Skip to main content

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

Project description

Overview

acrilib is a python library providing useful programming patterns and tools. acrilib started as Acrisel’s internal idioms and utilities for programmers. The main key is that this library is completely independent. It does not use any external packages beside what provided by Python.

It includes:
  1. programming idioms that are repeatedly used by programmers.

  2. helpers functions for logging and other utilities.

We decided to contribute this library to Python community as a token of appreciation to what this community enables us.

We hope that you will find this library useful and helpful as we find it.

If you have comments or insights, please don’t hesitate to contact us at support@acrisel.com

Programming Idoms

threaded

decorator for methods that can be executed as a thread. RetriveAsycValue callable class used in the example below provide means to access results. One can provide their own callable to pass results.

example

from acris import threaded, RetriveAsycValue
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

example output

print("starting workers")
te1 = ThreadedExample().proc('TE1', 3, 1)
te2 = ThreadedExample().proc('TE2', 3, 1)

print("collecting results")
te1_callback = RetriveAsycValue('te1')
te1.addCallback(te1_callback)
te2_callback = RetriveAsycValue('te2')
te2.addCallback(te2_callback)

print('joining t1')
te1.join()
print('joined t1')
print('%s callback result: %s' % (te1_callback.name, te1_callback.result))
result = te1.syncResult()
print('te1 syncResult : %s' %result)

result = te2.syncResult()
print('te2 syncResult : %s' % result)
print('%s callback result: %s' % (te2_callback.name, te2_callback.result))

will produce:

starting workers
TE1: 3
TE2: 3
collecting results
joining t1
TE1: 4
TE2: 4
TE1: 5
TE2: 5
TE1: 6
TE2: 6
joined t1
te1 callback result: 6
te1 syncResult : 6
te2 syncResult : 6
te2 callback result: 6

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

TBD

Decorators

Useful decorators for production and debug.

traced_method

logs entry and exit of function or method.

from acris import traced_method

traced = traced_method(print, print_args=True, print_result=True)

class Oper(object):
    def __init__(self, value):
        self.value = value

    def __repr__(self):
        return str(self.value)

    @traced
    def mul(self, value):
        self.value *= value
        return self

    @traced
    def add(self, value):
        self.value += value
        return self

o=Oper(3)
print(o.add(2).mul(5).add(7).mul(8))

would result with the following output:

[ add ][ entering][ args: (2) ][ kwargs: {} ][ trace_methods.py.Oper(39) ]
[ add ][ exiting ] [ time span: 0:00:00.000056][ result: 5 ][ trace_methods.py.Oper(39) ]
[ mul ][ entering][ args: (5) ][ kwargs: {} ][ trace_methods.py.Oper(34) ]
[ mul ][ exiting ] [ time span: 0:00:00.000010][ result: 25 ][ trace_methods.py.Oper(34) ]
[ add ][ entering][ args: (7) ][ kwargs: {} ][ trace_methods.py.Oper(39) ]
[ add ][ exiting ] [ time span: 0:00:00.000007][ result: 32 ][ trace_methods.py.Oper(39) ]
[ mul ][ entering][ args: (8) ][ kwargs: {} ][ trace_methods.py.Oper(34) ]
[ mul ][ exiting ] [ time span: 0:00:00.000008][ result: 256 ][ trace_methods.py.Oper(34) ]
256

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}

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

Change History

Version 1.0.2

  1. Initial publication to open source

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

acrilib-1.0.2.tar.gz (17.0 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page