acris 1.1.12

acris is a python library of programming patterns that we use, at acrisel, in Python projects and choose to contribute to Python community

Overview

acris is a python library providing useful programming patterns.

threaded

decorator for methods that can be executed as a thread.

example

from acris import threaded
from time import sleep

class ThreadedExample(object):
    @threaded
    def proc(self, id_, num, stall):
        s=num
        while num > 0:
            print("%s: %s" % (id_, s))
            num -= 1
            s += stall
            sleep(stall)
        print("%s: %s" % (id_, s))
        return s

class RetVal(object):
    def __init__(self, name):
        self.name=name

    def __call__(self, retval):
        print(self.name, ':', retval)

example output

te1=ThreadedExample().proc(1, 3, 1)
te2=ThreadedExample().proc(2, 3, 5)

te1.addCallback(RetVal('te1'))
te2.addCallback(RetVal('te2'))

will produce:

1: 3
2: 3
1: 4
1: 5
1: 6
te1 : 6
2: 8
2: 13
2: 18
te2 : 18

Singleton and NamedSingleton

meta class that creates singleton footprint of classes inheriting from it.

Singleton example

from acris import Singleton

class Sequence(Singleton):
    step_id=0

    def __call__(self):
        step_id=self.step_id
        self.step_id += 1
        return step_id

example output

A=Sequence()
print('A', A())
print('A', A())
B=Sequence()
print('B', B())

will produce:

A 0
A 1
B 2

NamedSingleton example

from acris import Singleton

class Sequence(NamedSingleton):
    step_id=0

    def __init__(self, name=''):
        self.name=name

    def __call__(self,):
        step_id=self.step_id
        self.step_id += 1
        return step_id

example output

A=Sequence('A')
print(A.name, A())
print(A.name, A())
B=Sequence('B')
print(B.name, B())

will produce:

A 0
A 1
B 0

Sequence

meta class to produce sequences. Sequence allows creating different sequences using name tags.

example

from acris import Sequence

A=Sequence('A')
print('A', A())
print('A', A())
B=Sequence('B')
print('B', B())

A=Sequence('A')
print('A', A())
print('A', A())
B=Sequence('B')
print('B', B())

example output

A 0
A 1
B 0
A 2
A 3
B 1

TimedSizedRotatingHandler

Use TimedSizedRotatingHandler is combining TimedRotatingFileHandler with RotatingFileHandler. Usage as handler with logging is as defined in Python’s logging how-to

example

import logging

# create logger
logger = logging.getLogger('simple_example')
logger.setLevel(logging.DEBUG)

# create console handler and set level to debug
ch = logging.TimedRotatingFileHandler()
ch.setLevel(logging.DEBUG)

# create formatter
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')

# add formatter to ch
ch.setFormatter(formatter)

# add ch to logger
logger.addHandler(ch)

# 'application' code
logger.debug('debug message')
logger.info('info message')
logger.warn('warn message')
logger.error('error message')
logger.critical('critical message')

MpLogger and LevelBasedFormatter

Multiprocessor logger using QueueListener and QueueHandler It uses TimedSizedRotatingHandler as its logging handler

It also uses acris provided LevelBasedFormatter which facilitate message formats based on record level. LevelBasedFormatter inherent from logging.Formatter and can be used as such in customized logging handlers.

example

Within main process

import logging
import time

logger=logging.getLogger(__name__)

level_formats={logging.DEBUG:"[ %(asctime)s ][ %(levelname)s ][ %(message)s ][ %(module)s.%(funcName)s.%(lineno)d ]",
                'default':   "[ %(asctime)s ][ %(levelname)s ][ %(message)s ]",
                }


mplogger=MpLogger(logging_level=logging.DEBUG, level_formats=level_formats)
mplogger.start()

logger.debug("starting sub processes")
# running processes
logger.debug("joining sub processes")

mplogger.stop()

Within individual process

import logging

logger=logging.getLogger(__name__)
logger.debug("logging from sub process")

Example output

[ 2016-12-06 13:39:56,196 ][ DEBUG ][ starting sub processes ][ mptest.<module>.178 ]
[ 2016-12-06 13:39:56,630 ][ INFO ][ proc [2663]: 0/1 - sleep 0.42sec ]
[ 2016-12-06 13:39:56,802 ][ INFO ][ proc [2664]: 0/1 - sleep  0.6sec ]
[ 2016-12-06 13:39:56,805 ][ DEBUG ][ sub processes completed ][ mptest.<module>.189 ]

Data Types

varies derivative of Python data types

MergeChainedDict

Similar to ChainedDict, but merged the keys and is actually derivative of dict.

a={1:11, 2:22}
b={3:33, 4:44}
c={1:55, 4:66}
d=MergedChainedDict(c, b, a)
print(d)

Will output:

{1: 55, 2: 22, 3: 33, 4: 66}

ResourcePool

Resource pool provides program with interface to manager resource pools. This is used as means to funnel processing.

