def test_html_layout(self):
html_parser = self.get_html_parser()
reader = HTMLReader(html_parser)
elements = tuple(
reader.iter_elements(io.BytesIO(b'''
<html>
<head>
<title>hi</title>
</head>
<body>
<img>
</body>
</html>'''), encoding='ascii')
)
print(elements)
self.assertEqual('html', elements[0].tag)
self.assertEqual('head', elements[1].tag)
self.assertEqual('title', elements[2].tag)
self.assertEqual('title', elements[3].tag)
self.assertEqual('head', elements[4].tag)
self.assertEqual('body', elements[5].tag)
self.assertEqual('img', elements[6].tag)
if isinstance(html_parser, LxmlHTMLParser):
self.assertEqual('img', elements[7].tag)
self.assertEqual('body', elements[8].tag)
self.assertEqual('html', elements[9].tag)
else:
self.assertEqual('body', elements[7].tag)
self.assertEqual('html', elements[8].tag)
def test_html_encoding(self):
html_parser = self.get_html_parser()
is_lxml = isinstance(html_parser, LxmlHTMLParser)
reader = HTMLReader(html_parser)
bom_map = {
'utf_16_le': codecs.BOM_UTF16_LE,
'utf_16_be': codecs.BOM_UTF16_BE,
'utf_32_le': codecs.BOM_UTF32_LE,
'utf_32_be': codecs.BOM_UTF32_BE,
}
for name in CODEC_NAMES:
if name in EBCDIC or name == 'utf_8_sig':
# XXX: we're assuming that all codecs are ASCII backward
# compatable
continue
if is_lxml and (name.startswith('utf_16') or name.startswith('utf_32')):
# FIXME: libxml/lxml doesn't like it when we pass in a codec
# name but don't specify the endian but BOM is included
continue
print('->', name)
data = io.BytesIO( bom_map.get(name, b'') +'<img>'.encode(name))
elements = tuple(reader.iter_elements(data, encoding=name))
html_element = elements[0]
if isinstance(html_parser, LxmlHTMLParser):
self.assertEqual('html', html_element.tag)
else:
self.assertEqual('img', html_element.tag)
def test_html_encoding(self):
html_parser = HTMLParser()
reader = HTMLReader(html_parser)
bom_map = {
'utf_16_le': codecs.BOM_UTF16_LE,
'utf_16_be': codecs.BOM_UTF16_BE,
'utf_32_le': codecs.BOM_UTF32_LE,
'utf_32_be': codecs.BOM_UTF32_BE,
}
for name in CODEC_NAMES:
if name in EBCDIC or name == 'utf_8_sig':
# XXX: we're assuming that all codecs are ASCII backward
# compatable
continue
if name.startswith('utf_16') or name.startswith('utf_32'):
# FIXME: libxml/lxml doesn't like it when we pass in a codec
# name but don't specify the endian but BOM is included
continue
print('->', name)
data = io.BytesIO(bom_map.get(name, b'') + '<img>'.encode(name))
elements = tuple(reader.iter_elements(data, encoding=name))
html_element = elements[0]
self.assertEqual('html', html_element.tag)
def test_html_script_comment(self):
test_string = b'''<script><!-- blah --></script>'''
reader = HTMLReader(self.get_html_parser())
elements = reader.iter_elements(io.BytesIO(test_string),
encoding='ascii')
elements = tuple(elements)
self.assertTrue(
all(isinstance(element, Element) for element in elements))
def test_html_script_comment(self):
test_string = b'''<script><!-- blah --></script>'''
reader = HTMLReader(HTMLParser())
elements = reader.iter_elements(io.BytesIO(test_string),
encoding='ascii')
elements = tuple(elements)
self.assertTrue(
all(isinstance(element, Element) for element in elements))
def scrape_document(self, item_session):
response = item_session.response
url_info = item_session.request.url_info
url = url_info.raw
if response_body_size(response) < 30 * 1024 * 1024:
dupes_db = self.dupes_db
body = response.body.content()
if HTMLReader.is_response(response):
body = dupespotter.process_body(body, url)
digest = hashlib.md5(body).digest()
if dupes_db is not None:
dupe_of = dupes_db.get_old_url(digest)
else:
dupe_of = None
if dupe_of is not None:
# Don't extract links from pages we've already seen
# to avoid loops that descend a directory endlessly
print("DUPE {}\n OF {}".format(url, dupe_of))
return
else:
if dupes_db is not None:
dupes_db.set_old_url(digest, url)
super().scrape_document(item_session)
def process(self, url_item, request, response, file_writer_session):
'''Process PhantomJS.
