Class: DQReadability::Document
- Inherits:
-
Object
- Object
- DQReadability::Document
- Defined in:
- lib/dq-readability.rb
Constant Summary collapse
- DEFAULT_OPTIONS =
{ :retry_length => 250, :min_text_length => 25, :remove_unlikely_candidates => true, :weight_classes => true, :clean_conditionally => true, :remove_empty_nodes => true, :min_image_width => 130, :min_image_height => 80, :ignore_image_format => [], :bypass => false }.freeze
- REGEXES =
{ :unlikelyCandidatesRe => /combx|comment|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup/i, :okMaybeItsACandidateRe => /and|article|body|column|main|shadow/i, :positiveRe => /article|body|content|entry|hentry|main|page|pagination|post|text|blog|story/i, :negativeRe => /combx|comment|com-|contact|foot|footer|footnote|masthead|media|meta|outbrain|promo|related|scroll|shoutbox|sidebar|sponsor|shopping|tags|tool|widget/i, :divToPElementsRe => /<(a|blockquote|dl|div|img|ol|p|pre|table|ul)/i, :replaceBrsRe => /(<br[^>]*>[ \n\r\t]*){2,}/i, :replaceFontsRe => /<(\/?)font[^>]*>/i, :trimRe => /^\s+|\s+$/, :normalizeRe => /\s{2,}/, :killBreaksRe => /(<br\s*\/?>(\s| ?)*){1,}/, :videoRe => /http:\/\/(www\.)?(youtube|vimeo)\.com/i }
Instance Attribute Summary collapse
-
#best_candidate ⇒ Object
Returns the value of attribute best_candidate.
-
#best_candidate_has_image ⇒ Object
Returns the value of attribute best_candidate_has_image.
-
#candidates ⇒ Object
Returns the value of attribute candidates.
-
#html ⇒ Object
Returns the value of attribute html.
-
#options ⇒ Object
Returns the value of attribute options.
Instance Method Summary collapse
-
#author ⇒ Object
Look through the @html document looking for the author Precedence Information here on the wiki: (TODO attach wiki URL if it is accepted) Returns nil if no author is detected.
- #class_weight(e) ⇒ Object
- #clean_conditionally(node, candidates, selector) ⇒ Object
- #content(remove_unlikely_candidates = :default) ⇒ Object
- #debug(str) ⇒ Object
- #get_article(candidates, best_candidate) ⇒ Object
- #get_image_size(url) ⇒ Object
- #get_link_density(elem) ⇒ Object
- #image_meets_criteria?(image) ⇒ Boolean
- #images(content = nil, reload = false) ⇒ Object
- #images_with_fqdn_uris(document = @html.dup, source_uri) ⇒ Object
- #images_with_fqdn_uris!(source_uri) ⇒ Object
-
#initialize(input, options = {}) ⇒ Document
constructor
A new instance of Document.
- #make_html ⇒ Object
- #prepare_candidates ⇒ Object
- #remove_unlikely_candidates! ⇒ Object
- #sanitize(node, candidates, options = {}) ⇒ Object
- #score_node(elem) ⇒ Object
- #score_paragraphs(min_text_length) ⇒ Object
- #select_best_candidate(candidates) ⇒ Object
- #title ⇒ Object
- #transform_misused_divs_into_paragraphs! ⇒ Object
Constructor Details
#initialize(input, options = {}) ⇒ Document
Returns a new instance of Document.
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
# File 'lib/dq-readability.rb', line 40 def initialize(input, = {}) @options = DEFAULT_OPTIONS.merge() @input = open(input).read @url = input if RUBY_VERSION =~ /^(1\.9|2)/ && !@options[:encoding] @input = GuessHtmlEncoding.encode(@input, @options[:html_headers]) unless @options[:do_not_guess_encoding] @options[:encoding] = @input.encoding.to_s end @input = @input.gsub(REGEXES[:replaceBrsRe], '</p><p>').gsub(REGEXES[:replaceFontsRe], '<\1span>') @remove_unlikely_candidates = @options[:remove_unlikely_candidates] @weight_classes = @options[:weight_classes] @clean_conditionally = @options[:clean_conditionally] @best_candidate_has_image = true @bypass = @options[:bypass] make_html end |
Instance Attribute Details
#best_candidate ⇒ Object
Returns the value of attribute best_candidate.
38 39 40 |
# File 'lib/dq-readability.rb', line 38 def best_candidate @best_candidate end |
#best_candidate_has_image ⇒ Object
Returns the value of attribute best_candidate_has_image.