ResourcePoolRequestor object can be used to request resource set resides in multiple pools.

ResourcePoolRequestors object manages multiple requests for multiple resources.

Sync Example

import time
from acris import resource_pool as rp
from acris import Threaded
import queue
from datetime import datetime

class MyResource1(rp.Resource): pass

class MyResource2(rp.Resource): pass

rp1=rp.ResourcePool('RP1', resource_cls=MyResource1, policy={'resource_limit': 2, }).load()
rp2=rp.ResourcePool('RP2', resource_cls=MyResource2, policy={'resource_limit': 1, }).load()

@Threaded()
def worker_awaiting(name, rp):
    print('[ %s ] %s getting resource' % (str(datetime.now()), name ) )
    r=rp.get()
    print('[ %s ] %s doing work (%s)' % (str(datetime.now()), name, repr(r)))
    time.sleep(4)
    print('[ %s ] %s returning %s' % (str(datetime.now()), name, repr(r)))
    rp.put(*r)


r1=worker_awaiting('>>> w11-direct', rp1)
r2=worker_awaiting('>>> w21-direct', rp2)
r3=worker_awaiting('>>> w22-direct', rp2)
r4=worker_awaiting('>>> w12-direct', rp1)

Sync Example Output

[ 2016-12-11 13:06:14.659569 ] >>> w11-direct getting resource
[ 2016-12-11 13:06:14.659640 ] >>> w11-direct doing work ([Resource(name:MyResource1)])
[ 2016-12-11 13:06:14.659801 ] >>> w21-direct getting resource
[ 2016-12-11 13:06:14.659834 ] >>> w21-direct doing work ([Resource(name:MyResource2)])
[ 2016-12-11 13:06:14.659973 ] >>> w22-direct getting resource
[ 2016-12-11 13:06:14.660190 ] >>> w12-direct getting resource
[ 2016-12-11 13:06:14.660260 ] >>> w12-direct doing work ([Resource(name:MyResource1)])
[ 2016-12-11 13:06:18.662362 ] >>> w11-direct returning [Resource(name:MyResource1)]
[ 2016-12-11 13:06:18.662653 ] >>> w21-direct returning [Resource(name:MyResource2)]
[ 2016-12-11 13:06:18.662826 ] >>> w12-direct returning [Resource(name:MyResource1)]
[ 2016-12-11 13:06:18.662998 ] >>> w22-direct doing work ([Resource(name:MyResource2)])
[ 2016-12-11 13:06:22.667149 ] >>> w22-direct returning [Resource(name:MyResource2)]

Async Example

import time
from acris import resource_pool as rp
from acris import Threaded
import queue
from datetime import datetime

class MyResource1(rp.Resource): pass

class MyResource2(rp.Resource): pass

rp1=rp.ResourcePool('RP1', resource_cls=MyResource1, policy={'resource_limit': 2, }).load()
rp2=rp.ResourcePool('RP2', resource_cls=MyResource2, policy={'resource_limit': 1, }).load()

class Callback(object):
    def __init__(self, notify_queue):
        self.q=notify_queue
    def __call__(self, resources=None):
        self.q.put(resources)

@Threaded()
def worker_callback(name, rp):
    print('[ %s ] %s getting resource' % (str(datetime.now()), name))
    notify_queue=queue.Queue()
    r=rp.get(callback=Callback(notify_queue))

    if not r:
        print('[ %s ] %s doing work before resource available' % (str(datetime.now()), name,))
        print('[ %s ] %s waiting for resources' % (str(datetime.now()), name,))
        ticket=notify_queue.get()
        r=rp.get(ticket=ticket)

    print('[ %s ] %s doing work (%s)' % (str(datetime.now()), name, repr(r)))
    time.sleep(2)
    print('[ %s ] %s returning (%s)' % (str(datetime.now()), name, repr(r)))
    rp.put(*r)

r1=worker_callback('>>> w11-callback', rp1)
r2=worker_callback('>>> w21-callback', rp2)
r3=worker_callback('>>> w22-callback', rp2)
r4=worker_callback('>>> w12-callback', rp1)

