def downloadDeferWorkQueue(urls):
"""
Uses a DeferWorkQueue to download URLs with
dynamic number of concurrent downloads, here we
will download 10 concurrently.
NOTE: In this case we aren't handling failures like in the others
"""
ret = {}
def _setUrl(url, content):
ret[url] = content
getPage = lambda url : http.getPage(url,
connectionTimeout=30,
timeout=30
).addCallback(lambda content : _setUrl(url, content))
dwq = defer_work_queue.DeferWorkQueue(10)
for url in urls:
dwq.add(getPage, url)
return defer_work_queue.waitForCompletion(dwq).addCallback(lambda _ : ret)
def run(options):
logging.logPrint('Starting')
batchConfig = config.configFromStream(open(options.configFile), lazy=True)
machineConf = config.configFromStream(open('/tmp/machine.conf'))
state = State(
options.workflowConfig, options.batchStatesFile,
_validateWrapper(
batchConfig('batch_pipeline.pipeline.PIPELINE_TEMPLATE'),
pipeline_misc.determineWrapper(
machineConf,
batchConfig('batch_pipeline.pipeline.PIPELINE_TEMPLATE'))),
_interpretBatchFile(options.batchFile),
_extractInnerPipelineConfig(batchConfig),
batchConfig('pipeline.PIPELINE_WRAPPER_NAME'),
int(batchConfig('batch.options.CONCURRENT_PRERUN')),
int(batchConfig('batch.options.CONCURRENT_PIPELINES')),
int(batchConfig('batch.options.CONCURRENT_POSTRUN')))
logging.logPrint('Queuing any incomplete work')
queueCount = _queueIncompleteWork(state)
logging.logPrint('Queued: %d' % queueCount)
if state.pipelinesQueue.hasWork():
yield defer_work_queue.waitForCompletion(state.pipelinesQueue)
for batchState in state.batchStates.values():
if 'state' not in batchState or batchState['state'] == 'failed':
raise JobFailed()
def run(options):
logging.logPrint('Starting')
batchConfig = config.configFromStream(open(options.configFile), lazy=True)
machineConf = config.configFromStream(open('/tmp/machine.conf'))
state = State(options.workflowConfig,
options.batchStatesFile,
_validateWrapper(batchConfig('batch_pipeline.pipeline.PIPELINE_TEMPLATE'),
pipeline_misc.determineWrapper(machineConf,
batchConfig('batch_pipeline.pipeline.PIPELINE_TEMPLATE'))),
_interpretBatchFile(options.batchFile),
_extractInnerPipelineConfig(batchConfig),
batchConfig('pipeline.PIPELINE_WRAPPER_NAME'),
int(batchConfig('batch.options.CONCURRENT_PRERUN')),
int(batchConfig('batch.options.CONCURRENT_PIPELINES')),
int(batchConfig('batch.options.CONCURRENT_POSTRUN')))
logging.logPrint('Queuing any incomplete work')
queueCount = _queueIncompleteWork(state)
logging.logPrint('Queued: %d' % queueCount)
if state.pipelinesQueue.hasWork():
yield defer_work_queue.waitForCompletion(state.pipelinesQueue)
for batchState in state.batchStates.values():
if 'state' not in batchState or batchState['state'] == 'failed':
raise JobFailed()
def initialize(self):
self.cache = yield mongo_cache.createCache("tags_cache", lambda d: func.updateDict(d, {"_id": d["tag_name"]}))
self.persistManager.addDependent(self)
# If there are a lot of tags we want to parallelize caching them
self.workQueue.parallel = 100
# Force all tags to be cached
tags = yield self.persistManager.listTags()
for tagName in tags:
yield self.persistManager.loadTag(tagName)
yield defer_work_queue.waitForCompletion(self.workQueue)
# Now that we are done, set it back to 1
self.workQueue.parallel = 1
def initialize(self):
self.cache = yield mongo_cache.createCache('tags_cache',
lambda d : func.updateDict(d, {'_id': d['tag_name']}))
self.persistManager.addDependent(self)
# If there are a lot of tags we want to parallelize caching them
self.workQueue.parallel = 100
# Force all tags to be cached
tags = yield self.persistManager.listTags()
for tagName in tags:
yield self.persistManager.loadTag(tagName)
yield defer_work_queue.waitForCompletion(self.workQueue)
# Now that we are done, set it back to 1
self.workQueue.parallel = 1
def validate(state, protocolConf, pipelineConf):
success = []
failure = []
def _validateElement(elm):
d = _validateType(state, pipelineConf, elm)
d.addCallback(lambda v : success.extend(v))
d.addErrback(lambda f : failure.append(f))
return d
def _validateElementFunc(elm):
return lambda : _validateElement(elm)
dwq = defer_work_queue.DeferWorkQueue(10)
dwq.extend([_validateElementFunc(e) for e in protocolConf])
yield defer_work_queue.waitForCompletion(dwq)
if failure:
failure[0].raiseException()
defer.returnValue(success)
def validate(state, protocolConf, pipelineConf):
success = []
failure = []
def _validateElement(elm):
d = _validateType(state, pipelineConf, elm)
d.addCallback(lambda v: success.extend(v))
d.addErrback(lambda f: failure.append(f))
return d
def _validateElementFunc(elm):
return lambda: _validateElement(elm)
dwq = defer_work_queue.DeferWorkQueue(10)
dwq.extend([_validateElementFunc(e) for e in protocolConf])
yield defer_work_queue.waitForCompletion(dwq)
if failure:
failure[0].raiseException()
defer.returnValue(success)
def downloadDeferWorkQueue(urls):
"""
Uses a DeferWorkQueue to download URLs with
dynamic number of concurrent downloads, here we
will download 10 concurrently.
NOTE: In this case we aren't handling failures like in the others
"""
ret = {}
def _setUrl(url, content):
ret[url] = content
getPage = lambda url: http.getPage(url, connectionTimeout=30, timeout=30
).addCallback(lambda content: _setUrl(
url, content))
dwq = defer_work_queue.DeferWorkQueue(10)
for url in urls:
dwq.add(getPage, url)
return defer_work_queue.waitForCompletion(dwq).addCallback(lambda _: ret)
