From 3347f16d939d397c23793490c7c6e91538801f53 Mon Sep 17 00:00:00 2001 From: Richard Harding Date: Fri, 20 Apr 2012 06:33:42 -0400 Subject: [PATCH] Fix the flipped nature of the wrapping setting --- src/readability_lxml/readability.py | 12 ++++++------ src/tests/test_article_only.py | 2 +- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/src/readability_lxml/readability.py b/src/readability_lxml/readability.py index 4bcf8df..255b626 100755 --- a/src/readability_lxml/readability.py +++ b/src/readability_lxml/readability.py @@ -139,7 +139,7 @@ class Document: def short_title(self): return shorten_title(self.html) - def summary(self, enclose_with_html_tag=False): + def summary(self, enclose_with_html_tag=True): """Generate the summary of the html docuemnt :param enclose_with_html_tag: return only the div of the document, @@ -197,7 +197,7 @@ class Document: raise Unparseable(str(e)), None, sys.exc_info()[2] def get_article(self, candidates, best_candidate, - enclose_with_html_tag=False): + enclose_with_html_tag=True): # 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. @@ -206,9 +206,9 @@ class Document: best_candidate['content_score'] * 0.2]) # create a new html document with a html->body->div if enclose_with_html_tag: - output = fragment_fromstring('
') - else: output = document_fromstring('
') + else: + output = fragment_fromstring('
') best_elem = best_candidate['elem'] for sibling in best_elem.getparent().getchildren(): # in lxml there no concept of simple text @@ -238,9 +238,9 @@ class Document: # We don't want to append directly to output, but the div # in html->body->div if enclose_with_html_tag: - output.append(sibling) - else: output.getchildren()[0].getchildren()[0].append(sibling) + else: + output.append(sibling) #if output is not None: # output.append(best_elem) return output diff --git a/src/tests/test_article_only.py b/src/tests/test_article_only.py index d8595c9..858d9c6 100644 --- a/src/tests/test_article_only.py +++ b/src/tests/test_article_only.py @@ -34,5 +34,5 @@ class TestArticleOnly(unittest.TestCase): """Using the si sample, make sure we can get the article alone.""" sample = load_sample('si-game.sample.html') doc = Document(sample, url='http://sportsillustrated.cnn.com/baseball/mlb/gameflash/2012/04/16/40630_preview.html') - res = doc.summary(enclose_with_html_tag=True) + res = doc.summary(enclose_with_html_tag=False) self.assertEqual('