38 39 40 |
# File 'lib/dq-readability.rb', line 38 def best_candidate_has_image @best_candidate_has_image end |
#candidates ⇒ Object
Returns the value of attribute candidates.
38 39 40 |
# File 'lib/dq-readability.rb', line 38 def candidates @candidates end |
#html ⇒ Object
Returns the value of attribute html.
38 39 40 |
# File 'lib/dq-readability.rb', line 38 def html @html end |
#options ⇒ Object
Returns the value of attribute options.
38 39 40 |
# File 'lib/dq-readability.rb', line 38 def @options end |
Instance Method Details
#author ⇒ Object
Look through the @html document looking for the author Precedence Information here on the wiki: (TODO attach wiki URL if it is accepted) Returns nil if no author is detected
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 |
# File 'lib/dq-readability.rb', line 260 def # Let's grab this author: # <meta name="dc.creator" content="Finch - http://www.getfinch.com" /> = @html.xpath('//meta[@name = "dc.creator"]') unless .empty? .each do |element| if element['content'] return element['content'].strip end end end # Now let's try to grab this # <span class="byline author vcard"><span>By</span><cite class="fn">Austin Fonacier</cite></span> # <div class="author">By</div><div class="author vcard"><a class="url fn" href="http://austinlivesinyoapp.com/">Austin Fonacier</a></div> = @html.xpath('//*[contains(@class, "vcard")]//*[contains(@class, "fn")]') unless .empty? .each do |element| if element.text return element.text.strip end end end # Now let's try to grab this # <a rel="author" href="http://dbanksdesign.com">Danny Banks (rel)</a> # TODO: strip out the (rel)? = @html.xpath('//a[@rel = "author"]') unless .empty? .each do |element| if element.text return element.text.strip end end end = @html.xpath('//*[@id = "author"]') unless .empty? .each do |element| if element.text return element.text.strip end end end end |
#class_weight(e) ⇒ Object
423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 |
# File 'lib/dq-readability.rb', line 423 def class_weight(e) weight = 0 return weight unless @weight_classes if e[:class] && e[:class] != "" if e[:class] =~ REGEXES[:negativeRe] weight -= 25 end if e[:class] =~ REGEXES[:positiveRe] weight += 25 end end if e[:id] && e[:id] != "" if e[:id] =~ REGEXES[:negativeRe] weight -= 25 end if e[:id] =~ REGEXES[:positiveRe] weight += 25 end end weight end |
#clean_conditionally(node, candidates, selector) ⇒ Object
568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 |
# File 'lib/dq-readability.rb', line 568 def clean_conditionally(node, candidates, selector) return unless @clean_conditionally node.css(selector).each do |el| weight = class_weight(el) content_score = candidates[el] ? candidates[el][:content_score] : 0 name = el.name.downcase if weight + content_score < 0 el.remove debug("Conditionally cleaned #{name}##{el[:id]}.#{el[:class]} with weight #{weight} and content score #{content_score} because score + content score was less than zero.") elsif el.text.count(",") < 10 counts = %w[p img li a embed input].inject({}) { |m, kind| m[kind] = el.css(kind).length; m } counts["li"] -= 100 # For every img under a noscript tag discount one from the count to avoid double counting counts["img"] -= el.css("noscript").css("img").length content_length = el.text.strip.length # Count the text length excluding any surrounding whitespace link_density = get_link_density(el) to_remove = false reason = "" if (counts["img"] > counts["p"]+2) reason = "too many images" to_remove = true elsif counts["li"] > counts["p"] && name != "ul" && name != "ol" reason = "more <li>s than <p>s" to_remove = true elsif counts["input"] > (counts["p"] / 3).to_i reason = "less than 3x <p>s than <input>s" to_remove = true elsif (content_length < [:min_text_length]) && (counts["img"] != 1) reason = "too short a content length without a single image" to_remove = true elsif weight < 25 && link_density > 0.2 reason = "too many links for its weight (#{weight})" to_remove = true elsif weight >= 25 && link_density > 0.5 reason = "too many links for its weight (#{weight})" to_remove = true elsif (counts["embed"] == 1 && content_length < 75) || counts["embed"] > 1 reason = "<embed>s with too short a content length, or too many <embed>s" to_remove = true end if to_remove debug("Conditionally cleaned #{name}##{el[:id]}.#{el[:class]} with weight #{weight} and content score #{content_score} because it has #{reason}.") el.remove end end end end |
#content(remove_unlikely_candidates = :default) ⇒ Object
