stream_parser 0.2 → 0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1b99a423c7791ac03b0e45b390d22c10b455b45e51066edb1912eb4c9aec7a65
-  data.tar.gz: edb3c225aeaff8d4400e0929a8bde7120326ab95c39fe42c11c8d4bf6f8ba2ee
+  metadata.gz: 1467f9df4e5b2bcf033a774a97ea24f61635a8249be9a1c7daad54d77251ede6
+  data.tar.gz: dbb178a552f4d194b25a1df946bd10ae546225bc140ce5047695370310562937
 SHA512:
-  metadata.gz: 24ae770c7090a31b80abc9b7c216de440e61ca38d87476e17ef855ba37aa90a30bd1cb7f990831b0ebd68fead35a385d98c7307b608e9f34759d54dfcfd43fe3
-  data.tar.gz: 2ce893bdd9e79aa3b094ba15bc612ac598f1e11b0325bf5fa3e44a521ff678ada20705e805cef40314eeb4a274d7e877603b5dfc4ee6ffbebc96b23980e370aa
+  metadata.gz: faadb8fe5f2706851da234ff47e7a6d01f78a13cc06f663d213e31cf179c35773ef30ae483619070822dad456500e66e1b5a22ba98eb4afb5935bdb6f8fd715f
+  data.tar.gz: 56ace2ac66d7ccdbf217d1e21250cc3d91f54320501a7d75232f613e49fb3d8920eddef54eeb6f20797c9cf5e57807256119a4fc92c0f5676624a25ef556cde3

data/README.md

@@ -55,5 +55,7 @@ QuotedStringFinder.parse(%q{Here "ar})
 # => SyntaxError "Unbalanced Quotes in string"
 ```
 
+Although we grab quoted values ourselfs in this example there is a `quoted_value` helper as well as a
+`StreamParser::HTML` which provides additional helpers such as `next_tag`, `scan_for_tag`, `next_end_tag` and others.
 
 

data/lib/stream_parser/html/tag.rb

@@ -0,0 +1,39 @@
+class StreamParser::HTML::Tag
+  attr_reader :name, :attributes
+  attr_writer :self_closing
+  
+  def initialize(name, closing=false)
+    @name = name
+    @attributes = {}
+    @closing = closing
+    @self_closing = false
+  end
+
+  def [](key)
+    @attributes[key.to_sym]
+  end
+
+  def []=(key, value)
+    @attributes[key.to_sym] = value
+  end
+  
+  def self_closing?
+    @self_closing
+  end
+  
+  def closing?
+    @closing
+  end
+  
+  def opening?
+    !@closing
+  end
+  
+  def match(name: nil, closing: nil, attributes: nil)
+    return false if name && @name != name
+    return false if !closing.nil? && @closing != closing
+    return false if attributes && !attributes.all? { |k,v| @attributes[k] == v }
+    
+    true
+  end
+end

data/lib/stream_parser/html.rb

@@ -0,0 +1,112 @@
+module StreamParser::HTML
+  
+  autoload :Tag, File.expand_path('../html/tag', __FILE__)
+  
+  def self.included(base)
+    base.include(StreamParser)
+  end
+  
+  def next_tag(old_index: nil)
+    old_index ||= @index
+    return unless scan_until(/<\s*/)
+    start_index = @index-1
+
+    while peek(3) == '!--'
+      forward(3)
+      scan_until(/-->\s*/)
+      scan_until(/<\s*/)
+    end
+    
+    # HTMLComment.new(pre_match)
+    if peek(1) == '/'
+      scan_until(/[^>\s\/]+/)
+      scan_tag(match, old_index: old_index, start_index: start_index, closing: true)
+    else
+      scan_until(/[^>\s\/]+/)
+      scan_tag(match, old_index: old_index, start_index: start_index) 
+    end
+  end
+  
+  def scan_for_tag(name, closing: nil, **attributes)
+    old_index ||= @index
+    tag = next_tag
+    while tag && !tag.match(name: name, closing: closing, attributes: attributes)
+      tag = next_tag(old_index: old_index)
+    end
+    tag
+  end
+  
+  def scan_for_closing_tag
+    old_index = @index
+    heap = []
+    
+    tag = next_tag
+    puts tag.inspect
+    while tag && !(tag.closing? && heap.empty?)
+      if !tag.closing? && !tag.self_closing?
+        heap << tag
+      elsif !tag.self_closing?
