Class: Newstile::Parser::Html::ElementConverter

Inherits:
Object
  • Object
show all
Includes:
Constants, Utils::Entities
Defined in:
lib/newstile/parser/html.rb

Overview

Converts HTML elements to native elements if possible.

Constant Summary collapse

REMOVE_TEXT_CHILDREN = 
%w{html head hgroup ol ul dl table colgroup tbody thead tfoot tr select optgroup}
WRAP_TEXT_CHILDREN = 
%w{body section nav article aside header footer address div li dd blockquote figure
figcaption fieldset form}
REMOVE_WHITESPACE_CHILDREN = 
%w{body section nav article aside header footer address
div li dd blockquote figure figcaption td th fieldset form}
STRIP_WHITESPACE = 
%w{address article aside blockquote body caption dd div dl dt fieldset figcaption form footer
header h1 h2 h3 h4 h5 h6 legend li nav p section td th}
SIMPLE_ELEMENTS = 
%w{em strong blockquote hr br img p thead tbody tfoot tr td th ul ol dl li dl dt dd}

Constants included from Utils::Entities

Utils::Entities::ENTITY_MAP, Utils::Entities::ENTITY_TABLE

Constants included from Constants

Constants::HTML_ATTRIBUTE_RE, Constants::HTML_BLOCK_ELEMENTS, Constants::HTML_COMMENT_RE, Constants::HTML_DOCTYPE_RE, Constants::HTML_ELEMENTS_WITHOUT_BODY, Constants::HTML_ENTITY_RE, Constants::HTML_INSTRUCTION_RE, Constants::HTML_PARSE_AS, Constants::HTML_PARSE_AS_BLOCK, Constants::HTML_PARSE_AS_RAW, Constants::HTML_PARSE_AS_SPAN, Constants::HTML_SPAN_ELEMENTS, Constants::HTML_TAG_CLOSE_RE, Constants::HTML_TAG_RE

Instance Method Summary collapse

Methods included from Utils::Entities

entity

Constructor Details

#initialize(doc) ⇒ ElementConverter

Returns a new instance of ElementConverter.



175
176
177
 
# File 'lib/newstile/parser/html.rb', line 175

def initialize(doc)
  @doc = doc
end

Instance Method Details

#convert_a(el) ⇒ Object 



315
316
317
318
319
320
321
322
 
# File 'lib/newstile/parser/html.rb', line 315

def convert_a(el)
  if el.attr['href']
    set_basics(el, :a, :span)
    process_children(el)
  else
    process_html_element(el, false)
  end
end

#convert_b(el) ⇒ Object 



324
325
326
327
 
# File 'lib/newstile/parser/html.rb', line 324

def convert_b(el)
  set_basics(el, :strong, :span)
  process_children(el)
end

#convert_code(el) ⇒ Object Also known as: convert_pre 



343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
 
# File 'lib/newstile/parser/html.rb', line 343

def convert_code(el)
  raw = ''
  extract_text(el, raw)
  result = process_text(raw, true)
  begin
    str = result.inject('') do |mem, c|
      if c.type == :text
        mem << c.value
      elsif c.type == :entity
        if RUBY_VERSION >= '1.9'
          mem << c.value.char.encode(@doc.parse_infos[:encoding])
        elsif [60, 62, 34, 38].include?(c.value.code_point)
          mem << c.value.code_point.chr
        end
      elsif c.type == :smart_quote || c.type == :typographic_sym
        mem << entity(c.value.to_s).char.encode(@doc.parse_infos[:encoding])
      else
        raise "Bug - please report"
      end
    end
    result.clear
    result << Element.new(:text, str)
  rescue
  end
  if result.length > 1 || result.first.type != :text
    process_html_element(el, false, true)
  else
    if el.value == 'code'
      set_basics(el, :codespan, :span)
    else
      set_basics(el, :codeblock, :block)
    end
    el.value = result.first.value
    el.children.clear
  end
end

#convert_div(el) ⇒ Object Also known as: convert_span 



430
431
432
433
434
435
436
 
# File 'lib/newstile/parser/html.rb', line 430

def convert_div(el)
  if !is_math_tag?(el)
    process_html_element(el)
  else
    handle_math_tag(el)
  end
end

#convert_h1(el) ⇒ Object 



334
335
336
337
338
 
# File 'lib/newstile/parser/html.rb', line 334

def convert_h1(el)
  set_basics(el, :header, :block, :level => el.value[1..1].to_i)
  extract_text(el, el.options[:raw_text] = '')
  process_children(el)
end

