def story_top_image(stories_id):
story = mc.story(stories_id)
# use the tool key so anyone can see these images
story_html = apicache.story_raw_1st_download(TOOL_API_KEY, stories_id)
article = newspaper.Article(url=story['url'])
article.set_html(story_html)
article.parse()
return jsonify({
'top': article.top_image,
'all': list(article.images),
})
def scrap_article(article_link):
article = newspaper.Article(article_link)
article.download()
article.parse()
raw_paragraph = article.text
paragraph_list = [p for p in raw_paragraph.split("\n") if p]
return {
"title": article.title,
"paragraph_list": paragraph_list
}
def all_the_content(content, article_database_ref, reload_pundits=False):
"""
:param content: this is the mongo object containing our content up to now
:param reload_pundits: if true, pundits are re-scraped every time
:return: returns keywords, entities, and newpundits, as well as storing them in the mongo object for the article
"""
reload_pundits = True
article = newspaper.Article(content['url'])
article.download()
article.parse()
article.nlp()
print "HERE ARE THE NEWSPAPER KEYWORDS", article.keywords
content['keywords'] = ""
content['entities'] = ""
# if not 'keywords' in content:
# content['keywords'] = [x for x in get_keywords(content['text'])
# if x['count'] > 2]
# _content.update({'_id': bson.ObjectId(content['id'])},
# {'$set': {'keywords': content['keywords']}})
#
# if not 'entities' in content:
# content['entities'] = get_entities(content['text'])
# _content.update({'_id': bson.ObjectId(content['id'])},
# {'$set': {'entities': content['entities']}})
if not 'newpundits' in content or reload_pundits:
content['newpundits'] = []
dupe_list = []
snippets, ratios = pundits.keyword_match(article_database_ref,
article.keywords)
content['newpundits'] = snippets
_content.update({'_id': bson.ObjectId(content['id'])},
{'$set': {
'newpundits': content['newpundits']
}})
if not len(content['newpundits']):
print "nothing to see here!"
failed_snippet = {}
failed_snippet['name'] = "#shambles"
failed_snippet['text'] = "we can't seem to find anything."
content['newpundits'] = [[failed_snippet]]
else:
print "HERE ARE NEW PUNDITS:", content['newpundits']
return content['keywords'], content['entities'], content['newpundits']
def get_article(url):
article = newspaper.Article(url, language=u'zh')
try:
article.download()
article.parse()
except Exception as e:
print(u"Something go wrong...Cannot download it...")
pass
article_title = article.title
article_text = article.text
return article_title, article_text
def getKeywords(self):
# UGLY HACK WARNING
# if a site has a specific scraper written for it, Newspaper is never involved - but Newspaper's keyword functionality is really good and I don't want to write my own function for it
# so I'm creating a newspaper.Article object and forcibly setting attributes to allow the natural language processing to work and give me keywords
a = newspaper.Article(self.url)
a.text = self.text
a.title = self.title
a.download_state = 2 # nlp() function below uses direct comparisons to check for download state so I'm getting away with setting it to something arbitrary
a.is_parsed = True
a.nlp()
return a.keywords
def get_text_from_url(url,
session,
cleanwriter,
errorwriter,
allow_redirects=False,
verify=True):
url_idx = url[0]
url_str = url[1]
try:
response = session.get(url_str,
allow_redirects=allow_redirects,
verify=verify)
response.close()
except (ConnectionError, InvalidSchema) as e:
errorwriter.writerow([url_str, e.__class__.__name__])
response = None
print(("#%s:" % url_idx), e.__class__.__name__, url_str)
pass
except (MissingSchema, TooManyRedirects, RetryError) as e:
errorwriter.writerow([url_str, e.__class__.__name__])
response = None
print(("#%s:" % url_idx), e.__class__.__name__, url_str)
pass
if response is not None:
if response.ok:
article = newspaper.Article(url_str)
article.download()
# See https://github.com/codelucas/newspaper/blob/master/newspaper/article.py#L31
if article.download_state == 2:
article.parse()
article.nlp()
date, time = get_date_time(article.html)
