def process(self, block_iter):
(self._limit_files_per_ds, self._files_per_ds) = ({}, {}
) # reset counters
if self._limit_files_fraction >= 0:
block_list = list(DataProcessor.process(self, block_iter))
goal_per_ds = {} # calculate file limit per dataset
for (dataset_name, fn_list_len) in self._files_per_ds.items():
goal_per_ds[dataset_name] = int(
self._limit_files_fraction * fn_list_len) or 1
for block in block_list:
self._reduce_fn_list(block, goal_per_ds)
yield block
else:
for block in DataProcessor.process(self, block_iter):
yield block
def process(self, block_iter): if self.enabled() and self._config: block_list = list(DataProcessor.process(self, block_iter)) if (self._target_jobs > 0) or (self._target_jobs_ds > 0): self._set_split_opt(self._config, 'files per job', dict(self._files), self._target_jobs, self._target_jobs_ds) self._set_split_opt(self._config, 'events per job', dict(self._entries), self._target_jobs, self._target_jobs_ds) self._config = None return block_list return block_iter
def process(self, block_iter):
if self.enabled() and self._config:
block_list = list(DataProcessor.process(self, block_iter))
if (self._target_jobs > 0) or (self._target_jobs_ds > 0):
self._set_split_opt(self._config, 'files per job',
dict(self._files), self._target_jobs,
self._target_jobs_ds)
self._set_split_opt(self._config, 'events per job',
dict(self._entries), self._target_jobs,
self._target_jobs_ds)
self._config = None
return block_list
return block_iter
def process(self, blockIter):
self._recordedURL = set()
self._recordedBlock = set()
return DataProcessor.process(self, blockIter)
def process(self, blockIter): self._recordedURL = set() self._recordedBlock = set() return DataProcessor.process(self, blockIter)
def process(self, block_iter): self._recorded_url = set() # reset records self._recorded_block = set() return DataProcessor.process(self, block_iter)
def process(self, block_iter):
self._recorded_url = set() # reset records
self._recorded_block = set()
return DataProcessor.process(self, block_iter)