Async Example Output

[ 2016-12-11 13:08:24.410447 ] >>> w11-callback getting resource
[ 2016-12-11 13:08:24.410539 ] >>> w11-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-11 13:08:24.410682 ] >>> w21-callback getting resource
[ 2016-12-11 13:08:24.410762 ] >>> w21-callback doing work ([Resource(name:MyResource2)])
[ 2016-12-11 13:08:24.410945 ] >>> w22-callback getting resource
[ 2016-12-11 13:08:24.411227 ] >>> w22-callback doing work before resource available
[ 2016-12-11 13:08:24.411273 ] >>> w12-callback getting resource
[ 2016-12-11 13:08:24.411334 ] >>> w22-callback waiting for resources
[ 2016-12-11 13:08:24.411452 ] >>> w12-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-11 13:08:26.411901 ] >>> w11-callback returning ([Resource(name:MyResource1)])
[ 2016-12-11 13:08:26.412200 ] >>> w21-callback returning ([Resource(name:MyResource2)])
[ 2016-12-11 13:08:26.412505 ] >>> w22-callback doing work ([Resource(name:MyResource2)])
[ 2016-12-11 13:08:26.416130 ] >>> w12-callback returning ([Resource(name:MyResource1)])
[ 2016-12-11 13:08:28.416001 ] >>> w22-callback returning ([Resource(name:MyResource2)])

Requestor Example

import time
from acris import resource_pool as rp
from acris import Threaded
import queue
from datetime import datetime

class MyResource1(rp.Resource): pass

class MyResource2(rp.Resource): pass

rp1=rp.ResourcePool('RP1', resource_cls=MyResource1, policy={'resource_limit': 2, }).load()
rp2=rp.ResourcePool('RP2', resource_cls=MyResource2, policy={'resource_limit': 2, }).load()

class Callback(object):
    def __init__(self, notify_queue):
        self.q=notify_queue
    def __call__(self, ready=False):
        self.q.put(ready)

@Threaded()
def worker_callback(name, rps):
    print('[ %s ] %s getting resource' % (str(datetime.now()), name))
    notify_queue=queue.Queue()
    callback=Callback(notify_queue, name=name)
    request=rp.Requestor(request=rps, callback=callback)

    if request.is_reserved():
        resources=request.get()
    else:
        print('[ %s ] %s doing work before resource available' % (str(datetime.now()), name,))
        print('[ %s ] %s waiting for resources' % (str(datetime.now()), name,))
        notify_queue.get()
        resources=request.get()

    print('[ %s ] %s doing work (%s)' % (str(datetime.now()), name, repr(resources)))
    time.sleep(2)
    print('[ %s ] %s returning (%s)' % (str(datetime.now()), name, repr(resources)))
    request.put(*resources)

r1=worker_callback('>>> w11-callback', [(rp1,1),])
r2=worker_callback('>>> w21-callback', [(rp1,1),(rp2,1)])
r3=worker_callback('>>> w22-callback', [(rp1,1),(rp2,1)])
r4=worker_callback('>>> w12-callback', [(rp1,1),])

Requestor Example Output

[ 2016-12-13 06:27:54.924629 ] >>> w11-callback getting resource
[ 2016-12-13 06:27:54.925094 ] >>> w21-callback getting resource
[ 2016-12-13 06:27:54.925453 ] >>> w22-callback getting resource
[ 2016-12-13 06:27:54.926188 ] >>> w12-callback getting resource
[ 2016-12-13 06:27:54.932922 ] >>> w11-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-13 06:27:54.933709 ] >>> w12-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-13 06:27:54.938425 ] >>> w22-callback doing work before resource available
[ 2016-12-13 06:27:54.938548 ] >>> w22-callback waiting for resources
[ 2016-12-13 06:27:54.939256 ] >>> w21-callback doing work before resource available
[ 2016-12-13 06:27:54.939267 ] >>> w21-callback waiting for resources
[ 2016-12-13 06:27:56.936881 ] >>> w11-callback returning ([Resource(name:MyResource1)])
[ 2016-12-13 06:27:56.937543 ] >>> w12-callback returning ([Resource(name:MyResource1)])
[ 2016-12-13 06:27:56.947615 ] >>> w22-callback doing work ([Resource(name:MyResource2), Resource(name:MyResource1)])
[ 2016-12-13 06:27:56.948587 ] >>> w21-callback doing work ([Resource(name:MyResource2), Resource(name:MyResource1)])
[ 2016-12-13 06:27:58.949812 ] >>> w22-callback returning ([Resource(name:MyResource2), Resource(name:MyResource1)])
[ 2016-12-13 06:27:58.950064 ] >>> w21-callback returning ([Resource(name:MyResource2), Resource(name:MyResource1)])

Virtual ResourcePool

Like ResourcePool, VResourcePool manages resources. The main difference between the two is that ResourcePool manages physical resource objects. VResourcePool manages virtual resources (VResource) that only represent physical resources. VResources can not be activated or deactivated.