cleanwriter.writerow([
article.text, article.title, article.keywords, url_str,
article.tags, article.meta_keywords, date, time
])
else:
errorwriter.writerow([url_str, response.status_code])
print("#%s: Error with status code %s for URL: %s" %
(url_idx, response.status_code, url_str))
else:
print("%s is not a valid URL" % url_str)
def get_article(self, url):
article = newspaper.Article(url, language=u'zh')
try:
article.download()
article.parse()
except Exception as e:
self.log.info(
u"Something go wrong:\n{}\nCannot download it...".format(e))
pass
article_title = article.title
article_text = article.text
return article_title, article_text
def save_raw_html_files(graph, page_id, num_posts, html_save_loc):
#Unrelated to project. Saving raw html files for TT's work
page_posts = graph.get_connections(page_id, "posts", limit=num_posts)
page_posts_data = page_posts['data']
article_index = 0
success = 0
driver = webdriver.Firefox()
driver.set_page_load_timeout(30)
for post in page_posts_data:
try:
print("Examining article {n}".format(n=article_index))
attachments = graph.get_connections(post['id'], "attachments")
link_in_post = attachments['data'][0]['url']
driver.get(link_in_post)
redirected_url = driver.current_url #Javascript redirect smh
if "facebook.com" in redirected_url:
print("Redirect failed - probably not an article")
continue
article = newspaper.Article(redirected_url, "en")
article.download()
html_file = article.html
#final_save_loc = html_save_loc + str(article_index) + ".html"
final_save_loc = html_save_loc + post['id'] + ".html"
f = open(final_save_loc, 'w')
f.write(html_file)
f.close()
article_index = article_index + 1
success = success + 1
except KeyError:
print("This article has no attachments or url on facebook")
continue
except TimeoutException:
print(
"Time out exception has been thrown. Just go to next article, since we don't care about any particular article."
)
continue
except UnicodeEncodeError:
print(
"UnicodeEncodeError thrown. Just go to the next one lol, too many possible reasons why"
)
continue
print("Number of successful html files downloaded: {n}".format(n=success))
driver.close()
return
def extract_headline(self):
try:
self.article = newspaper.Article(self.news_url)
self.article.download()
self.article.parse()
except newspaper.article.ArticleException: #List possible errors in case of any exception
print("\nCONNECTION/URL ERROR: There may be a problem with your connection or the URL entered may be invalid")
article.title = "Invalid URL/Could not extract title"
return self.article.title.strip()
def parse(self, response):
newsItem = newsPaperItem()
article = newspaper.Article(response.url)
article.download()
article.parse()
nltk.download('punkt')
article.nlp()
newsItem["Author_Name"] = article.authors
newsItem['Publication_Date'] = article.publish_date
newsItem['Keywords'] = article.keywords
newsItem['Article_text'] = article.text
yield newsItem
def link_pull(self, url):
parser = newspaper.Article(url, request_timeout=10)
try:
parser.download()
parser.parse()
except (newspaper.article.ArticleException, ValueError):
return (None, None, [])
article = parser.text
title = parser.title
img = parser.top_image
return (article, title, [img])
def findArticle(links):
articles = []
for l in links:
url = l.strip()
a = newspaper.Article(url, language='en')
a.download()
try:
a.parse()
except newspaper.article.ArticleException:
print("Article not found")
return articles, links
def scrape(self):
for source in [self.queue[i].get_urls() for i in range(len(self.queue))]:
if source:
for date in source:
for url in date:
entry = {}
article = nw.Article(url)
article.download()
article.parse()
src = re.search('www.(.*).co',url).group(0).replace('.','')
entry.update(title=article.title, source=src, text=article.text)
self.results = self.results.append(entry,ignore_index=True)
self.save()
def extract_article(url):
"""Function that takes the url string of a news article and returns the
title and text of the article as a Python dictionary. Built on top of
Newspaper's article scraping & curation library."""
