diff options
Diffstat (limited to 'indra/lib')
| -rw-r--r-- | indra/lib/python/indra/ipc/webdav.py | 577 |
1 files changed, 577 insertions, 0 deletions
diff --git a/indra/lib/python/indra/ipc/webdav.py b/indra/lib/python/indra/ipc/webdav.py new file mode 100644 index 0000000000..d50b3f2b26 --- /dev/null +++ b/indra/lib/python/indra/ipc/webdav.py @@ -0,0 +1,577 @@ +""" +@file webdav.py +@brief Classes to make manipulation of a webdav store easier. + +Copyright (c) 2006-2007, Linden Research, Inc. +$License$ +""" + +import sys, os, httplib, urlparse +import socket, time +import xml.dom.minidom +import syslog +# import signal + +__revision__ = '0' + +dav_debug = False + + +# def urlsafe_b64decode (enc): +# return base64.decodestring (enc.replace ('_', '/').replace ('-', '+')) + +# def urlsafe_b64encode (str): +# return base64.encodestring (str).replace ('+', '-').replace ('/', '_') + + +class DAVError (Exception): + """ Base class for exceptions in this module. """ + def __init__ (self, status=0, message='', body='', details=''): + self.status = status + self.message = message + self.body = body + self.details = details + Exception.__init__ (self, '%d:%s:%s%s' % (self.status, self.message, + self.body, self.details)) + + def print_to_stderr (self): + """ print_to_stderr docstring """ + print >> sys.stderr, str (self.status) + ' ' + self.message + print >> sys.stderr, str (self.details) + + +class Timeout (Exception): + """ Timeout docstring """ + def __init__ (self, arg=''): + Exception.__init__ (self, arg) + + +def alarm_handler (signum, frame): + """ alarm_handler docstring """ + raise Timeout ('caught alarm') + + +class WebDAV: + """ WebDAV docstring """ + def __init__ (self, url, proxy=None, retries_before_fail=6): + self.init_url = url + self.init_proxy = proxy + self.retries_before_fail = retries_before_fail + url_parsed = urlparse.urlsplit (url) + + self.top_path = url_parsed[ 2 ] + # make sure top_path has a trailing / + if self.top_path == None or self.top_path == '': + self.top_path = '/' + elif len (self.top_path) > 1 and self.top_path[-1:] != '/': + self.top_path += '/' + + if dav_debug: + syslog.syslog ('new WebDAV %s : %s' % (str (url), str (proxy))) + + if proxy: + proxy_parsed = urlparse.urlsplit (proxy) + self.host_header = url_parsed[ 1 ] + host_and_port = proxy_parsed[ 1 ].split (':') + self.host = host_and_port[ 0 ] + if len (host_and_port) > 1: + self.port = int(host_and_port[ 1 ]) + else: + self.port = 80 + else: # no proxy + host_and_port = url_parsed[ 1 ].split (':') + self.host_header = None + self.host = host_and_port[ 0 ] + if len (host_and_port) > 1: + self.port = int(host_and_port[ 1 ]) + else: + self.port = 80 + + self.connection = False + self.connect () + + + def log (self, msg, depth=0): + """ log docstring """ + if dav_debug and depth == 0: + host = str (self.init_url) + if host == 'http://int.tuco.lindenlab.com:80/asset/': + host = 'tuco' + if host == 'http://harriet.lindenlab.com/asset-keep/': + host = 'harriet/asset-keep' + if host == 'http://harriet.lindenlab.com/asset-flag/': + host = 'harriet/asset-flag' + if host == 'http://harriet.lindenlab.com/asset/': + host = 'harriet/asset' + if host == 'http://ozzy.lindenlab.com/asset/': + host = 'ozzy/asset' + if host == 'http://station11.lindenlab.com:12041/:': + host = 'station11:12041' + proxy = str (self.init_proxy) + if proxy == 'None': + proxy = '' + if proxy == 'http://int.tuco.lindenlab.com:3128/': + proxy = 'tuco' + syslog.syslog ('WebDAV (%s:%s) %s' % (host, proxy, str (msg))) + + + def connect (self): + """ connect docstring """ + self.log ('connect') + self.connection = httplib.HTTPConnection (self.host, self.port) + + def __err (self, response, details): + """ __err docstring """ + raise DAVError (response.status, response.reason, response.read (), + str (self.init_url) + ':' + \ + str (self.init_proxy) + ':' + str (details)) + + def request (self, method, path, body=None, headers=None, + read_all=True, body_hook = None, recurse=0, allow_cache=True): + """ request docstring """ + # self.log ('request %s %s' % (method, path)) + if headers == None: + headers = {} + if not allow_cache: + headers['Pragma'] = 'no-cache' + headers['cache-control'] = 'no-cache' + try: + if method.lower () != 'purge': + if path.startswith ('/'): + path = path[1:] + if self.host_header: # use proxy + headers[ 'host' ] = self.host_header + fullpath = 'http://%s%s%s' % (self.host_header, + self.top_path, path) + else: # no proxy + fullpath = self.top_path + path + else: + fullpath = path + + self.connection.request (method, fullpath, body, headers) + if body_hook: + body_hook () + + # signal.signal (signal.SIGALRM, alarm_handler) + # try: + # signal.alarm (120) + # signal.alarm (0) + # except Timeout, e: + # if recurse < 6: + # return self.retry_request (method, path, body, headers, + # read_all, body_hook, recurse) + # else: + # raise DAVError (0, 'timeout', self.host, + # (method, path, body, headers, recurse)) + + response = self.connection.getresponse () + + if read_all: + while len (response.read (1024)) > 0: + pass + if (response.status == 500 or \ + response.status == 503 or \ + response.status == 403) and \ + recurse < self.retries_before_fail: + return self.retry_request (method, path, body, headers, + read_all, body_hook, recurse) + return response + except (httplib.ResponseNotReady, + httplib.BadStatusLine, + socket.error): + # if the server hangs up on us (keepalive off, broken pipe), + # we need to reconnect and try again. + if recurse < self.retries_before_fail: + return self.retry_request (method, path, body, headers, + read_all, body_hook, recurse) + raise DAVError (0, 'reconnect failed', self.host, + (method, path, body, headers, recurse)) + + + def retry_request (self, method, path, body, headers, + read_all, body_hook, recurse): + """ retry_request docstring """ + time.sleep (10.0 * recurse) + self.connect () + return self.request (method, path, body, headers, + read_all, body_hook, recurse+1) + + + + def propfind (self, path, body=None, depth=1): + """ propfind docstring """ + # self.log ('propfind %s' % path) + headers = {'Content-Type':'text/xml; charset="utf-8"', + 'Depth':str(depth)} + response = self.request ('PROPFIND', path, body, headers, False) + if response.status == 207: + return response # Multi-Status + self.__err (response, ('PROPFIND', path, body, headers, 0)) + + + def purge (self, path): + """ issue a squid purge command """ + headers = {'Accept':'*/*'} + response = self.request ('PURGE', path, None, headers) + if response.status == 200 or response.status == 404: + # 200 if it was purge, 404 if it wasn't there. + return response + self.__err (response, ('PURGE', path, None, headers)) + + + def get_file_size (self, path): + """ + Use propfind to ask a webdav server what the size of + a file is. If used on a directory (collection) return 0 + """ + self.log ('get_file_size %s' % path) + # "getcontentlength" property + # 8.1.1 Example - Retrieving Named Properties + # http://docs.python.org/lib/module-xml.dom.html + nsurl = 'http://apache.org/dav/props/' + doc = xml.dom.minidom.Document () + propfind_element = doc.createElementNS (nsurl, "D:propfind") + propfind_element.setAttributeNS (nsurl, 'xmlns:D', 'DAV:') + doc.appendChild (propfind_element) + prop_element = doc.createElementNS (nsurl, "D:prop") + propfind_element.appendChild (prop_element) + con_len_element = doc.createElementNS (nsurl, "D:getcontentlength") + prop_element.appendChild (con_len_element) + + response = self.propfind (path, doc.toxml ()) + doc.unlink () + + resp_doc = xml.dom.minidom.parseString (response.read ()) + cln = resp_doc.getElementsByTagNameNS ('DAV:','getcontentlength')[ 0 ] + try: + content_length = int (cln.childNodes[ 0 ].nodeValue) + except IndexError: + return 0 + resp_doc.unlink () + return content_length + + + def file_exists (self, path): + """ + do an http head on the given file. return True if it succeeds + """ + self.log ('file_exists %s' % path) + expect_gzip = path.endswith ('.gz') + response = self.request ('HEAD', path) + got_gzip = response.getheader ('Content-Encoding', '').strip () + if got_gzip.lower () == 'x-gzip' and expect_gzip == False: + # the asset server fakes us out if we ask for the non-gzipped + # version of an asset, but the server has the gzipped version. + return False + return response.status == 200 + + + def mkdir (self, path): + """ mkdir docstring """ + self.log ('mkdir %s' % path) + headers = {} + response = self.request ('MKCOL', path, None, headers) + if response.status == 201: + return # success + if response.status == 405: + return # directory already existed? + self.__err (response, ('MKCOL', path, None, headers, 0)) + + + def delete (self, path): + """ delete docstring """ + self.log ('delete %s' % path) + headers = {'Depth':'infinity'} # collections require infinity + response = self.request ('DELETE', path, None, headers) + if response.status == 204: + return # no content + if response.status == 404: + return # hmm + self.__err (response, ('DELETE', path, None, headers, 0)) + + + def list_directory (self, path, dir_filter=None, allow_cache=True, + minimum_cache_time=False): + """ + Request an http directory listing and parse the filenames out of lines + like: '<LI><A HREF="X"> X</A>'. If a filter function is provided, + only return filenames that the filter returns True for. + + This is sort of grody, but it seems faster than other ways of getting + this information from an isilon. + """ + self.log ('list_directory %s' % path) + + def try_match (lline, before, after): + """ try_match docstring """ + try: + blen = len (before) + asset_start_index = lline.index (before) + asset_end_index = lline.index (after, asset_start_index + blen) + asset = line[ asset_start_index + blen : asset_end_index ] + + if not dir_filter or dir_filter (asset): + return [ asset ] + return [] + except ValueError: + return [] + + if len (path) > 0 and path[-1:] != '/': + path += '/' + + response = self.request ('GET', path, None, {}, False, + allow_cache=allow_cache) + + if allow_cache and minimum_cache_time: # XXX + print response.getheader ('Date') + # s = "2005-12-06T12:13:14" + # from datetime import datetime + # from time import strptime + # datetime(*strptime(s, "%Y-%m-%dT%H:%M:%S")[0:6]) + # datetime.datetime(2005, 12, 6, 12, 13, 14) + + if response.status != 200: + self.__err (response, ('GET', path, None, {}, 0)) + assets = [] + for line in response.read ().split ('\n'): + lline = line.lower () + if lline.find ("parent directory") == -1: + # isilon file + assets += try_match (lline, '<li><a href="', '"> ') + # apache dir + assets += try_match (lline, 'alt="[dir]"> <a href="', '/">') + # apache file + assets += try_match (lline, 'alt="[ ]"> <a href="', '">') + return assets + + + def __tmp_filename (self, path_and_file): + """ __tmp_filename docstring """ + head, tail = os.path.split (path_and_file) + if head != '': + return head + '/.' + tail + '.' + str (os.getpid ()) + else: + return head + '.' + tail + '.' + str (os.getpid ()) + + + def __put__ (self, filesize, body_hook, remotefile): + """ __put__ docstring """ + headers = {'Content-Length' : str (filesize)} + remotefile_tmp = self.__tmp_filename (remotefile) + response = self.request ('PUT', remotefile_tmp, None, + headers, True, body_hook) + if not response.status in (201, 204): # created, no content + self.__err (response, ('PUT', remotefile, None, headers, 0)) + if filesize != self.get_file_size (remotefile_tmp): + try: + self.delete (remotefile_tmp) + except: + pass + raise DAVError (0, 'tmp upload error', remotefile_tmp) + # move the file to its final location + try: + self.rename (remotefile_tmp, remotefile) + except DAVError, exc: + if exc.status == 403: # try to clean up the tmp file + try: + self.delete (remotefile_tmp) + except: + pass + raise + if filesize != self.get_file_size (remotefile): + raise DAVError (0, 'file upload error', str (remotefile_tmp)) + + + def put_string (self, strng, remotefile): + """ put_string docstring """ + self.log ('put_string %d -> %s' % (len (strng), remotefile)) + filesize = len (strng) + def body_hook (): + """ body_hook docstring """ + self.connection.send (strng) + self.__put__ (filesize, body_hook, remotefile) + + + def put_file (self, localfile, remotefile): + """ + Send a local file to a remote webdav store. First, upload to + a temporary filename. Next make sure the file is the size we + expected. Next, move the file to its final location. Next, + check the file size at the final location. + """ + self.log ('put_file %s -> %s' % (localfile, remotefile)) + filesize = os.path.getsize (localfile) + def body_hook (): + """ body_hook docstring """ + handle = open (localfile) + while True: + data = handle.read (1300) + if len (data) == 0: + break + self.connection.send (data) + handle.close () + self.__put__ (filesize, body_hook, remotefile) + + + def create_empty_file (self, remotefile): + """ create an empty file """ + self.log ('touch_file %s' % (remotefile)) + headers = {'Content-Length' : '0'} + response = self.request ('PUT', remotefile, None, headers) + if not response.status in (201, 204): # created, no content + self.__err (response, ('PUT', remotefile, None, headers, 0)) + if self.get_file_size (remotefile) != 0: + raise DAVError (0, 'file upload error', str (remotefile)) + + + def __get_file_setup (self, remotefile, check_size=True): + """ __get_file_setup docstring """ + if check_size: + remotesize = self.get_file_size (remotefile) + response = self.request ('GET', remotefile, None, {}, False) + if response.status != 200: + self.__err (response, ('GET', remotefile, None, {}, 0)) + try: + content_length = int (response.getheader ("Content-Length")) + except TypeError: + content_length = None + if check_size: + if content_length != remotesize: + raise DAVError (0, 'file DL size error', remotefile) + return (response, content_length) + + + def __get_file_read (self, writehandle, response, content_length): + """ __get_file_read docstring """ + if content_length != None: + so_far_length = 0 + while so_far_length < content_length: + data = response.read (content_length - so_far_length) + if len (data) == 0: + raise DAVError (0, 'short file download') + so_far_length += len (data) + writehandle.write (data) + while len (response.read ()) > 0: + pass + else: + while True: + data = response.read () + if (len (data) < 1): + break + writehandle.write (data) + + + def get_file (self, remotefile, localfile, check_size=True): + """ + Get a remote file from a webdav server. Download to a local + tmp file, then move into place. Sanity check file sizes as + we go. + """ + self.log ('get_file %s -> %s' % (remotefile, localfile)) + (response, content_length) = \ + self.__get_file_setup (remotefile, check_size) + localfile_tmp = self.__tmp_filename (localfile) + handle = open (localfile_tmp, 'w') + self.__get_file_read (handle, response, content_length) + handle.close () + if check_size: + if content_length != os.path.getsize (localfile_tmp): + raise DAVError (0, 'file DL size error', + remotefile+','+localfile) + os.rename (localfile_tmp, localfile) + + + def get_file_as_string (self, remotefile, check_size=True): + """ + download a file from a webdav server and return it as a string. + """ + self.log ('get_file_as_string %s' % remotefile) + (response, content_length) = \ + self.__get_file_setup (remotefile, check_size) + # (tmp_handle, tmp_filename) = tempfile.mkstemp () + tmp_handle = os.tmpfile () + self.__get_file_read (tmp_handle, response, content_length) + tmp_handle.seek (0) + ret = tmp_handle.read () + tmp_handle.close () + # os.unlink (tmp_filename) + return ret + + + def get_post_as_string (self, remotefile, body): + """ + Do an http POST, send body, get response and return it. + """ + self.log ('get_post_as_string %s' % remotefile) + # headers = {'Content-Type':'application/x-www-form-urlencoded'} + headers = {'Content-Type':'text/xml; charset="utf-8"'} + # b64body = urlsafe_b64encode (asset_url) + response = self.request ('POST', remotefile, body, headers, False) + if response.status != 200: + self.__err (response, ('POST', remotefile, body, headers, 0)) + try: + content_length = int (response.getheader ('Content-Length')) + except TypeError: + content_length = None + tmp_handle = os.tmpfile () + self.__get_file_read (tmp_handle, response, content_length) + tmp_handle.seek (0) + ret = tmp_handle.read () + tmp_handle.close () + return ret + + + def __destination_command (self, verb, remotesrc, dstdav, remotedst): + """ + self and dstdav should point to the same http server. + """ + if len (remotedst) > 0 and remotedst[ 0 ] == '/': + remotedst = remotedst[1:] + headers = {'Destination': 'http://%s:%d%s%s' % (dstdav.host, + dstdav.port, + dstdav.top_path, + remotedst)} + response = self.request (verb, remotesrc, None, headers) + if response.status == 201: + return # created + if response.status == 204: + return # no content + self.__err (response, (verb, remotesrc, None, headers, 0)) + + + def rename (self, remotesrc, remotedst): + """ rename a file on a webdav server """ + self.log ('rename %s -> %s' % (remotesrc, remotedst)) + self.__destination_command ('MOVE', remotesrc, self, remotedst) + def xrename (self, remotesrc, dstdav, remotedst): + """ rename a file on a webdav server """ + self.log ('xrename %s -> %s' % (remotesrc, remotedst)) + self.__destination_command ('MOVE', remotesrc, dstdav, remotedst) + + + def copy (self, remotesrc, remotedst): + """ copy a file on a webdav server """ + self.log ('copy %s -> %s' % (remotesrc, remotedst)) + self.__destination_command ('COPY', remotesrc, self, remotedst) + def xcopy (self, remotesrc, dstdav, remotedst): + """ copy a file on a webdav server """ + self.log ('xcopy %s -> %s' % (remotesrc, remotedst)) + self.__destination_command ('COPY', remotesrc, dstdav, remotedst) + + +def put_string (data, url): + """ + upload string s to a url + """ + url_parsed = urlparse.urlsplit (url) + dav = WebDAV ('%s://%s/' % (url_parsed[ 0 ], url_parsed[ 1 ])) + dav.put_string (data, url_parsed[ 2 ]) + + +def get_string (url, check_size=True): + """ + return the contents of a url as a string + """ + url_parsed = urlparse.urlsplit (url) + dav = WebDAV ('%s://%s/' % (url_parsed[ 0 ], url_parsed[ 1 ])) + return dav.get_file_as_string (url_parsed[ 2 ], check_size) |