def test_iso2seconds(self):
"""iso2seconds should be able to parse most iso 8601 duration strings"""
durations = (('P1Y', 3600*24*365),
('P1M', 2628000),
('P1W', 3600*24*7),
('P1D', 3600*24),
('PT10H20M30S', 10*3600+20*60+30),
('PT1H2M4.567S', 3600+2*60+5),
('PT1H2M4,567S', 3600+2*60+5),
)
for iso_in, seconds_out in durations:
self.assertEqual(iso2seconds(iso_in), seconds_out,
"Duration '%s' converts to %s when it should be %s"
% (iso_in, iso2seconds(iso_in), seconds_out))
def parse_car(self, xml_record):
'''
Main function for parsing CAR record.
Interesting data can be fetched from 2 places:
* as a content of node (here called text node)
* as a attribute value (extracted by getAttr)
'''
functions = {
'Site' : lambda nodes: self.getText(nodes['Site'][0].childNodes),
'SubmitHost' : lambda nodes: self.getText(nodes['SubmitHost'][0].childNodes),
'MachineName' : lambda nodes: self.getText(nodes['MachineName'][0].childNodes),
'Queue' : lambda nodes: self.getText(nodes['Queue'][0].childNodes),
'LocalJobId' : lambda nodes: self.getText(nodes['LocalJobId'][0].childNodes),
'LocalUserId' : lambda nodes: self.getText(nodes['LocalUserId'][0].childNodes),
'GlobalUserName' : lambda nodes: self.getText(nodes['GlobalUserName'][0].childNodes),
'FQAN' : lambda nodes: self.getText(
self.getTagByAttr(nodes['GroupAttribute'],
'type', 'FQAN')[0].childNodes),
'VO' : lambda nodes: self.getText(nodes['Group'][0].childNodes),
'VOGroup' : lambda nodes: self.getText(
self.getTagByAttr(nodes['GroupAttribute'],
'type', 'group')[0].childNodes),
'VORole' : lambda nodes: self.getText(
self.getTagByAttr(nodes['GroupAttribute'],
'type', 'role')[0].childNodes),
'WallDuration' : lambda nodes: iso2seconds(self.getText(
nodes['WallDuration'][0].childNodes)),
'CpuDuration' : lambda nodes: iso2seconds(self.retrieve_cpu(nodes)),
'Processors' : lambda nodes: self.getText(nodes['Processors'][0].childNodes),
'NodeCount' : lambda nodes: self.getText(nodes['NodeCount'][0].childNodes),
'MemoryReal' : lambda nodes: None,
'MemoryVirtual' : lambda nodes: None,
'StartTime' : lambda nodes: parse_timestamp(self.getText(
nodes['StartTime'][0].childNodes)),
'EndTime' : lambda nodes: parse_timestamp(self.getText(
nodes['EndTime'][0].childNodes)),
'InfrastructureDescription' : lambda nodes: self.getAttr(nodes['Infrastructure'][0], 'description'),
'InfrastructureType' : lambda nodes: self.getAttr(nodes['Infrastructure'][0], 'type'),
'ServiceLevelType' : lambda nodes: self.getAttr(
nodes['ServiceLevel'][0], 'type'),
'ServiceLevel' : lambda nodes: self.getText(
nodes['ServiceLevel'][0].childNodes),
}
tags = ['Site', 'SubmitHost', 'MachineName', 'Queue', 'LocalJobId', 'LocalUserId',
'GlobalUserName', 'GroupAttribute',
'Group', 'WallDuration', 'CpuDuration', 'Memory',
'Processors', 'NodeCount', 'StartTime', 'EndTime', 'Infrastructure',
'ServiceLevel']
# Create a dictionary of all the tags we want to retrieve from the XML
nodes = {}.fromkeys(tags)
data = {}
for node in nodes:
# Create a list of nodes which match the tags we want.
# Note that this only matches the one namespace we have defined.
nodes[node] = xml_record.getElementsByTagNameNS(self.NAMESPACE, node)
for field in functions:
try:
data[field] = functions[field](nodes)
except (IndexError, KeyError, AttributeError), e:
log.debug('Failed to parse field %s: %s' % (field, e))
def parseAurRecord(self, xml_record):
'''
Main function for parsing AUR record.
