def extract_one(self, item, response):
name = extract_text(item.css('.child_title').get(), guess_layout=False)
link = item.css('.click_for_more_container>a::attr(href)').get()
if link:
link = urljoin(response.url, link)
else:
link = ''
price = item.css('.centered .styleColor::text').get()
price = price.split('€')[-1].strip().replace(',', '')
uid = item.xpath(
'.//div[@class="item-box-desc"]/p/i[contains(text(), "Product Code")]/text()'
).get()
if uid:
uid = uid.split(':')[-1].strip()
ean = item.xpath(
'.//div[@class="item-box-desc"]/p/i[contains(text(), "Barcode/EAN")]/text()'
).get()
if ean:
ean = ean.split(':')[-1].strip()
return {
'ean': ean,
'link': link,
'name': name,
'price': float(price),
'spider': self.name,
'T': utc_time(),
'uid': uid,
}
def printContentOnConsole(text):
if not "html-text" in sys.modules:
import html_text
print(html_text.extract_text(text))
#subprocess.Popen(["echo", html_text.extract_text(text)])
#sys.stdout.flush()
return text
def post(masto, body, instance, title=None, direction='ltr', in_reply_to=None):
# Markdown more than we need, to [hopefully] discard chopped markup.
summary = extract_text(markdown(body.strip()))[:140]
hashtags = get_hashtags(body, ignore=summary)
mentions = get_mentions(body, ignore=summary)
irt_id = in_reply_to and in_reply_to.get('id') or None
body = linkify_hashtags(linkify_mentions(body), instance)
if direction == 'rtl':
body = u"""<div dir="rtl">
{}
</div>""".format(markdown(body))
if in_reply_to:
body = u"""#### In reply to [@{}]({}):
{}""".format(in_reply_to['account']['username'], in_reply_to['url'], body)
gist = make_gist(
title
or u"A gistodon toot, {} GMT".format(time.asctime(time.gmtime())),
body + u"""
###### Generated by [Gistodon](https://github.com/thedod/gistodon/#readme).""")
if NO_TOOTING:
return gist
status = u'{}... {}'.format(summary, gist)
if hashtags or mentions:
status += u'\n' + u' '.join(hashtags.union(mentions))
return masto.status_post(status, spoiler_text=title,
in_reply_to_id=irt_id)['url']
def switchToCli():
if not "html-text" in sys.modules:
import html_text
# Print content where is context menu was called from
if config.pluginContext == "study":
print(html_text.extract_text(config.studyWindowContent))
else:
print(html_text.extract_text(config.bibleWindowContent))
# Hide gui
if platform.system() == "Darwin":
config.mainWindow.showMinimized()
else:
config.mainWindow.hide()
# Cli input
config.cli = True
toQuit = False
#config.printContentOnConsole = printContentOnConsole
config.bibleWindowContentTransformers.append(printContentOnConsole)
config.studyWindowContentTransformers.append(printContentOnConsole)
while config.cli:
print("--------------------")
print("Enter '.bible' to read bible text, '.study' to read study resource, '.help' to read UBA command reference, '.gui' to launch gui, '.quit' to quit,")
command = input("or UBA command: ").strip()
if command == ".gui":
del config.bibleWindowContentTransformers[-1]
del config.studyWindowContentTransformers[-1]
config.cli = False
elif command == ".bible":
print(html_text.extract_text(config.bibleWindowContent))
elif command == ".study":
print(html_text.extract_text(config.studyWindowContent))
elif command in (".help", ".command"):
getCommandDocumentation()
elif command == ".quit":
toQuit = True
config.cli = False
else:
config.mainWindow.runTextCommand(command)
if toQuit:
config.mainWindow.quitApp()
else:
app.setApplicationName("UniqueBible.app")
if platform.system() == "Darwin":
config.mainWindow.showMaximized()
config.mainWindow.raise_()
else:
config.mainWindow.show()
def get_body(self, response):
xpath = "//div[@id = 'article']/descendant::p[not(ancestor::div/@class='aboutAuthor')]"