#convert_i(el) ⇒ Object 



329
330
331
332
 
# File 'lib/newstile/parser/html.rb', line 329

def convert_i(el)
  set_basics(el, :em, :span)
  process_children(el)
end

#convert_table(el) ⇒ Object 



381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
 
# File 'lib/newstile/parser/html.rb', line 381

def convert_table(el)
  if !is_simple_table?(el)
    process_html_element(el, false)
    return
  end
  process_children(el)
  set_basics(el, :table, :block)
  el.options[:alignment] = []
  calc_alignment = lambda do |c|
    if c.type == :tr && el.options[:alignment].empty?
      el.options[:alignment] = [:default] * c.children.length
      break
    else
      c.children.each {|cc| calc_alignment.call(cc)}
    end
  end
  calc_alignment.call(el)
  if el.children.first.type == :tr
    tbody = Element.new(:tbody, nil, nil, :category => :block)
    tbody.children = el.children
    el.children = [tbody]
  end
end

#extract_text(el, raw) ⇒ Object 



310
311
312
313
 
# File 'lib/newstile/parser/html.rb', line 310

def extract_text(el, raw)
  raw << el.value.to_s if el.type == :text
  el.children.each {|c| extract_text(c, raw)}
end

#handle_math_tag(el) ⇒ Object 



444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
 
# File 'lib/newstile/parser/html.rb', line 444

def handle_math_tag(el)
  set_basics(el, :math, (el.value == 'div' ? :block : :span))
  el.value = el.children.shift.value
  if el.attr['class'] =~ /^\s*math\s*$/
    el.attr.delete('class')
  else
    el.attr['class'].sub!(/\s?math/, '')
  end
  el.value.gsub!(/&(amp|quot|gt|lt);/) do |m|
    case m
    when '&amp;'   then '&'
    when '&quot;'  then '"'
    when '&gt;'    then '>'
    when '&lt;'    then '<'
    end
  end
end

#is_math_tag?(el) ⇒ Boolean

Returns:

  • (Boolean) 


439
440
441
442
 
# File 'lib/newstile/parser/html.rb', line 439

def is_math_tag?(el)
  el.attr['class'].to_s =~ /\bmath\b/ &&
    el.children.size == 1 && el.children.first.type == :text
end

#is_simple_table?(el) ⇒ Boolean

Returns:

  • (Boolean) 


405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
 
# File 'lib/newstile/parser/html.rb', line 405

def is_simple_table?(el)
  only_phrasing_content = lambda do |c|
    c.children.all? do |cc|
      (cc.type == :text || !HTML_BLOCK_ELEMENTS.include?(cc.value)) && only_phrasing_content.call(cc)
    end
  end
  check_cells = Proc.new do |c|
    if c.value == 'th' || c.value == 'td'
      return false if !only_phrasing_content.call(c)
    else
      c.children.each {|cc| check_cells.call(cc)}
    end
  end
  check_cells.call(el)

  check_rows = lambda do |t, type|
    t.children.all? {|r| (r.value == 'tr' || r.type == :text) && r.children.all? {|c| c.value == type || c.type == :text}}
  end
  check_rows.call(el, 'td') ||
    (el.children.all? do |t|
       t.type == :text || (t.value == 'thead' && check_rows.call(t, 'th')) ||
         ((t.value == 'tfoot' || t.value == 'tbody') && check_rows.call(t, 'td'))
     end && el.children.any? {|t| t.value == 'tbody'})
end

#process(el, do_conversion = true, preserve_text = false, parent = nil) ⇒ Object

Convert the element el and its children.



180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
 
# File 'lib/newstile/parser/html.rb', line 180

def process(el, do_conversion = true, preserve_text = false, parent = nil)
  case el.type
  when :xml_comment, :xml_pi, :html_doctype
    ptype = if parent.nil?
              'div'
            else
              case parent.type
              when :html_element then parent.value
              when :code_span then 'code'
              when :code_block then 'pre'
              when :header then 'h1'
              else parent.type.to_s
              end
            end
    el.options = {:category => HTML_PARSE_AS_SPAN.include?(ptype) ? :span : :block}
    return
  when :html_element
  else return
  end

  type = el.value
  remove_text_children(el) if REMOVE_TEXT_CHILDREN.include?(type)

  mname = "convert_#{el.value}"
  if do_conversion && self.class.method_defined?(mname)
    send(mname, el)
  elsif do_conversion && SIMPLE_ELEMENTS.include?(type)
    set_basics(el, type.intern, HTML_SPAN_ELEMENTS.include?(type) ? :span : :block)
    process_children(el, do_conversion, preserve_text)
  else
    process_html_element(el, do_conversion, preserve_text)
  end

  strip_whitespace(el) if STRIP_WHITESPACE.include?(type)
  remove_whitespace_children(el) if REMOVE_WHITESPACE_CHILDREN.include?(type)
  wrap_text_children(el) if WRAP_TEXT_CHILDREN.include?(type)
end

