def _request_file(url, base_url):
retries = HttpDirectory.MAX_RETRIES
while retries > 0:
try:
curl = HttpDirectory._curl_handle()
raw_headers = BytesIO()
curl.setopt(pycurl.URL, url.encode("utf-8", errors="ignore"))
curl.setopt(pycurl.HEADERFUNCTION, raw_headers.write)
curl.perform()
stripped_url = url[len(base_url) - 1:]
headers = HttpDirectory._parse_dict_header(
raw_headers.getvalue().decode("utf-8", errors="ignore"))
raw_headers.close()
path, name = os.path.split(stripped_url)
date = headers.get("Last-Modified", "1970-01-01")
curl.close()
return File(path=unquote(path).strip("/"),
name=unquote(name),
size=int(headers.get("Content-Length", -1)),
mtime=int(parse_date(date).timestamp()),
is_dir=False)
except pycurl.error:
retries -= 1
logger.debug("TimeoutError - _request_file")
raise TimeoutError
def _process_listings(self, url: str, in_q: Queue, files_q: Queue):
directory = RemoteDirectoryFactory.get_directory(url)
timeout_retries = 20 # If any worker threads reaches 20 retries, the whole queue is emptied
while directory:
try:
path = in_q.get(timeout=2000)
except Empty:
logger.debug("in_q is Empty")
directory.close()
break
if path is None:
break
try:
path_id, listing = directory.list_dir(path)
if len(listing) > 0 and path_id not in self.crawled_paths:
self.crawled_paths.add(path_id)
for f in listing:
if f.is_dir:
in_q.put(urljoin(f.path, f.name))
else:
files_q.put(f)
logger.debug("LISTED " + urljoin(self.url, path))
except TooManyConnectionsError:
logger.debug(
"Too many connections, this thread will be killed and path resubmitted"
)
# Kill worker and resubmit listing task
directory.close()
in_q.put(path)
# TODO: If all workers are killed the queue will never get processed and
# TODO: the crawler will be stuck forever
break
except TimeoutError:
logger.error("Directory listing timed out, " +
str(timeout_retries) + " retries left")
if timeout_retries > 0:
timeout_retries -= 1
in_q.put(path)
else:
logger.error("Dropping website " + url)
self.status_code = "Timeout during website listing"
directory.close()
logger.debug("Emptying queue")
while True:
try:
in_q.get_nowait()
in_q.task_done()
except Empty:
break
logger.debug("Emptied queue")
break
finally:
in_q.task_done()
def _fetch_body(self, url: str):
retries = HttpDirectory.MAX_RETRIES
while retries > 0:
try:
content = BytesIO()
self.curl.setopt(pycurl.URL,
url.encode("utf-8", errors="ignore"))
self.curl.setopt(pycurl.WRITEDATA, content)
self.curl.perform()
return content.getvalue().decode("utf-8", errors="ignore")
except pycurl.error:
self.close()
retries -= 1
logger.debug("TimeoutError - _fetch_body")
raise TimeoutError
def stop_when_connected(self):
failed_attempts = 0
while failed_attempts < self.max_attempts:
try:
self._connect()
logger.debug("New FTP connection @ " + self.base_url)
return True
except ftputil.error.FTPError as e:
if e.errno == 530 or e.errno == 421:
break
failed_attempts += 1
print("Connection error; reconnecting..." + e.strerror + " " +
str(e.errno))
time.sleep(2)
return False
def error(self, message):
logger.debug("HTML Parser error: " + message)
def close(self):
if self.ftp:
self.ftp.close()
self.ftp = None
logger.debug("Closing FtpRemoteDirectory for " + self.base_url)
def reconnect(self):
if self.ftp:
self.ftp.close()
success = self.stop_when_connected()
logger.debug("Reconnecting to FTP server " + self.base_url +
(" (OK)" if success else " (ERR)"))
