928ce6f408fb37a10ef66c6827aca06823a2804c0ec6a2476c221279e846b22d1c8e8e8f66fcf47c9c2122367
+  metadata.gz: ecf144cf3b49ee96f9eeb1c12808e204dec22c7ef996382097c4e0f3e361f7d3945e17fdabe049903ce931c59238c975394d850916dbdbad26e227dc86ec6fb6
+  data.tar.gz: 1d640a4be4e4c8ca58d1061d4a05d7188492893185e8dcfc8c5d36ce93a5c2e8a3286e864bbfe0fd45ac1e8c6813dbaea23ea2a3e47529dd38d21af1bf845822

data/README.md

@@ -40,6 +40,9 @@ Allowed options:
  - `:self_recognition` - Symbols that will be represented as itself. Can be string or array
  - `:look` - `:ahead` or `:behind` (`:behind` by default). To generate regex with text after or before the value
  - `:strict_count` - When it `true` regex will generated with strict chars count
+ - `:title` - Regex will generated for provided title. If `:title` is provided
+ as Hash (i.e. to generate regex with name capturing groups), `:title` must
+ contains the same keys as target
 
 ## Contributing
 

data/bin/generate_regex

@@ -19,10 +19,8 @@ OptionParser.new do |parser|
     exit
   end
 
-  parser.on('--self_recognition=CHARS', 'Recognizes chargs as itself') do |chars|
-    chars
-  end
-
+  parser.on('--self_recognition=CHARS', 'Recognizes chargs as itself', &:chars)
+  parser.on('--title=TITLE', 'Generates regex for provided title', &:title)
   parser.on('--strict_count', 'Generates regex with a strcit chars count')
   parser.on('--exact_target', 'Generates regex with exact target value')
 end.parse!(into: params)

data/lib/regex_generator.rb

@@ -17,6 +17,7 @@ module RegexGenerator
   #   or after the target
   # @option options [true, false] :strict_count to generate regex with a strict
   #   chars count
+  # @option options [String, Hash] :title to generate regex for provided title
   # @return [Regexp]
   #
   # @example Generate regex

data/lib/regex_generator/characters_recognizer.rb

@@ -1,31 +1,56 @@
 module RegexGenerator
   class CharactersRecognizer
-    PATTERNS = [/[A-Z]/, /[a-z]/, /\d/, /\n/, /\s/, /./].freeze
-
-    # Creates array with regex representation for each char from the text
-    #
-    # @param text [String]
-    # @param options [Hash] options to recognize regex patterns with
-    # @option options [Array] :self_recognition to recognize chars as itself
-    # @return [Array]
-    def self.recognize(text, options = {})
-      return [] unless text
-
-      result = []
-      text.chars.each do |char|
-        PATTERNS.each do |pattern|
-          next unless char[pattern]
-
-          escaped_char = Regexp.escape(char)
-          res_pattern = escaped_char
-          res_pattern = pattern.source if (char.eql?(escaped_char) &&
-            !options[:self_recognition]&.include?(escaped_char)) || char[/\s/]
-
-          break result << res_pattern
+    class << self
+      PATTERNS = [/[A-Z]/, /[a-z]/, /\d/, /\n/, /\s/, /./].freeze
+
+      # Creates array with regex representation for each char from the text
+      #
+      # @param text [String]
+      # @param options [Hash] options to recognize regex patterns with
+      # @option options [Array] :self_recognition to recognize chars as itself
+      # @option options [String, Hash] :title to recognize all chars excluding
+      #   title
+      # @return [Array]
+      def recognize(text, options = {})
+        return [] unless text
+
+        result = []
+
+        for_each_char(text, options) do |char, title|
+          next result << title if title
+
+          PATTERNS.each do |pattern|
+            next unless char[pattern]
+
+            escaped_char = Regexp.escape(char)
+            res_pattern = escaped_char
+            res_pattern = pattern.source if (char.eql?(escaped_char) &&
+              !options[:self_recognition]&.include?(escaped_char)) || char[/\s/]
+
+            break result << res_pattern
+          end
         end
+
+        result
       end
 