Interesting data can be fetched from 2 places:
* as a content of node (here called text node)
* as a attribute value (extracted by getAttr)
'''
functions = {
'Site' : lambda nodes: self.getText(nodes['Site'][0].childNodes),
'Month' : lambda nodes: self.getText(nodes['Month'][0].childNodes),
'Year' : lambda nodes: self.getText(nodes['Year'][0].childNodes),
'GlobalUserName' : lambda nodes: self.getText(nodes['GlobalUserName'][0].childNodes),
'VO' : lambda nodes: self.getText(nodes['Group'][0].childNodes),
'VOGroup' : lambda nodes: self.getText(
self.getTagByAttr(nodes['GroupAttribute'],
'type', 'vo-group', CarParser.NAMESPACE)[0].childNodes),
'VORole' : lambda nodes: self.getText(
self.getTagByAttr(nodes['GroupAttribute'],
'type', 'role', CarParser.NAMESPACE)[0].childNodes),
'MachineName' : lambda nodes: self.getText(nodes['MachineName'][0].childNodes),
'SubmitHost' : lambda nodes: self.getText(nodes['SubmitHost'][0].childNodes),
'Infrastructure' : lambda nodes: self.getAttr(nodes['Infrastructure'][0], 'type', CarParser.NAMESPACE),
'EarliestEndTime' : lambda nodes: parse_timestamp(self.getText(
nodes['EarliestEndTime'][0].childNodes)),
'LatestEndTime' : lambda nodes: parse_timestamp(self.getText(
nodes['LatestEndTime'][0].childNodes)),
'WallDuration' : lambda nodes: iso2seconds(self.getText(
nodes['WallDuration'][0].childNodes)),
'CpuDuration' : lambda nodes: iso2seconds(self.getText(
nodes['CpuDuration'][0].childNodes)),
'NormalisedWallDuration': lambda nodes: iso2seconds(self.getText(
nodes['NormalisedWallDuration'][0].childNodes)),
'NormalisedCpuDuration': lambda nodes: iso2seconds(self.getText(
nodes['NormalisedCpuDuration'][0].childNodes)),
'NumberOfJobs' : lambda nodes: self.getText(nodes['NumberOfJobs'][0].childNodes),
'NodeCount' : lambda nodes: self.getText(nodes['NodeCount'][0].childNodes),
'Processors' : lambda nodes: self.getText(nodes['Processors'][0].childNodes),
}
tags = ['Site', 'Month', 'Year', 'GlobalUserName', 'Group',
'GroupAttribute', 'SubmitHost', 'Infrastructure',
'EarliestEndTime', 'LatestEndTime', 'WallDuration', 'CpuDuration',
'NormalisedWallDuration', 'NormalisedCpuDuration',
'NumberOfJobs', 'NodeCount', 'Processors']
nodes = {}.fromkeys(tags)
data = {}
for node in nodes:
if node in ('GroupAttribute',):
# For these attributes we need to dig into the GroupAttribute
# elements to get the values so we save the whole elements.
nodes[node] = xml_record.getElementsByTagNameNS(
CarParser.NAMESPACE, 'GroupAttribute')
else:
nodes[node] = xml_record.getElementsByTagNameNS(self.NAMESPACE, node)
# Some of the nodes are in the CAR namespace.
nodes[node].extend(xml_record.getElementsByTagNameNS(CarParser.NAMESPACE, node))
for field in functions:
try:
data[field] = functions[field](nodes)
except IndexError, e:
log.debug('Failed to parse field %s: %s', field, e)
except KeyError, e:
log.debug('Failed to parse field %s: %s', field, e)
def test_iso2seconds(self):
txt1, txt2, txt3 = 'P1Y', 'P1M', 'P1D'
self.assertTrue(iso2seconds(txt1), 3600*24*365)
self.assertTrue(iso2seconds(txt2), 3600*24*30)
self.assertTrue(iso2seconds(txt3), 3600*24)
def test_iso2seconds(self):
txt1, txt2, txt3 = 'P1Y', 'P1M', 'P1D'
self.assertTrue(iso2seconds(txt1), 3600 * 24 * 365)
self.assertTrue(iso2seconds(txt2), 3600 * 24 * 30)
self.assertTrue(iso2seconds(txt3), 3600 * 24)