Coroutine.
'''
if response.status_code != 200:
return
if not HTMLReader.is_supported(request=request, response=response):
return
_logger.debug('Starting PhantomJS processing.')
self._file_writer_session = file_writer_session
# FIXME: this is a quick hack for crashes. See #137.
attempts = int(os.environ.get('WPULL_PHANTOMJS_TRIES', 5))
for dummy in range(attempts):
try:
yield From(self._run_driver(url_item, request, response))
except trollius.TimeoutError:
_logger.warning(_('Waiting for page load timed out.'))
break
except PhantomJSCrashed as error:
_logger.exception(__('PhantomJS crashed: {}', error))
else:
break
else:
_logger.warning(
__(_('PhantomJS failed to fetch ‘{url}’. I am sorry.'),
url=request.url_info.url))
def process(self, item_session: ItemSession, request, response, file_writer_session):
'''Process PhantomJS.
Coroutine.
'''
if response.status_code != 200:
return
if not HTMLReader.is_supported(request=request, response=response):
return
_logger.debug('Starting PhantomJS processing.')
self._file_writer_session = file_writer_session
# FIXME: this is a quick hack for crashes. See #137.
attempts = int(os.environ.get('WPULL_PHANTOMJS_TRIES', 5))
for dummy in range(attempts):
try:
yield from self._run_driver(item_session, request, response)
except asyncio.TimeoutError:
_logger.warning(_('Waiting for page load timed out.'))
break
except PhantomJSCrashed as error:
_logger.exception(__('PhantomJS crashed: {}', error))
else:
break
else:
_logger.warning(__(
_('PhantomJS failed to fetch ‘{url}’. I am sorry.'),
url=request.url_info.url
))
def _append_filename_extension(self, response: BaseResponse):
'''Append an HTML/CSS file suffix as needed.'''
if not self._filename:
return
if response.request.url_info.scheme not in ('http', 'https'):
return
if not re.search(r'\.[hH][tT][mM][lL]?$', self._filename) and \
HTMLReader.is_response(response):
self._filename += '.html'
elif not re.search(r'\.[cC][sS][sS]$', self._filename) and \
CSSReader.is_response(response):
self._filename += '.css'
def test_html_encoding(self):
html_parser = self.get_html_parser()
reader = HTMLReader(html_parser)
for name in CODEC_NAMES:
if name in EBCDIC or name == 'utf_8_sig':
# XXX: we're assuming that all codecs are ASCII backward
# compatable
continue
if name.endswith('_le') or name.endswith('_be'):
# XXX: Assume BOM is always included
continue
print(name)
data = io.BytesIO('<img>'.encode(name))
elements = tuple(reader.iter_elements(data, encoding=name))
html_element = elements[0]
if isinstance(html_parser, LxmlHTMLParser):
self.assertEqual('html', html_element.tag)
else:
self.assertEqual('img', html_element.tag)
def _append_filename_extension(self, response):
'''Append an HTML/CSS file suffix as needed.'''