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 |
# File 'lib/dq-readability.rb', line 306 def content(remove_unlikely_candidates = :default) if @bypass == false @remove_unlikely_candidates = false if remove_unlikely_candidates == false prepare_candidates article = get_article(@candidates, @best_candidate) cleaned_article = sanitize(article, @candidates, ) if article.text.strip.length < [:retry_length] if @remove_unlikely_candidates @remove_unlikely_candidates = false elsif @weight_classes @weight_classes = false elsif @clean_conditionally @clean_conditionally = false else # nothing we can do return cleaned_article end make_html content else cleaned_article end else make_html s = Nokogiri::XML::Node::SaveOptions save_opts = s::NO_DECLARATION | s::NO_EMPTY_TAGS | s::AS_XHTML html = @html.serialize(:save_with => save_opts) return html end end |
#debug(str) ⇒ Object
465 466 467 |
# File 'lib/dq-readability.rb', line 465 def debug(str) puts str if [:debug] end |
#get_article(candidates, best_candidate) ⇒ Object
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 |
# File 'lib/dq-readability.rb', line 340 def get_article(candidates, best_candidate) # Now that we have the top candidate, look through its siblings for content that might also be related. # Things like preambles, content split by ads that we removed, etc. sibling_score_threshold = [10, best_candidate[:content_score] * 0.2].max output = Nokogiri::XML::Node.new('div', @html) best_candidate[:elem].parent.children.each do |sibling| append = false append = true if sibling == best_candidate[:elem] append = true if candidates[sibling] && candidates[sibling][:content_score] >= sibling_score_threshold if sibling.name.downcase == "p" link_density = get_link_density(sibling) node_content = sibling.text node_length = node_content.length if node_length > 80 && link_density < 0.25 append = true elsif node_length < 80 && link_density == 0 && node_content =~ /\.( |$)/ append = true end end if append sibling_dup = sibling.dup # otherwise the state of the document in processing will change, thus creating side effects sibling_dup.name = "div" unless %w[div p].include?(sibling.name.downcase) output << sibling_dup end end output end |
#get_image_size(url) ⇒ Object
236 237 238 239 240 241 242 243 244 245 |
# File 'lib/dq-readability.rb', line 236 def get_image_size(url) begin w, h = FastImage.size(url) raise "Couldn't get size." if w.nil? || h.nil? {:width => w, :height => h} rescue => e debug("Image error: #{e}") nil end end |
#get_link_density(elem) ⇒ Object
387 388 389 390 391 |
# File 'lib/dq-readability.rb', line 387 def get_link_density(elem) link_length = elem.css("a").map(&:text).join("").length text_length = elem.text.length link_length / text_length.to_f end |
#image_meets_criteria?(image) ⇒ Boolean
247 248 249 250 |
# File 'lib/dq-readability.rb', line 247 def image_meets_criteria?(image) return false if [:ignore_image_format].include?(image[:format].downcase) image[:width] >= ([:min_image_width] || 0) && image[:height] >= ([:min_image_height] || 0) end |
#images(content = nil, reload = false) ⇒ Object
162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 |
# File 'lib/dq-readability.rb', line 162 def images(content=nil, reload=false) begin require 'fastimage' rescue LoadError raise "Please install fastimage in order to use the #images feature." end @best_candidate_has_image = false if reload prepare_candidates list_images = [] tested_images = [] content = @best_candidate[:elem] unless reload return list_images if content.nil? elements = content.css("img").map(&:attributes) elements.each do |element| next unless element["src"] url = element["src"].value height = element["height"].nil? ? 0 : element["height"].value.to_i width = element["width"].nil? ? 0 : element["width"].value.to_i if url =~ /\Ahttps?:\/\//i && (height.zero? || width.zero?) image = get_image_size(url) next unless image else image = {:width => width, :height => height} end image[:format] = File.extname(url).gsub(".", "") if tested_images.include?(url) debug("Image was tested: #{url}") next end tested_images.push(url) if image_meets_criteria?(image) list_images << url else debug("Image discarded: #{url} - height: #{image[:height]} - width: #{image[:width]} - format: #{image[:format]}") end end (list_images.empty? and content != @html) ? images(@html, true) : list_images end |
#images_with_fqdn_uris(document = @html.dup, source_uri) ⇒ Object
215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 |