link = newspaper.Article(url)
link.download()
link.parse()
article = {}
article["title"] = link.title
article["text"] = link.text
return (article)
def extract(self, url, html_text: str):
doc = newspaper.Article(url)
doc.download(input_html=html_text)
doc.parse()
self.content = {
'url': url,
'text': doc.text,
'title': doc.title,
'publish_date': doc.publish_date,
'top_image_url': doc.top_image,
'authors': doc.authors,
'extraction_method': METHOD_NEWSPAPER_3k,
}
def xrun(suburls):
sh = StructHtml(feilds, 'ex001.xlsx')
for v in suburls:
# print type(v)
url = getInnerPageURLs(v)[0]
url = "%s%s" % (mainHTTP, url)
art = newspaper.Article(url, language='zh')
art.download()
art.parse()
sh.getSruct(art.text, art.html)
yield (v, sh.format_txt)
pass
pass
def retrieveContent(self, link):
try:
a = newspaper.Article(link)
a.download()
a.parse()
text = a.text
a.nlp()
self.keywords = a.keywords
self.retrieved = str(datetime.datetime.utcnow())
return a.text
except Exception as e:
log.error("Exception retrieving %s" % (link))
log.exception(e)
def set_article(self):
try:
self.article = newspaper.Article(self.url, keep_article_html=True)
except Exception as e:
print e.message
self.article.download()
if not self.article.is_downloaded:
time.sleep(1)
self.article.parse()
if not self.article.is_parsed:
time.sleep(1)
self.article.nlp()
def getArticle(url):
import dateutil.parser
article = newspaper.Article(url, keep_article_html=True)
try:
article.download()
article.parse()
except Exception as e:
raise e
date = article.publish_date or dateutil.parser.parse(
extractDate(article.html)).strftime('%Y-%m-%d %H:%M:%S')
if not date: raise Exception("Cannot find date")
return (url, article.title, ','.join(article.authors), str(date),
article.text.replace('\n', ' '))
def getText(url):
try:
time.sleep(1 / 5)
a = newspaper.Article(url, language='zh', memoize_articles=False)
a.download()
a.parse()
text = ' ' + a.text
except Exception as e:
time.sleep(random.randint(1, 10) / 10)
print("下载失败:" + url)
print(e)
text = getText(url)
return text.replace('\n\n', '\n\n ')
def getText(self, url):
try:
time.sleep(3)
a = newspaper.Article(url, language='zh')
a.download()
a.parse()
text = ' ' + a.text
except Exception as e:
time.sleep(random.randint(1, 10))
print("下载失败:" + url)
print(e)
text = self.getText(url)
return text.replace('\n\n', '\n\n ')
def build_article_object(article_url):
'''Build a formatted string with the article title, summary, and url'''
log.debug("Building article object for article {0}".format(article_url))
article = newspaper.Article(article_url)
log.debug("Downloading article {0}".format(article_url))
article.download()
log.debug("Finished downloading article {0}, parsing".format(article_url))
article.parse()
log.debug("Finished debugging {0}, running nlp".format(article_url))
article.nlp()
article_str = "{0} ({1})\n{2}\n".format(
article.title.encode('ascii', 'ignore'), article_url, article.summary)
output_strs.append(article_str)
def testurl(url, newsClassifier):
a = newspaper.Article(url)
a.download()
a.parse()
a.nlp()
l1 = get_named_entities(a.text)
author = "default"
try:
author = a.author[0]
except:
print "Not found"
art = add_article(a.title, a.summary, url, author, l1)
test_keywords(art, newsClassifier)
def __init__(self, title=None, url=None, pubDate=None, rssFeed=None):
super(Article, self).__init__()
# self._title = title if title != None else ''
# self._url = url if url != None else ''
# self._pub_date = pubDate if pubDate != None else ''
self._title = title
self._url = url
self._pub_date = pubDate
self._rssFeed = rssFeed
if self._title != None and self._url != None and self._pub_date != None:
self._articleParse = newspaper.Article(self.url)
def get_string_data():
# 读取文本数据
# 获取文章
article = newspaper.Article('http://www.bjnews.com.cn/news/2019/07/02/598100.html', language='zh')
# 下载文章
article.download()
# 解析文章
article.parse()
# 对文章进行nlp处理
article.nlp()