if not self._filename:
return
if response.request.url_info.scheme not in ('http', 'https'):
return
if not re.search(r'\.[hH][tT][mM][lL]?$', self._filename) and \
HTMLReader.is_response(response):
self._filename += '.html'
elif not re.search(r'\.[cC][sS][sS]$', self._filename) and \
CSSReader.is_response(response):
self._filename += '.css'
def process(self, url_item, request, response, file_writer_session):
if response.status_code != 200:
return
if not HTMLReader.is_supported(request=request, response=response):
return
session = Session(self._proxy_address, self._youtube_dl_path,
self._root_path, url_item, file_writer_session,
self._user_agent, self._warc_recorder,
self._inet_family, self._check_certificate)
url = url_item.url_info.url
_logger.info(__(_('youtube-dl fetching ‘{url}’.'), url=url))
with contextlib.closing(session):
yield From(session.run())
_logger.info(__(_('youtube-dl fetched ‘{url}’.'), url=url))
def process(self, item_session: ItemSession, request, response, file_writer_session):
if response.status_code != 200:
return
if not HTMLReader.is_supported(request=request, response=response):
return
session = Session(
self._proxy_address, self._youtube_dl_path, self._root_path,
item_session, file_writer_session, self._user_agent,
self._warc_recorder, self._inet_family, self._check_certificate
)
url = item_session.url_record.url
_logger.info(__(_('youtube-dl fetching ‘{url}’.'), url=url))
with contextlib.closing(session):
yield from session.run()
_logger.info(__(_('youtube-dl fetched ‘{url}’.'), url=url))
def scrape_document(self, request, response, url_item):
if response.body.size() < 30*1024*1024:
dupes_db = self.dupes_db
body = response.body.content()
if HTMLReader.is_response(response):
body = dupespotter.process_body(body, response.request.url)
digest = hashlib.md5(body).digest()
if dupes_db is not None:
dupe_of = dupes_db.get_old_url(digest)
else:
dupe_of = None
if dupe_of is not None:
# Don't extract links from pages we've already seen
# to avoid loops that descend a directory endlessly
print(" DUPE {}\n OF {}".format(response.request.url, dupe_of))
return
else:
if dupes_db is not None:
dupes_db.set_old_url(digest, response.request.url)
super().scrape_document(request, response, url_item)
def scrape_document(self, request, response, url_item):
if response.body.size() < 30*1024*1024:
dupes_db = self.dupes_db
body = response.body.content()
if HTMLReader.is_response(response):
body = archivebot.dupespotter.dupespotter.process_body(body, response.request.url)
digest = hashlib.md5(body).digest()
if dupes_db is not None:
dupe_of = dupes_db.get_old_url(digest)
else:
dupe_of = None
if dupe_of is not None:
# Don't extract links from pages we've already seen
# to avoid loops that descend a directory endlessly
print(" DUPE {}\n OF {}".format(response.request.url, dupe_of))
sys.stdout.flush()
return
else:
if dupes_db is not None:
dupes_db.set_old_url(digest, response.request.url)
super().scrape_document(request, response, url_item)
def test_html_early_html(self):
reader = HTMLReader(self.get_html_parser())
for test_string in [
b'''<!DOCTYPE HTML><html></html><img>''',
b'''<html></html><img>''',
b'''<!DOCTYPE HTML><img><html></html>''',
b'''<img><html></html>''',
b'''<!DOCTYPE HTML>
<html><body></body></html><p><img>''',
b'''
<html><body></body></html><p><img>''',
b'''
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<title>Download</title>
</head>
<body>
<br />
</body>
</html>
<pre><img></pre>
''',
b'''
<html>
<head>
<title>Download</title>
</head>
<body>
<br />
</body>
</html>
<pre><img></pre>
''',
b'''
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<body>
<br />
</body>
<head>
<title>Download</title>
</head>
</html>
<pre><img></pre>
''',
b'''
<html>
<body>
<br />
</body>
<head>
<title>Download</title>
</head>
</html>
<pre><img></pre>
''',
]:
elements = []
print()
print('a' * 10)
print(test_string)