# File 'lib/dq-readability.rb', line 215 def images_with_fqdn_uris(document = @html.dup, source_uri) uri = URI.parse(source_uri) host = uri.host scheme = uri.scheme port = uri.port # defaults to 80 base = "#{scheme}://#{host}:#{port}/" images = [] document.css("img").each do |elem| begin elem['src'] = URI.join(base,elem['src']).to_s if URI.parse(elem['src']).host == nil images << elem['src'].to_s rescue URI::InvalidURIError => exc elem.remove end end images(document,true) end |
#images_with_fqdn_uris!(source_uri) ⇒ Object
211 212 213 |
# File 'lib/dq-readability.rb', line 211 def images_with_fqdn_uris!(source_uri) images_with_fqdn_uris(@html, source_uri) end |
#make_html ⇒ Object
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 |
# File 'lib/dq-readability.rb', line 68 def make_html @html = Nokogiri::HTML(@input, nil, @options[:encoding]) # In case document has no body, such as from empty string or redirect @html = Nokogiri::HTML('<body />', nil, @options[:encoding]) if @html.css('body').length == 0 # Remove html comment tags @html.xpath('//comment()').each { |i| i.remove } # making all the headings of same format @html.css("h1").each do |h| h.name = "h2" end @html.css("h2").each do |h| h.name = "h3" end @html.css("h4").each do |h| h.name = "h3" end uri = URI.parse(@url) host = uri.host scheme = uri.scheme port = uri.port # defaults to 80 base = "#{scheme}://#{host}:#{port}/" # changing img src @html.css("img").each do |elem| begin if elem['src'][0] == '/' if elem['src'][1] == '/' elem['src'] = 'http:'+elem['src'] else elem['src'] = URI.join(base,elem['src']).to_s if URI.parse(elem['src']).host == nil end else if @url.split('').last == '/' elem['src'] = URI.join(@url,elem['src']).to_s if URI.parse(elem['src']).host == nil else x = @url.split('/') x.delete_at(x.length-1) y = '' x.each{|i| y += i+'/'} elem['src'] = URI.join(y,elem['src']).to_s if URI.parse(elem['src']).host == nil end end rescue elem.remove end end # changing certain tags to <p> tags x = @html.css("ol") x.each do |t| t.name = "p" end len = @html.css('ol').length debug("length of ol tag #{len}") #changing the 'a' href @html.css("a").each do |elem| begin if elem['href'][0] == '/' elem['href'] = URI.join(base,elem['href']).to_s if URI.parse(elem['href']).host == nil else if @url.split('').last == '/' elem['href'] = URI.join(@url,elem['href']).to_s if URI.parse(elem['href']).host == nil else x = @url.split('/') x.delete_at(x.length-1) y = '' x.each{|i| y += i+'/'} elem['href'] = URI.join(y,elem['href']).to_s if URI.parse(elem['href']).host == nil end end rescue end end # removing edit spans @html.css('span').each do |elem| if elem.text.downcase == "[edit]" elem.remove end end end |
#prepare_candidates ⇒ Object
59 60 61 62 63 64 65 66 |
# File 'lib/dq-readability.rb', line 59 def prepare_candidates @html.css("script, style").each { |i| i.remove } remove_unlikely_candidates! if @remove_unlikely_candidates transform_misused_divs_into_paragraphs! @candidates = score_paragraphs([:min_text_length]) @best_candidate = select_best_candidate(@candidates) end |
#remove_unlikely_candidates! ⇒ Object
469 470 471 472 473 474 475 476 477 |
# File 'lib/dq-readability.rb', line 469 def remove_unlikely_candidates! @html.css("*").each do |elem| str = "#{elem[:class]}#{elem[:id]}" if str =~ REGEXES[:unlikelyCandidatesRe] && str !~ REGEXES[:okMaybeItsACandidateRe] && (elem.name.downcase != 'html') && (elem.name.downcase != 'body') debug("Removing unlikely candidate - #{str}") elem.remove end end end |
#sanitize(node, candidates, options = {}) ⇒ Object
499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 |
# File 'lib/dq-readability.rb', line 499 def sanitize(node, candidates, = {}) node.css("h1, h2, h3, h4, h5, h6").each do |header| header.remove if class_weight(header) < 0 || get_link_density(header) > 0.33 end node.css("form, object, iframe, embed").each do |elem| elem.remove end # if @options[:remove_empty_nodes] # # remove <p> tags that have no text content - this will also remove p tags that contain only images. # node.css("p").each do |elem| # elem.remove if elem.content.strip.empty? # end # end # Conditionally clean <table>s, <ul>s, and <div>s clean_conditionally(node, candidates, "table, ul, div") # We'll sanitize all elements using a whitelist base_whitelist = @options[:tags] || %w[div p] # We'll add whitespace instead of block elements, # so a<br>b will have a nice space between them base_replace_with_whitespace = %w[br hr h1 h2 h3 h4 h5 h6 dl dd ol li ul address blockquote center] # Use a hash for speed (don't want to make a million calls to include?) whitelist = Hash.new base_whitelist.each {|tag| whitelist[tag] = true } replace_with_whitespace = Hash.new base_replace_with_whitespace.each { |tag| replace_with_whitespace[tag] = true } ([node] + node.css("*")).each do |el| # If element is in whitelist, delete all its attributes if whitelist[el.node_name] el.attributes.each { |a, x| el.delete(a) unless @options[:attributes] && @options[:attributes].include?