One unique property VResourcePool enables is that request could be returned by quantity.

Virtual Requestors Example

import time
from acris import virtual_resource_pool as rp
from acris.threaded import Threaded
from acris.mplogger import create_stream_handler
import queue
from datetime import datetime

class MyResource1(rp.Resource): pass
class MyResource2(rp.Resource): pass

rp1=rp.ResourcePool('RP1', resource_cls=MyResource1, policy={'resource_limit': 2, }).load()
rp2=rp.ResourcePool('RP2', resource_cls=MyResource2, policy={'resource_limit': 1, }).load()

class Callback(object):
    def __init__(self, notify_queue, name=''):
        self.q=notify_queue
        self.name=name
    def __call__(self,received=False):
        self.q.put(received)

requestors=rp.Requestors()

@Threaded()
def worker_callback(name, rps):
    print('[ %s ] %s getting resource' % (str(datetime.now()), name))
    notify_queue=queue.Queue()
    callback=Callback(notify_queue, name=name)
    request_id=requestors.reserve(request=rps, callback=callback)

    if not requestors.is_reserved(request_id):
        print('[ %s ] %s doing work before resource available' % (str(datetime.now()), name,))
        notify_queue.get()
    resources=requestors.get(request_id)

    print('[ %s ] %s doing work (%s)' % (str(datetime.now()), name, repr(resources)))
    time.sleep(1)
    print('[ %s ] %s returning (%s)' % (str(datetime.now()), name, repr(resources)))
    requestors.put_requested(rps)

r2=worker_callback('>>> w21-callback', [(rp1,1), (rp2,1)])
r1=worker_callback('>>> w11-callback', [(rp1,1),])
r3=worker_callback('>>> w22-callback', [(rp1,1), (rp2,1)])
r4=worker_callback('>>> w12-callback', [(rp1,1),])

Virtual Requestor Example Output

[ 2016-12-16 14:27:53.224110 ] >>> w21-callback getting resource
[ 2016-12-16 14:27:53.224750 ] >>> w11-callback getting resource
[ 2016-12-16 14:27:53.225567 ] >>> w22-callback getting resource
[ 2016-12-16 14:27:53.226220 ] >>> w12-callback getting resource
[ 2016-12-16 14:27:53.237146 ] >>> w11-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-16 14:27:53.238361 ] >>> w12-callback doing work before resource available
[ 2016-12-16 14:27:53.241046 ] >>> w21-callback doing work before resource available
[ 2016-12-16 14:27:53.242350 ] >>> w22-callback doing work ([Resource(name:MyResource1), Resource(name:MyResource2)])
[ 2016-12-16 14:27:54.238443 ] >>> w11-callback returning ([Resource(name:MyResource1)])
[ 2016-12-16 14:27:54.246868 ] >>> w22-callback returning ([Resource(name:MyResource1), Resource(name:MyResource2)])
[ 2016-12-16 14:27:54.257040 ] >>> w12-callback doing work ([Resource(name:MyResource1)])
[ 2016-12-16 14:27:54.259858 ] >>> w21-callback doing work ([Resource(name:MyResource1), Resource(name:MyResource2)])
[ 2016-12-16 14:27:55.258659 ] >>> w12-callback returning ([Resource(name:MyResource1)])
[ 2016-12-16 14:27:55.262741 ] >>> w21-callback returning ([Resource(name:MyResource1), Resource(name:MyResource2)])

Mediator

Class interface to generator allowing query of has_next()

Example

from acris import Mediator

def yrange(n):
    i = 0
    while i < n:
        yield i
        i += 1

n=10
m=Mediator(yrange(n))
for i in range(n):
    print(i, m.has_next(3), next(m))
print(i, m.has_next(), next(m))

Example Output

0 True 0
1 True 1
2 True 2
3 True 3
4 True 4
5 True 5
6 True 6
7 True 7
8 False 8
9 False 9
Traceback (most recent call last):
  File "/private/var/acrisel/sand/acris/acris/acris/example/mediator.py", line 19, in <module>
    print(i, m.has_next(), next(m))
  File "/private/var/acrisel/sand/acris/acris/acris/acris/mediator.py", line 38, in __next__
    value=next(self.generator)
StopIteration