for element in reader.iter_elements(io.BytesIO(test_string),
encoding='ascii'):
if isinstance(element, Element):
print(element)
elements.append(element)
element_tags = tuple(element.tag for element in elements)
self.assertIn('img', element_tags)
def test_html_detect(self):
self.assertTrue(
HTMLReader.is_file(
io.BytesIO('<html><body>hi</body></html>'.encode('utf-16le'))))
self.assertFalse(
HTMLReader.is_file(io.BytesIO('hello world!'.encode('utf-16le'))))
self.assertTrue(
HTMLReader.is_file(io.BytesIO(b'<title>hello</title>hi')))
self.assertTrue(HTMLReader.is_file(io.BytesIO(b'<html><body>hello')))
self.assertTrue(
HTMLReader.is_file(
io.BytesIO(
b'The document has moved <a href="somewhere.html">here</a>'
)))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.htm')))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.html')))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.dhtm')))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.xhtml')))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.xht')))
self.assertFalse(
HTMLReader.is_url(URLInfo.parse('example.com/image.jpg')))
self.assertTrue(
HTMLReader.is_request(Request('example.com/index.html')))
self.assertFalse(
HTMLReader.is_request(Request('example.com/image.jpg')))
response = Response(200, 'OK')
response.fields['Content-Type'] = 'text/html'
self.assertTrue(HTMLReader.is_response(response))
response = Response(200, 'OK')
response.fields['Content-Type'] = 'image/png'
self.assertFalse(HTMLReader.is_response(response))
def test_html_early_html(self):
reader = HTMLReader(HTMLParser())
for test_string in [
b'''<!DOCTYPE HTML><html></html><img>''',
b'''<html></html><img>''',
b'''<!DOCTYPE HTML><img><html></html>''',
b'''<img><html></html>''',
b'''<!DOCTYPE HTML>
<html><body></body></html><p><img>''',
b'''
<html><body></body></html><p><img>''',
b'''
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<title>Download</title>
</head>
<body>
<br />
</body>
</html>
<pre><img></pre>
''',
b'''
<html>
<head>
<title>Download</title>
</head>
<body>
<br />
</body>
</html>
<pre><img></pre>
''',
b'''
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<body>
<br />
</body>
<head>
<title>Download</title>
</head>
</html>
<pre><img></pre>
''',
b'''
<html>
<body>
<br />
</body>
<head>
<title>Download</title>
</head>
</html>
<pre><img></pre>
''',
]:
elements = []
print()
print('a' * 10)
print(test_string)
for element in reader.iter_elements(io.BytesIO(test_string),
encoding='ascii'):
if isinstance(element, Element):
print(element)
elements.append(element)
element_tags = tuple(element.tag for element in elements)
self.assertIn('img', element_tags)
def test_html_detect(self):
self.assertTrue(HTMLReader.is_file(
io.BytesIO('<html><body>hi</body></html>'.encode('utf-16le'))
))
self.assertFalse(HTMLReader.is_file(
io.BytesIO('hello world!'.encode('utf-16le'))
))
self.assertTrue(HTMLReader.is_file(
io.BytesIO(b'<title>hello</title>hi')
))
self.assertTrue(HTMLReader.is_file(
io.BytesIO(b'<html><body>hello')
))
self.assertTrue(HTMLReader.is_file(
io.BytesIO(
b'The document has moved <a href="somewhere.html">here</a>'
)
))
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.htm'))
)
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.html'))
)
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.dhtm'))
)
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.xhtml'))
)
self.assertTrue(
HTMLReader.is_url(URLInfo.parse('example.com/index.xht'))
)
self.assertFalse(
HTMLReader.is_url(URLInfo.parse('example.com/image.jpg'))
)
self.assertTrue(
HTMLReader.is_request(Request('example.com/index.html'))
)
self.assertFalse(
HTMLReader.is_request(Request('example.com/image.jpg'))
)
response = Response(200, 'OK')
response.fields['Content-Type'] = 'text/html'
self.assertTrue(HTMLReader.is_response(response))
response = Response(200, 'OK')
response.fields['Content-Type'] = 'image/png'
self.assertFalse(HTMLReader.is_response(response))