(a.to_s) } # Otherwise, replace the element with its contents else # If element is root, replace the node as a text node if el.parent.nil? node = Nokogiri::XML::Text.new(el.text, el.document) break else if replace_with_whitespace[el.node_name] el.swap(Nokogiri::XML::Text.new(' ' << el.text << ' ', el.document)) else el.swap(Nokogiri::XML::Text.new(el.text, el.document)) end end end end s = Nokogiri::XML::Node::SaveOptions save_opts = s::NO_DECLARATION | s::NO_EMPTY_TAGS | s::AS_XHTML html = node.serialize(:save_with => save_opts) # Get rid of duplicate whitespace html = "<head><meta http-equiv='Content-Type' content='text/html; charset=utf-8'></head>" + "\n" + html.gsub(/[\r\n\f]+/, "\n" ) # get rid of incompitable characters if html.encode('utf-8').include?('Â') html = html.encode('utf-8').gsub('Â',' ') end return html end |
#score_node(elem) ⇒ Object
450 451 452 453 454 455 456 457 458 459 460 461 462 463 |
# File 'lib/dq-readability.rb', line 450 def score_node(elem) content_score = class_weight(elem) case elem.name.downcase when "div" content_score += 5 when "blockquote" content_score += 3 when "form" content_score -= 3 when "th" content_score -= 5 end { :content_score => content_score, :elem => elem } end |
#score_paragraphs(min_text_length) ⇒ Object
393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 |
# File 'lib/dq-readability.rb', line 393 def score_paragraphs(min_text_length) candidates = {} @html.css("p,td").each do |elem| parent_node = elem.parent grand_parent_node = parent_node.respond_to?(:parent) ? parent_node.parent : nil inner_text = elem.text # If this paragraph is less than 25 characters, don't even count it. next if inner_text.length < min_text_length candidates[parent_node] ||= score_node(parent_node) candidates[grand_parent_node] ||= score_node(grand_parent_node) if grand_parent_node content_score = 1 content_score += inner_text.split(',').length content_score += [(inner_text.length / 100).to_i, 3].min candidates[parent_node][:content_score] += content_score candidates[grand_parent_node][:content_score] += content_score / 2.0 if grand_parent_node end # Scale the final candidates score based on link density. Good content should have a # relatively small link density (5% or less) and be mostly unaffected by this operation. candidates.each do |elem, candidate| candidate[:content_score] = candidate[:content_score] * (1 - get_link_density(elem)) end candidates end |
#select_best_candidate(candidates) ⇒ Object
373 374 375 376 377 378 379 380 381 382 383 384 385 |
# File 'lib/dq-readability.rb', line 373 def select_best_candidate(candidates) sorted_candidates = candidates.values.sort { |a, b| b[:content_score] <=> a[:content_score] } debug("Top 5 candidates:") sorted_candidates[0...5].each do |candidate| debug("Candidate #{candidate[:elem].name}##{candidate[:elem][:id]}.#{candidate[:elem][:class]} with score #{candidate[:content_score]}") end best_candidate = sorted_candidates.first || { :elem => @html.css("body").first, :content_score => 0 } debug("Best candidate #{best_candidate[:elem].name}##{best_candidate[:elem][:id]}.#{best_candidate[:elem][:class]} with score #{best_candidate[:content_score]}") best_candidate end |
#title ⇒ Object
252 253 254 255 |
# File 'lib/dq-readability.rb', line 252 def title title = @html.css("title").first title ? title.text : nil end |
#transform_misused_divs_into_paragraphs! ⇒ Object
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 |
# File 'lib/dq-readability.rb', line 479 def transform_misused_divs_into_paragraphs! @html.css("*").each do |elem| if elem.name.downcase == "div" # transform <div>s that do not contain other block elements into <p>s if elem.inner_html !~ REGEXES[:divToPElementsRe] debug("Altering div(##{elem[:id]}.#{elem[:class]}) to p"); elem.name = "p" end else # wrap text nodes in p tags # elem.children.each do |child| # if child.text? # debug("wrapping text node with a p") # child.swap("<p>#{child.text}</p>") # end # end end end end |