+        heap.pop
+      end
+      tag = next_tag(old_index: old_index)
+    end
+    @old_index = old_index
+    tag
+  end
+  
+  def scan_tag(name, closing: false, old_index:, start_index:)
+    tag = Tag.new(name, closing)
+    
+    while !eos?
+      gobble(/\s+/)
+      key = case peek(1)
+      when '>'
+        forward(1)
+        @old_index = old_index
+        @match = @source[start_index...@index]
+        return tag
+      when '/'
+        forward(1)
+        gobble(/\s*\>/)
+        @old_index = old_index
+        @match = @source[start_index...@index]
+        tag.self_closing = true
+        return tag
+      when '"', "'"
+        quote_char = next_char
+        forward(1)
+        quoted_value(quote_char)
+      else
+        scan_until(/[^>\s\/=]+/)[0]
+      end
+      
+      tag[key] = if next?(/\s*=/)
+        gobble(/\s*=/)
+        html_tag_value
+      else
+        true
+      end
+    end
+
+    @old_index = old_index
+    @match = @source[start_index...@index]
+    tag
+  end
+  
+  def html_tag_value
+    gobble(/\s+/)
+    case peek(1)
+    when '"', "'"
+      quote_char = next_char
+      forward(1)
+      quoted_value(quote_char)
+    else
+      scan_until(/[^>\s\/=]+/)[0]
+    end
+  end
+  
+  def next_end_tag(name)
+    scan_until(/<\/\s*li>/)
+  end
+  
+end

data/lib/stream_parser/syntax_error.rb

@@ -0,0 +1 @@
+class StreamParser::SyntaxError < StandardError; end

data/lib/stream_parser/version.rb

@@ -1,3 +1,3 @@
 module StreamParser
-  VERSION = '0.2'
+  VERSION = '0.4'
 end

data/lib/stream_parser.rb

@@ -1,5 +1,8 @@
 module StreamParser
 
+  autoload :HTML, 'stream_parser/html'
+  autoload :SyntaxError, 'stream_parser/syntax_error'
+  
   def self.included(base)
     base.extend ClassMethods
   end
@@ -29,8 +32,8 @@ module StreamParser
 
   def scan_until(r)
     r = Regexp.new(Regexp.escape(r)) if r.is_a?(String)
-    index = @source.index(r, @index)
     match = @source.match(r, @index)
+    index = match&.begin(0)
     
     if match
       @match = match.to_s
@@ -43,6 +46,13 @@ module StreamParser
     end
     match
   end
+  
+  def gobble(r)
+    m = @source.match(r, @index)
+    if m&.begin(0) == @index
+      scan_until(r)
+    end
+  end
 
   def pre_match
     @source[@old_index...(@index-(@match&.size || 0))]
@@ -70,6 +80,18 @@ module StreamParser
     @source[@index-1]
   end
   
+  def next?(r)
+    @source.match(r, @index)&.begin(0) == @index
+  end
+
+  def peek(n=1)
+    if n.is_a?(Regexp)
+      @source.match(n, @index)
+    else
+      @source.slice(@index, n)
+    end
+  end
+
   def next_word
     nw = @source.match(/\s*(\S+)/, @index)
     nw.nil? ? nil : nw[1]
@@ -107,7 +129,7 @@ module StreamParser
     ret_value = ""
     while scan_until(/(#{quote_char}|\Z)/)
       if match != quote_char
-        raise Net::HTTPHeaderSyntaxError.new("Invalid Set-Cookie header format: unbalanced quotes (#{quote_char})")
+        raise StreamParser::SyntaxError.new("Unbalanced quotes #{quote_char}")
       elsif !escape_chars.include?(pre_match[-1])
         ret_value << pre_match
         return ret_value

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: stream_parser
 version: !ruby/object:Gem::Version
-  version: '0.2'
+  version: '0.4'
 platform: ruby
 authors:
 - Jon Bracy
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-05-07 00:00:00.000000000 Z
+date: 2024-04-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -118,6 +118,9 @@ extra_rdoc_files:
 files:
 - README.md
 - lib/stream_parser.rb
+- lib/stream_parser/html.rb
+- lib/stream_parser/html/tag.rb
+- lib/stream_parser/syntax_error.rb
 - lib/stream_parser/version.rb
 homepage: https://github.com/malomalo/stream_parser
 licenses:
@@ -140,7 +143,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.3
+rubygems_version: 3.5.4
 signing_key:
 specification_version: 4
 summary: SAX/Stream style parse helpers