# nlp处理后的文章拼接
string_data = "".join(article.keywords)
return string_data
def getArticle(self,url):
article = newspaper.Article(url)
for i in range(5):
article.download()
print("Retry:",i)
html = article.html
if html and len(html)>0:
break
article.parse()
article.nlp()
return article
def calculate_article_word_count(url):
config = newspaper.Config()
config.browser_user_agent = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_4) " \
"AppleWebKit/537.36 (KHTML, like Gecko) " \
"Chrome/64.0.3282.186 Safari/537.36"
article = newspaper.Article(url, config=config)
article.download()
article.parse()
if (len(article.text.split())) < 200:
raise ValidationError('Could not find article')
return len(article.text.split()) + len(article.title.split())
def newspaper_read(article_link, continue_var=0):
#continue variables
max_allowed = 7999
start_character = max_allowed * continue_var
end_character = start_character + max_allowed
text = ''
n_entity = newspaper.Article(article_link)
n_entity.download()
n_entity.parse()
text += (n_entity.text.encode('ascii',
errors='ignore').replace('\n', '').replace(
'\t', ''))[start_character:end_character]
return text
def main(argv=None):
urls = []
if len(argv) < 2:
raise FileNotFoundError(
'Please input a file with urls as the first argument')
with open(argv[1]) as f:
for line in f.readlines():
line = line.rstrip("\n")
# url will be the last "column" in line
# this leaves room for user-created tags
if line != '':
line = line.split()[-1]
urls.append(line)
articles = []
print("downloading")
for url in urls:
print(".", url)
article = newspaper.Article(url, language='en')
article.download()
article.parse()
articles.append([article.title, article.text])
print("done downloading")
data_matrix = get_data_matrix(articles)
# write_matrix(data_matrix)
# title_combinations = article_combinations(articles)
# euclidean = get_similarity(title_combinations, data_matrix, euclidean_similarity)
# cosine = get_similarity(title_combinations, data_matrix, cosine_similarity)
# jaccard = get_similarity(title_combinations, data_matrix, jaccard_similarity)
# similarity_matrix = get_similarity_matrix(title_combinations,
# euclidean,
# cosine,
# jaccard)
# write_matrix(similarity_matrix)
print("start euclidean")
euclidean_clusters = k_means(data_matrix, 5, euclidean_similarity)
print("euclidean", [i[1] for i in euclidean_clusters])
print("-------\nstart cosine")
cosine_clusters = k_means(data_matrix, 5, cosine_similarity)
print("cosine", [i[1] for i in cosine_clusters])
print("-------\nstart jaccard")
jaccard_clusters = k_means(data_matrix, 5, jaccard_similarity)
print("jaccard", [i[1] for i in jaccard_clusters])
euclidean_sse = sse(data_matrix, euclidean_clusters)
cosine_sse = sse(data_matrix, cosine_clusters)
jaccard_sse = sse(data_matrix, jaccard_clusters)
print(euclidean_sse)
print(cosine_sse)
print(jaccard_sse)
def parse_item(self, response):
sel = Selector(response)
try:
#正文
new = newspaper.Article(url=response.url, language='zh')
new.download()
new.parse()
content = re.sub('\s|\W', '', new.text)
print(content)
# 标题
if sel.xpath("//h1[@class='entry-title']/text()").extract_first():
title = sel.xpath(
"//h1[@class='entry-title']/text()").extract_first()
# print(title)
else:
pass
#时间
if sel.xpath("//a[1]/time[@class='entry-date']").extract_first():
time = sel.xpath("//a[1]/time[@class='entry-date']//text()"
).extract_first()
# print(time)
else:
pass
#img图片url
if sel.xpath(
"//img[@class='aligncenter size-full wp-image-1142']/@src"
).extract_first():
img_url = sel.xpath(
"//img[@class='aligncenter size-full wp-image-1142']/@src"
).extract()
for url in img_url:
print(url)
else:
pass
#来源
if sel.xpath("//div[@class='entry-content']/p/text()").extract():
#内容来源一起
content = sel.xpath(
"//div[@class='entry-content']/p/text()").extract()
if '来源' in content[-1]:
print(content[-1])
# print(content,response.url)
else:
pass
else:
pass
print(response.url)
except:
pass