body = response.xpath(xpath).getall()
body = [extract_text(para) for para in body]
# check if first line is real text
if len(body[0].split()) < 5:
body = body[1:]
return "\n\n".join(body)
def get_paragraphs_html_text(str_text, mode):
"""
using html_text
"""
try:
text = html_text.extract_text(str_text, guess_layout=False)
except TypeError:
return ['']
return re.split("\n", text)
def test_webpages():
webpages = sorted(glob.glob('./test_webpages/*.html'))
extracted = sorted(glob.glob('./test_webpages/*.txt'))
for page, extr in zip(webpages, extracted):
with open(page, 'r', encoding='utf8') as f_in:
html = f_in.read()
with open(extr, 'r', encoding='utf8') as f_in:
expected = f_in.read()
assert extract_text(html) == expected
async def security_check(page, response):
html = await page.content()
text = html_text.extract_text(html)
error_level = len([r for r in error_regs if r.search(text)])
if len(text) < 1000:
error_level += 1
if response and not response.ok:
error_level += 1
return error_level
def main():
output = {}
for path in Path('html').glob('*.html.gz'):
with gzip.open(path, 'rt', encoding='utf8') as f:
html = f.read()
item_id = path.stem.split('.')[0]
output[item_id] = {'articleBody': html_text.extract_text(html)}
(Path('output') / 'html-text.json').write_text(json.dumps(
output, sort_keys=True, ensure_ascii=False, indent=4),
encoding='utf8')
def _test(train_response):
model = decode_object(train_response.pop('model')) # type: BaseModel
pprint(train_response)
pprint(json.loads(train_response['quality']))
page_neg, page_pos = pages[:2]
pred_proba = lambda page: \
model.predict_proba([{'text': extract_text(page['html'])}])[0][1]
assert pred_proba(page_pos) > 0.5
assert pred_proba(page_neg) < 0.5
return train_response
def parse_dataprovider_publicmessagedetail(self, response):
# TODO
# print(self, response.text)
response_json = response.json()
msg = Message(
oid = response.meta['messageOID'],
date=response_json["time"], # TODO: Parse
title=html_text.extract_text(response_json["subject"]),
body=html_text.extract_text(response_json["body"]),
publication_id=response_json["tenderOID"],
data={
"authorityKey": response_json["authorityKey"],
"tender": response_json["tender"],
},
)
if "attachmentLink" in response_json:
msg['file_name']=response_json["fileName"]
msg['file_urls']=[response.urljoin(response_json["attachmentLink"])]
return msg
def text_worker(item, html_field):
url = item.get('url')
html = item.get(html_field)
try:
text = html_text.extract_text(html)
except UnicodeEncodeError:
return None
text_item = {'text': text}
if url is not None:
text_item['url'] = url
return text_item
def relevancy(self, response: Response) -> float:
if not isinstance(response, TextResponse):
# XXX: only text responses are supported
return 0.0
if self.classifier_input == 'vector':
x = self._page_vector(response)
elif self.classifier_input == 'text':
x = html_text.extract_text(response.text)
elif self.classifier_input == 'text_url':
x = {
'text': html_text.extract_text(response.text),
'url': response.url
}
elif self.classifier_input == 'html':
x = response.text
else:
raise ValueError("self.classifier_input is invalid")
return float(self.relevancy_clf.predict_proba([x])[0, 1])
def main():
output = {}
for path in Path('html').glob('*.html.gz'):
with gzip.open(path, 'rt', encoding='utf8') as f:
html = f.read()
item_id = path.stem.split('.')[0]
doc = Document(html)
text = html_text.extract_text(doc.summary(html_partial=True))
output[item_id] = {'articleBody': text}
(Path('output') / 'readability.json').write_text(json.dumps(
output, sort_keys=True, ensure_ascii=False, indent=4),
encoding='utf8')
def test_webpages(page, extracted):
html = _load_file(page)
if not six.PY3:
# FIXME: produces '\xa0' in Python 2, but ' ' in Python 3
# this difference is ignored in this test.
# What is the correct behavior?
html = html.replace(' ', ' ')
expected = _load_file(extracted)
assert extract_text(html) == expected
tree = cleaner.clean_html(parse_html(html))
assert etree_to_text(tree) == expected
def test_guess_layout():
html = (u'<title> title </title><div>text_1.<p>text_2 text_3</p>'
'<p id="demo"></p><ul><li>text_4</li><li>text_5</li></ul>'
'<p>text_6<em>text_7</em>text_8</p>text_9</div>'
'<script>document.getElementById("demo").innerHTML = '
'"This should be skipped";</script> <p>...text_10</p>')
text = 'title text_1. text_2 text_3 text_4 text_5 text_6 text_7 ' \
'text_8 text_9 ...text_10'
assert extract_text(html, guess_punct_space=False, guess_layout=False) == text
text = ('title\n\ntext_1.\n\ntext_2 text_3\n\ntext_4\ntext_5'
'\n\ntext_6 text_7 text_8\n\ntext_9\n\n...text_10')
assert extract_text(html, guess_punct_space=False, guess_layout=True) == text
text = 'title text_1. text_2 text_3 text_4 text_5 text_6 text_7 ' \
'text_8 text_9...text_10'
assert extract_text(html, guess_punct_space=True, guess_layout=False) == text
text = 'title\n\ntext_1.\n\ntext_2 text_3\n\ntext_4\ntext_5\n\n' \
'text_6 text_7 text_8\n\ntext_9\n\n...text_10'
assert extract_text(html, guess_punct_space=True, guess_layout=True) == text
def htmlToPlainText(content):
if isHtmlTextInstalled:
content = html_text.extract_text(content)
elif isBeautifulsoup4Installed:
content = re.sub("(</th>|</td>)", r"\1 ", content)
content = re.sub("(<br>|<br/>|</tr>)", r"\1\n", content)
content = re.sub("(</h[0-9]>|</p>|</div>|<hr>)", r"\1\n\n",
content)
content = BeautifulSoup(content, "html5lib").get_text()
else:
content = re.sub("<br/?>|<br>", "\n", content)
content = re.sub('<[^<]+?>', '', content)
return content
def get_text_html(href):
headers = {
'User-Agent':
'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36'
}
#req = requests.get(href, headers=headers)
req = reliable_request(href)
html = req.text
text = html_text.extract_text(html)
return text, html
def parse(self, response, date_check=True):
articles = response.css("div.m-headlineTease__info")
for article in articles:
url = article.xpath(".//a/@href").get()
if self.in_urls(url):
continue
dt = extract_text(article.xpath(".//time").get()).strip()
if "days" in dt and dt[0].isnumeric():
dt = self.today - timedelta(int(dt[0]))
if self.cutoff > dt and date_check:
break
yield scrapy.Request(url=url,
callback=self.art_parse,
headers=self.headers)
def extract_one(self, item):
name = extract_text(item.css('.desc-zone a[itemprop="url"]').get(), guess_layout=False)
link = item.css('.desc-zone h5 a::attr(href)').get()
mfr = item.css('.skus .sku[data-selenium="sku"]::text').get()
idata = json.loads(item.css('::attr(data-itemdata)').get())
return {
'T': utc_time(),
'link': link,
'mfr': mfr,
'name': name.strip(),
'price': float(idata['price']),
'spider': self.name,
'uid': idata['sku'],
}
async def fetch(spider, url, **spider_kwargs):
_, page = await spider.get(url, **spider_kwargs)
is_login = await is_login_page(page)
if is_login:
print("Detected login page.")
await login(page)
html = await page.content()
text = extract_text(html)
if "Let's do a quick security check" in text:
input("SOLVE SECURITY CHECK")
elif 'You’ve made too many requests in too short a time. Please try again later.' in text:
print("TOO MANY REQUESTS! sleeping")
await asyncio.sleep(60 * 11)
return await fetch(spider, url, **spider_kwargs)
return page
def startWithCli():
if not "html-text" in sys.modules:
import html_text
# Cli input
#config.mainWindow.hide()
#config.cli = True
#config.printContentOnConsole = printContentOnConsole
config.bibleWindowContentTransformers.append(printContentOnConsole)
config.studyWindowContentTransformers.append(printContentOnConsole)
while config.cli:
print("--------------------")
print(
"Enter '.bible' to read bible text, '.study' to read study resource, '.help' to read UBA command reference, '.gui' to launch gui, '.quit' to quit,"
)
command = input("or UBA command: ").strip()
if command == ".gui":
del config.bibleWindowContentTransformers[-1]
del config.studyWindowContentTransformers[-1]
config.cli = False
elif command == ".bible":
print(html_text.extract_text(config.bibleWindowContent))
# The following line does not work on Windows on UBA startup
#config.mainWindow.mainPage.runJavaScript("document.documentElement.outerHTML", 0, printContentOnConsole)
elif command == ".study":
print(html_text.extract_text(config.studyWindowContent))
# The following line does not work on Windows on UBA startup
#config.mainWindow.studyPage.runJavaScript("document.documentElement.outerHTML", 0, printContentOnConsole)
elif command in (".help", ".command"):
getCommandDocumentation()
elif command == ".quit":
exit()
