def crawl(self, crawlCandidate):
article = Article()
parseCandidate = URLHelper.getCleanedUrl(crawlCandidate.url)
rawHtml = self.getHTML(crawlCandidate, parseCandidate)
if rawHtml is None:
return article
doc = self.getDocument(parseCandidate.url, rawHtml)
extractor = self.getExtractor()
docCleaner = self.getDocCleaner()
outputFormatter = self.getOutputFormatter()
# article
article.finalUrl = parseCandidate.url
article.linkhash = parseCandidate.linkhash
article.rawHtml = rawHtml
article.doc = doc
article.rawDoc = deepcopy(doc)
article.title = extractor.getTitle(article)
article.metaLang = extractor.getMetaLang(article)
article.metaFavicon = extractor.getMetaFavicon(article)
article.metaDescription = extractor.getMetaDescription(article)
article.metaKeywords = extractor.getMetaKeywords(article)
article.canonicalLink = extractor.getCanonicalLink(article)
article.domain = extractor.getDomain(article.finalUrl)
article.tags = extractor.extractTags(article)
# if the user requested a full body response
article.doc = docCleaner.clean(article)
# big stuff
article.topNode = extractor.calculateBestNodeBasedOnClustering(article)
if article.topNode is not None and any(
[self.config.enableImageFetching, self.config.enableBodyAnalysis]):
if self.config.enableImageFetching:
imageExtractor = self.getImageExtractor(article)
article.topImage = imageExtractor.getBestImage(
article.rawDoc, article.topNode)
if self.config.enableBodyAnalysis:
article.topNode = extractor.postExtractionCleanup(
article.topNode)
article.cleanedArticleText = outputFormatter.getFormattedText(
article.topNode)
return article
def crawl(self, crawlCandidate):
article = Article()
parseCandidate = URLHelper.getCleanedUrl(crawlCandidate.url)
rawHtml = self.getHTML(crawlCandidate, parseCandidate)
if rawHtml is None:
return article
doc = self.getDocument(parseCandidate.url, rawHtml)
extractor = self.getExtractor()
docCleaner = self.getDocCleaner()
outputFormatter = self.getOutputFormatter()
# article
article.finalUrl = parseCandidate.url
article.linkhash = parseCandidate.linkhash
article.rawHtml = rawHtml
article.doc = doc
article.rawDoc = deepcopy(doc)
article.title = extractor.getTitle(article)
# TODO
# article.publishDate = config.publishDateExtractor.extract(doc)
# article.additionalData = config.getAdditionalDataExtractor.extract(doc)
article.metaLang = extractor.getMetaLang(article)
article.metaFavicon = extractor.getMetaFavicon(article)
article.metaDescription = extractor.getMetaDescription(article)
article.metaKeywords = extractor.getMetaKeywords(article)
article.canonicalLink = extractor.getCanonicalLink(article)
article.domain = extractor.getDomain(article.finalUrl)
article.tags = extractor.extractTags(article)
# # before we do any calcs on the body itself let's clean up the document
article.doc = docCleaner.clean(article)
# big stuff
article.topNode = extractor.calculateBestNodeBasedOnClustering(article)
if article.topNode is not None:
# TODO
# movies and images
# article.movies = extractor.extractVideos(article.topNode)
if self.config.enableImageFetching:
imageExtractor = self.getImageExtractor(article)
article.topImage = imageExtractor.getBestImage(
article.rawDoc, article.topNode)
article.topNode = extractor.postExtractionCleanup(article.topNode)
article.cleanedArticleText = outputFormatter.getFormattedText(
article)
return article
def crawl(self, crawlCandidate):
article = Article()
parseCandidate = URLHelper.getCleanedUrl(crawlCandidate.url)
rawHtml = self.getHTML(crawlCandidate, parseCandidate)
if rawHtml is None:
return article
doc = self.getDocument(parseCandidate.url, rawHtml)
extractor = self.getExtractor()
docCleaner = self.getDocCleaner()
outputFormatter = self.getOutputFormatter()
# article
article.finalUrl = parseCandidate.url
article.linkhash = parseCandidate.linkhash
article.rawHtml = rawHtml
article.doc = doc
article.rawDoc = deepcopy(doc)
article.title = extractor.getTitle(article)
# TODO
# article.publishDate = config.publishDateExtractor.extract(doc)
# article.additionalData = config.getAdditionalDataExtractor.extract(doc)
article.metaLang = extractor.getMetaLang(article)
article.metaFavicon = extractor.getMetaFavicon(article)
article.metaDescription = extractor.getMetaDescription(article)
article.metaKeywords = extractor.getMetaKeywords(article)
article.canonicalLink = extractor.getCanonicalLink(article)
article.domain = extractor.getDomain(article.finalUrl)
article.tags = extractor.extractTags(article)
# # before we do any calcs on the body itself let's clean up the document
article.doc = docCleaner.clean(article)