-      result
+      private
+
+      # Executes the block for each char in the text excluding the titles
+      def for_each_char(text, options)
+        title_regex = if options[:title].kind_of?(Hash)
+          /(#{options[:title].escape.join('|')})/
+        else
+          /(#{options[:title]})/
+        end
+
+        text.split(title_regex).each do |txt|
+          next if txt.empty?
+          next yield(nil, txt) if txt[title_regex].eql?(txt)
+
+          txt.chars.each { |char| yield char, nil }
+        end
+      end
     end
   end
 end

data/lib/regex_generator/exceptions.rb

@@ -5,6 +5,12 @@ module RegexGenerator
     end
   end
 
+  class TitleNotFoundError < StandardError
+    def message
+      'The title was not found in the provided text'
+    end
+  end
+
   class InvalidOption < StandardError
     def initialize(*options)
       @options = options

data/lib/regex_generator/generator.rb

@@ -12,18 +12,28 @@ module RegexGenerator
     #   or after the target
     # @option options [true, false] :strict_count to generate regex with a
     #   strict chars count
+    # @option options [String, Hash] :title to generate regex for provided title
     def initialize(target, text, options = {})
       @text = text
       @target = RegexGenerator::Target.new(target)
+      @title = RegexGenerator::Target.new(options[:title])
+      if options[:title] && !@title.keys_equal?(@target)
+        raise RegexGenerator::InvalidOption, :title
+      end
+
+      @title_str = @title.to_s
       @target_str = @target.to_s
       @options = options
     end
 
     # @return [Regexp]
     # @raise [TargetNotFoundError] if target text was not found in the text
-    # @raise [InvalidOption] if :look option is not :ahead or :behind
+    # @raise [InvalidOption] if :look option is not :ahead or :behind or :title
+    #   has different keys than target keys
+    # @raise [TitleNotFoundError] if :title was not found in the text
     def generate
       raise RegexGenerator::TargetNotFoundError unless @target.present?(@text)
+      raise RegexGenerator::TitleNotFoundError unless @title.present?(@text)
 
       string_regex_chars = recognize_text(cut_nearest_text, options)
       string_patterns_array = slice_to_identicals(string_regex_chars)
@@ -51,8 +61,8 @@ module RegexGenerator
 
     def text_regex_for_string
       {
-        behind: /[\w\W]*((?:\n|\A)[\w\W]+?)#{@target.escape}/,
-        ahead: /#{@target.escape}([\w\W]+?(?:\n|\Z))/
+        behind: /[\w\W]*((?:\n|\A)[\w\W]*?#{@title_str}\s*)#{@target.escape}/,
+        ahead: /#{@target.escape}([\w\W]*?#{@title_str}(?:\n|\Z))/
       }[options[:look]]
     end
 
@@ -82,6 +92,8 @@ module RegexGenerator
 
     # Prepares options
     def options
+      @options[:title] = @title
+
       if @options[:self_recognition].kind_of? String
         @options[:self_recognition] = @options[:self_recognition].chars
       end

data/lib/regex_generator/target.rb

@@ -45,5 +45,15 @@ module RegexGenerator
         result[key] = Regexp.escape value
       end
     end
+
+    # Compares keys (for strings always true)
+    #
+    # @param other_target [Target]
+    # @ return [true, false]
+    def keys_equal?(other_target)
+      return false unless to_s.class.eql?(other_target.to_s.class)
+
+      kind_of?(Hash) ? to_s.keys.eql?(other_target.to_s.keys) : true
+    end
   end
 end

data/lib/regex_generator/version.rb

@@ -1,3 +1,3 @@
 module RegexGenerator
-  VERSION = '0.3.4'
+  VERSION = '0.3.5'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: regex_generator
 version: !ruby/object:Gem::Version
-  version: 0.3.4
+  version: 0.3.5
 platform: ruby
 authors:
 - o.vykhor
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-04-16 00:00:00.000000000 Z
+date: 2019-04-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler