def UrlInHtmlMode(anUrl):
urlMode = lib_util.GetModeFromUrl(anUrl)
# sys.stderr.write("UrlInHtmlMode anUrl=%s urlMode=%s\n"%(anUrl,urlMode))
if urlMode:
return anUrl
else:
return lib_util.AnyUriModed(anUrl, "html")
def do_GET(self):
self.LogMsg("RdfQueue_HTTPRequestHandler do_GET path=" + self.path)
# Some browsers ask for this.
if self.path == "/favicon.ico":
return
if lib_tabular.ServeFile(self.path):
self.LogMsg("Tabular file: " + self.path)
return
# TODO: Use CGI arguments to return the size.
if self.path == "/infoqueues":
self.WriteQueueInformation()
return
# ET AUSSI: ON VA SERVIR LES FICHIERS CSV QU'ON CREE.
# ON LES RECONNAIT CAR YA "Tabular" AU DEBUT.
try:
# Maybe some ancillary tasks.
(entityType, entityId) = self.ParseQuery()
self.LogMsg("Getting queue entityId=" + entityId)
try:
theQ = self.server.m_queues[entityId]
theFeed = self.server.m_feeder[entityId]
except KeyError:
self.LogMsg("Caught when getting queue entityId=" + entityId)
lib_common.ErrorMessageHtml("No queue for entityId=" +
entityId)
# TODO: This test is already done in ParseQuery, so it could be simplified.
self.LogMsg("Getting pid")
pidFeeder = theFeed.pid
if not psutil.pid_exists(pidFeeder):
self.LogMsg("pidFeeder " + str(pidFeeder) + " not there.")
lib_common.ErrorMessageHtml("pidFeeder " + str(pidFeeder) +
" not there.")
# TODO: Restart the feeder if it crashed.
# TODO: Restart the feeder if it crashed.
# TODO: Restart the feeder if it crashed.
# OUI MAIS IL Y A DEJA EU UNE TENTATIVE AVEC ParseQuery()
self.send_response(200)
grph = cgiEnv.GetGraph()
self.LogMsg("do_GET Before insertion pidFeeder=%d nbtriples=%d" %
(pidFeeder, len(grph)))
# Add all the triples stored in the shared queue.
errMsg = None
while not theQ.empty():
self.LogMsg("Deserializing pidFeeder=%d szQ=%d" %
(pidFeeder, theQ.qsize()))
triple = theQ.get()
self.LogMsg("After get pidFeeder=%d szQ=%d" %
(pidFeeder, theQ.qsize()))
# If the object is NOT a triplet (In fact, a tuple) but a string,
# it can only be an error message, by convention.
if isinstance(triple, str):
errMsg = triple
self.LogMsg("Instead of tuple, str=" + errMsg)
break
try:
self.server.m_deserial(self.server.m_logFd, grph, triple)
except Exception:
exc = sys.exc_info()[1]
self.LogMsg("do_GET deserial:" + str(exc))
if errMsg != None:
self.LogMsg("do_GET Found error when reading queue=" + errMsg)
# TODO: Do we restart the feeder ? How to display this error message ?
else:
self.LogMsg("do_GET No error when reading queue")
self.LogMsg("After insertion pidFeeder=%d nbtriples=%d" %
(pidFeeder, len(grph)))
mode = lib_util.GetModeFromUrl(self.path)
self.LogMsg("GetModeFromUrl=" + mode)
topUrl = lib_util.TopUrl(entityType, entityId)
lib_util.SetDefaultOutput(self.wfile)
# TODO: How can we pass the edition parameters ? How to change them ?
# How to store them ? For the moment, the subservers cannot have any parameter.
# Maybe having parameters will imply restarting the feeder and maybe the subserver.
lib_common.OutCgiMode(grph,
topUrl,
mode,
self.server.m_page_title,
self.server.m_dot_layout,
errorMsg=errMsg,
isSubServer=True)
self.LogMsg("End of transmission pidFeeder=" + str(pidFeeder))
except Exception:
exc = sys.exc_info()[1]
# TODO: Send more appropriate error messages !!
self.send_error(
404, "go_GET caught:" + str(exc) + " pid=" + str(os.getpid()))
self.LogMsg("do_GET: Leaving")
def GetMimeTypeFromUrl(url):
urlMode = lib_util.GetModeFromUrl(url)
if urlMode and urlMode.startswith(mimeModePrefix):
return ModeToMimeType(urlMode)
else:
return ""
def ParseEntityUri(uriWithMode, longDisplay=True, force_entity_ip_addr=None):
#sys.stderr.write("ParseEntityUri uriWithMode=%s\n"%uriWithMode)
# Maybe there is a host name before the entity type. It can contain letters, numbers,
# hyphens, dots etc... but no ":" or "@".
# THIS CANNOT WORK WITH IPV6 ADDRESSES...
# WE MAY USE SCP SYNTAX: scp -6 osis@\[2001:db8:0:1\]:/home/osis/test.file ./test.file
# In the URI, we might have the CGI parameter "&mode=json". It must be removed otherwise
# it could be taken in entity_id, and the result of EntityToLabel() would be wrong.
uriWithModeClean = lib_util.UrlNoAmp(uriWithMode)
uri = lib_util.AnyUriModed(uriWithModeClean, "")
uriMode = lib_util.GetModeFromUrl(uriWithModeClean)
uprs = lib_util.survol_urlparse(uri)
filScript = os.path.basename(uprs.path)
# sys.stderr.write("ParseEntityUri filScript=%s\n"%filScript)
# This works for the scripts:
# entity.py xid=namespace/type:idGetNamespaceType
# objtypes_wbem.py Just extracts the namespace, as it prefixes the type: xid=namespace/type:id
# See variable lib_util.xidCgiDelimiter="?xid="
if uprs.query.startswith("xid="):
# TODO: Maybe the chain contains HTML codes and therefore cannot be parsed.
# Ex: "xid=%40%2F%3Aoracle_package." == "xid=@/:oracle_package."
(entity_type, entity_id,
entity_host) = lib_util.ParseXid(uprs.query[4:])
entity_graphic_class = entity_type
entity_label = CalcLabel(entity_host, entity_type, entity_id,
force_entity_ip_addr, filScript)
# TODO: Consider external_url_to_title, similar logic with different results.
if longDisplay:
entity_label = KnownScriptToTitle(filScript, uriMode, entity_host,
entity_label)
# Maybe an internal script, but not entity.py
# It has a special entity type as a display parameter
elif uri.startswith(lib_util.uriRoot):
# This is a bit of a special case which allows to display something if we know only
# the type of the entity but its id is undefined. Instead of displaying nothing,
# this attempts to display all available entities of this given type.
# source_top/enumerate_process.py etc... Not "." because this has a special role in Python.
mtch_enumerate = re.match(r"^.*/enumerate_([a-z0-9A-Z_]*)\.py$", uri)
if mtch_enumerate:
entity_graphic_class = mtch_enumerate.group(1)
entity_id = ""
# TODO: Change this label, not very nice.
# This indicates that a specific script can list all objects of a given entity type.
entity_label = entity_graphic_class + " enumeration"
else:
entity_graphic_class = "provider_script"
entity_id = ""
entity_label = KnownScriptToTitle(filScript, uriMode)
elif uri.split(':')[0] in ["ftp", "http", "https", "urn", "mail"]:
# Standard URLs. Example: lib_common.NodeUrl( "http://www.google.com" )
entity_graphic_class = ""
entity_id = ""
# Display the complete URL, otherwise it is not clickable.
entity_label = uriWithMode # uri # uri.split('/')[2]
else:
entity_graphic_class = ""
entity_id = "PLAINTEXTONLY"
entity_label = UriToTitle(uprs)
# TODO: " " are replaced by "%20". Why ? So change back.
entity_label = entity_label.replace("%20", " ")
return (entity_label, entity_graphic_class, entity_id)