#process_children(el, do_conversion = true, preserve_text = false) ⇒ Object 



218
219
220
221
222
223
224
225
226
227
 
# File 'lib/newstile/parser/html.rb', line 218

def process_children(el, do_conversion = true, preserve_text = false)
  el.children.map! do |c|
    if c.type == :text
      process_text(c.value, preserve_text)
    else
      process(c, do_conversion, preserve_text, el)
      c
    end
  end.flatten!
end

#process_html_element(el, do_conversion = true, preserve_text = false) ⇒ Object 



254
255
256
257
258
259
 
# File 'lib/newstile/parser/html.rb', line 254

def process_html_element(el, do_conversion = true, preserve_text = false)
  el.options = {:category => HTML_SPAN_ELEMENTS.include?(el.value) ? :span : :block,
    :parse_type => HTML_PARSE_AS[el.value]
  }
  process_children(el, do_conversion, preserve_text)
end

#process_text(raw, preserve = false) ⇒ Object

Process the HTML text raw: compress whitespace (if preserve is false) and convert entities in entity elements.



231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
 
# File 'lib/newstile/parser/html.rb', line 231

def process_text(raw, preserve = false)
  raw.gsub!(/\s+/, ' ') unless preserve
  src = StringScanner.new(raw)
  result = []
  while !src.eos?
    if tmp = src.scan_until(/(?=#{HTML_ENTITY_RE})/)
      result << Element.new(:text, tmp)
      src.scan(HTML_ENTITY_RE)
      val = src[1] || (src[2] && src[2].to_i) || src[3].hex
      result << if %w{lsquo rsquo ldquo rdquo}.include?(val)
                  Element.new(:smart_quote, val.intern)
                elsif %w{mdash ndash hellip laquo raquo}.include?(val)
                  Element.new(:typographic_sym, val.intern)
                else
                  Element.new(:entity, entity(val), nil, :original => src.matched)
                end
    else
      result << Element.new(:text, src.scan(/.*/m))
    end
  end
  result
end

#remove_text_children(el) ⇒ Object 



261
262
263
 
# File 'lib/newstile/parser/html.rb', line 261

def remove_text_children(el)
  el.children.delete_if {|c| c.type == :text}
end

#remove_whitespace_children(el) ⇒ Object 



294
295
296
297
298
299
300
301
302
 
# File 'lib/newstile/parser/html.rb', line 294

def remove_whitespace_children(el)
  i = -1
  el.children.delete_if do |c|
    i += 1
    c.type == :text && c.value.strip.empty? &&
      (i == 0 || i == el.children.length - 1 || (el.children[i-1].options[:category] == :block &&
                                                 el.children[i+1].options[:category] == :block))
  end
end

#set_basics(el, type, category, opts = {}) ⇒ Object 



304
305
306
307
308
 
# File 'lib/newstile/parser/html.rb', line 304

def set_basics(el, type, category, opts = {})
  el.type = type
  el.options = {:category => category}.merge(opts)
  el.value = nil
end

#strip_whitespace(el) ⇒ Object 



284
285
286
287
288
289
290
291
292
 
# File 'lib/newstile/parser/html.rb', line 284

def strip_whitespace(el)
  return if el.children.empty?
  if el.children.first.type == :text
    el.children.first.value.lstrip!
  end
  if el.children.last.type == :text
    el.children.last.value.rstrip!
  end
end

#wrap_text_children(el) ⇒ Object 



265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
 
# File 'lib/newstile/parser/html.rb', line 265

def wrap_text_children(el)
  tmp = []
  last_is_p = false
  el.children.each do |c|
    if c.options[:category] != :block || c.type == :text
      if !last_is_p
        tmp << Element.new(:p, nil, nil, :transparent => true)
        last_is_p = true
      end
      tmp.last.children << c
      tmp
    else
      tmp << c
      last_is_p = false
    end
  end
  el.children = tmp
end