# >>> arr = stamper([arr])
# Step 3 (representing computational resources)
# creates a resource that allows to utilize all local processors
local_computer = NuMap()
# Step 4 (creating processing nodes)
# this attaches a single computational resource to the two processing nodes
# the stamper_node will be tracked i.e. it will store the results of computation
# in memory.
cleaner_node = Piper(cleaner, parallel=local_computer)
stamper_node = Piper(stamper, parallel=local_computer, track=True)
# Step 5 (constructing a workflow graph)
# we construct a workflow graph add the two processing nodes and define the
# connection between them.
workflow = Plumber()
workflow.add_pipe((cleaner_node, stamper_node))
# Step 6 (execute the workflow)
# this starts the workflow, processes data in the "background" and waits
# until all data-items have been processed.
workflow.start([['AGA.TA', 'TG..AA']])
workflow.run()
workflow.wait()
results = workflow.stats['pipers_tracked'][stamper_node][0]
for seq in results.values():
print "Object \"%s\" has time stamp: %s " % (seq, seq.meta['timestamp'])
from papy.core import Plumber, Piper
def l33t(inbox):
word = inbox[0]
return word.replace('e', '3').replace('o', '0')
def l33ter(inbox):
word = inbox[0]
return word.replace('l', '1')
# execution endgine
numap = NuMap()
# function nodes
l33t_piper = Piper(l33t, parallel=numap)
l33ter_piper = Piper(l33ter, parallel=numap, track=True)
# topology
pipeline = Plumber()
pipeline.add_pipe((l33t_piper, l33ter_piper))
end = pipeline.get_outputs()[0]
# runtime
pipeline.start([['hello', 'world']])
pipeline.run()
pipeline.wait()
print pipeline.stats['pipers_tracked'][end]
assert [{0: 'h3110', 1: 'w0r1d'}] == pipeline.stats['pipers_tracked'][end]
def l33t(inbox):
word = inbox[0]
return word.replace('e', '3').replace('o', '0')
def l33ter(inbox):
word = inbox[0]
return word.replace('l', '1')
# execution endgine
numap = NuMap()
# function nodes
l33t_piper = Piper(l33t, parallel=numap)
l33ter_piper = Piper(l33ter, parallel=numap, track=True)
# topology
pipeline = Plumber()
pipeline.add_pipe((l33t_piper, l33ter_piper))
end = pipeline.get_outputs()[0]
# runtime
pipeline.start([['hello', 'world']])
pipeline.run()
pipeline.wait()
print pipeline.stats['pipers_tracked'][end]
assert [{0: 'h3110', 1: 'w0r1d'}] == pipeline.stats['pipers_tracked'][end]
# wraps timestamp
stamper = Worker(timestamp)
# >>> arr = stamper([arr])
# Step 3 (representing computational resources)
# creates a resource that allows to utilize all local processors
local_computer = NuMap()
# Step 4 (creating processing nodes)
# this attaches a single computational resource to the two processing nodes
# the stamper_node will be tracked i.e. it will store the results of computation
# in memory.
cleaner_node = Piper(cleaner, parallel=local_computer)
stamper_node = Piper(stamper, parallel=local_computer, track=True)
# Step 5 (constructing a workflow graph)
# we construct a workflow graph add the two processing nodes and define the
# connection between them.
workflow = Plumber()
workflow.add_pipe((cleaner_node, stamper_node))
# Step 6 (execute the workflow)
# this starts the workflow, processes data in the "background" and waits
# until all data-items have been processed.
workflow.start([['AGA.TA', 'TG..AA']])
workflow.run()
workflow.wait()
results = workflow.stats['pipers_tracked'][stamper_node][0]
for seq in results.values():
print "Object \"%s\" has time stamp: %s " % (seq, seq.meta['timestamp'])