# big stuff
article.topNode = extractor.calculateBestNodeBasedOnClustering(article)
if article.topNode is not None:
# TODO
# movies and images
# article.movies = extractor.extractVideos(article.topNode)
if self.config.enableImageFetching:
imageExtractor = self.getImageExtractor(article)
article.topImage = imageExtractor.getBestImage(article.rawDoc, article.topNode)
article.topNode = extractor.postExtractionCleanup(article.topNode)
article.cleanedArticleText = outputFormatter.getFormattedText(article.topNode)
return article
def crawl(self, crawlCandidate):
article = Article()
parseCandidate = URLHelper.getCleanedUrl(crawlCandidate.url)
rawHtml = self.getHTML(crawlCandidate, parseCandidate)
if rawHtml is None:
return article
doc = self.getDocument(parseCandidate.url, rawHtml)
if doc is None: return article
extractor = self.getExtractor()
docCleaner = self.getDocCleaner()
outputFormatter = self.getOutputFormatter()
# article
article.finalUrl = parseCandidate.url
article.linkhash = parseCandidate.linkhash
article.rawHtml = rawHtml
article.doc = doc
article.rawDoc = deepcopy(doc)
article.title = extractor.getTitle(article)
# TODO
# article.publishDate = config.publishDateExtractor.extract(doc)
# article.additionalData = config.getAdditionalDataExtractor.extract(doc)
article.metaLang = extractor.getMetaLang(article)
if not isinstance(self.config.targetLanguage,list): self.config.targetLanguage = [self.config.targetLanguage]
if article.metaLang:
self.config.targetLanguage.extend(article.metaLang)
extractor.setLanguage(self.config.targetLanguage)
article.metaFavicon = extractor.getMetaFavicon(article)
article.metaDescription = extractor.getMetaDescription(article)
article.metaKeywords = extractor.getMetaKeywords(article)
article.canonicalLink = extractor.getCanonicalLink(article)
article.domain = extractor.getDomain(article.finalUrl)
article.tags = extractor.extractTags(article)
# before we do any calcs on the body itself let's clean up the document
article.doc = docCleaner.clean(article)
# detects languages by unicode range
langs = get_languages(Parser.getText(article.doc))
self.config.targetLanguage.extend(langs)
extractor.setLanguage(self.config.targetLanguage)
# big stuff
article.h1 = ''
article.topNode = extractor.calculateBestNodeBasedOnClustering(article)
if article.topNode is not None:
# TODO
# movies and images
# article.movies = extractor.extractVideos(article.topNode)
if self.config.enableImageFetching:
imageExtractor = self.getImageExtractor(article)
article.topImage = imageExtractor.getBestImage(article.rawDoc, article.topNode)
article.topNode.attrib['rel'] = 'topnode' # mark html element
article.h1 = extractor.getH1(article)
article.topNode = extractor.postExtractionCleanup(article.topNode)
Parser.removeTitle(article.topNode,article.title,article.h1)
article.cleanedArticleText = outputFormatter.getFormattedText(article)
# cleanup tmp file
self.releaseResources(article)
return article
def test_instance(self):
a = Article()
self.assertIsInstance(a, Article)
def crawl(self, crawlCandidate):
article = Article()
parseCandidate = URLHelper.getCleanedUrl(crawlCandidate.url)
rawHtml = self.getHTML(crawlCandidate, parseCandidate)
if rawHtml is None:
return article
doc = self.getDocument(parseCandidate.url, rawHtml)
extractor = self.getExtractor()
docCleaner = self.getDocCleaner()
outputFormatter = self.getOutputFormatter()
# article
article.finalUrl = parseCandidate.url
article.linkhash = parseCandidate.linkhash
article.rawHtml = rawHtml
article.doc = doc
article.rawDoc = deepcopy(doc)
article.title = extractor.getTitle(article)
article.metaLang = extractor.getMetaLang(article)
article.metaFavicon = extractor.getMetaFavicon(article)
article.metaDescription = extractor.getMetaDescription(article)
article.metaKeywords = extractor.getMetaKeywords(article)
article.canonicalLink = extractor.getCanonicalLink(article)
article.domain = extractor.getDomain(article.finalUrl)
article.tags = extractor.extractTags(article)
# if the user requested a full body response
article.doc = docCleaner.clean(article)
# big stuff
article.topNode = extractor.calculateBestNodeBasedOnClustering(article)
if article.topNode is not None and any([self.config.enableImageFetching, self.config.enableBodyAnalysis]):
if self.config.enableImageFetching:
imageExtractor = self.getImageExtractor(article)
article.topImage = imageExtractor.getBestImage(article.rawDoc, article.topNode)
if self.config.enableBodyAnalysis:
article.topNode = extractor.postExtractionCleanup(article.topNode)
article.cleanedArticleText = outputFormatter.getFormattedText(article.topNode)
return article