else:
config.mainWindow.runTextCommand(command)
config.mainWindow.show()
if platform.system() == "Windows":
config.mainWindow.showMaximized()
def extract_features(doc):
html = doc['html'] or ''
if not doc_is_extra_sampled(doc):
try:
html = gzip.decompress(base64.b64decode(html)).decode('utf8')
except Exception:
pass # support not compressed html too
text = html_text.extract_text(html)
try:
lang = langdetect.detect(text)
except LangDetectException:
lang = None
return {
'text': text,
'language': lang,
}
def extract_text_from_pdf(pdf_path):
text = ""
resource_manager = PDFResourceManager()
fake_file_handle = io.BytesIO()
converter = HTMLConverter(resource_manager, fake_file_handle)
page_interpreter = PDFPageInterpreter(resource_manager, converter)
with open(pdf_path, 'rb') as fh:
for page in PDFPage.get_pages(fh, caching=True,
check_extractable=True):
page_interpreter.process_page(page)
text = fake_file_handle.getvalue().decode()
converter.close()
fake_file_handle.close()
text = html_text.extract_text(text)
file = open(pdf_path.replace('.pdf', '.txt'), 'w')
file.write(text)
file.close()
def parse(self, response, date_check=True):
articles = response.css(".archive-item-component__info")
for article in articles:
dt = extract_text(article.css("time").get())
dt = self.strptime(dt, "%B %d, %Y")
if dt and date_check:
if dt < self.cutoff:
break
url = urljoin(self.baseurl, article.xpath("a/@href").get())
if self.in_urls(url):
continue
yield scrapy.Request(
url=url,
callback=self.art_parse,
headers=self.headers,
cb_kwargs=dict(dt=dt),
)
def parse(self, response, date_check=True):
articles = response.css(".post_wrapper")
for art in articles:
dt = extract_text(
art.xpath(".//*[contains(@class, 'post_detail')]/a").get())
dt = self.strptime(dt, "%B %d, %Y")
if dt and date_check:
if dt < self.cutoff:
break
url = art.xpath(".//h3/a/@href").get()
if self.in_urls(url):
continue
yield scrapy.Request(
url=url,
headers=self.headers,
callback=self.art_parse,
cb_kwargs=dict(dt=dt),
)
def get_info(self, callback):
items = MovieScrapItem3()
name = callback.css('#firstHeading::text')[0].extract()
try:
image = 'http:' + callback.css(
'.vcard img::attr(src)')[0].extract()
except Exception as e:
image = None
born = callback.css('tr:nth-child(3) td span::text')[1].extract()
content = callback.css('.vcard+ p')
for para in content:
html = para.get()
bio = extract_text(html)
items['name'] = name
items['image'] = image
items['born'] = born
items['bio'] = bio
yield items
def parse(self, response):
cols = "rank name gain pips drawdown trades type monthly chart price age added action"
skip = [7, 8, 11, 12]
def age_to_months(t):
t = t.replace('m', 'M')
d = durations.Duration(t);
return d.to_months()
postprocess = {
'age': lambda t: age_to_months(t)
}
td = dict()
for i, col in enumerate(cols.split()):
td[i] = col
Behold().show('td')
for provider in response.xpath("//div[@class='row']//tr"):
data_row = dict()
Behold().show('provider')
details_url = None
for i, datum in enumerate(provider.xpath('td')):
Behold().show('i', 'datum')
if i == 1: # name
details_url = datum.css("a::attr(href)").get()
if i in skip:
print(".....skipping")
continue
text = html_text.extract_text(datum.get())
column_name = td[i]
if column_name in postprocess:
text = postprocess[column_name](text)
data_row[column_name] = text
if details_url:
yield scrapy.Request(url=details_url, callback=self.parse_details, meta={'data_row': data_row})
# next_page = response.css('.fa-angle-right').get()
# if next_page is not None:
# yield response.follow(next_page, self.parse)
async def _get_probability_blocked(self, response: Any) -> int:
"""Return an ingeger representing the probability of this worker encountered an blocking-related error.
i.e. server admin has blocked IP address for making too many requests.
Any value > 1 can be assumed to mean that the page is blocked.
"""
try:
self._text = html_text.extract_text(self._html)
error_level = len(
[r for r in self.block_error_regs if r.search(self._text)])
if len(self._text) < 1000:
error_level += 1
if response.status > 399:
error_level += 1
if error_level > 1:
self.logger.warning(
f"You may have been blocked! (Block probability level {error_level})")
return error_level
except Exception as e:
self.logger.error(f"Could not get block probability!: {e}")
def parse(self, response):
y = response.css(".article-thumb")
for el in y:
# meta is a string of how many days ago the article was published
# e.g. "1 week ago," "5 days ago" etc.
meta = extract_text(el.css(".article-thumb__meta").get())
# discard any article that is more than 4 days old
words = ["6", "week"]
if any([k in meta for k in words]):
break
url = el.css("a::attr(href)").get()
if self.in_urls(url):
continue
yield scrapy.Request(
url=url,
callback=self.art_parse,
headers=self.headers,
cb_kwargs=dict(dt=None),
)
