zope.minmax

Contents:

Conflict Resolution using Maximum or Minimum Values

The zope.minmax.AbstractValue class provides a super class which can be subclassed to store arbitrary homogeneous values in a persistent storage and apply different conflict resolution policies.

class zope.minmax.interfaces.IAbstractValue[source]

Bases: persistent.interfaces.IPersistent

A persistent value with the conflict resolution.

The values are expected to be homogeneous.

class zope.minmax.AbstractValue(value=None)[source]

Bases: persistent.Persistent

Abstract implementation of zope.minmax.interfaces.IAbstractValue.

Subclasses must implement _p_resolveConflict.

_p_resolveConflict(old, commited, new)[source]

Subclasses must implement this method.

Raises:NotImplementedError – Unless subclasses override.

The subclasses defined here are resolving the conflicts using always either the maximum or the minimum of the conflicting values.

Maximum

class zope.minmax.Maximum(value=None)[source]

Bases: zope.minmax._minmax.AbstractValue

The zope.minmax.Maximum class always resolves conflicts favoring the maximum value. Let’s instantiate one object and verify that it satisfies the interface.

>>> import zope.minmax
>>> import zope.interface.verify
>>> max_favored = zope.minmax.Maximum()
>>> zope.interface.verify.verifyObject(
...     zope.minmax.interfaces.IAbstractValue, max_favored)
True

We can confirm that the initial value is zero.

>>> bool(max_favored)
False
>>> print(max_favored.value)
None

Now, we can store a new value in the object.

>>> max_favored.value = 11
>>> print(max_favored.value)
11
>>> bool(max_favored)
True

Or we can use the methods.

>>> max_favored.__setstate__(4532)
>>> max_favored.__getstate__()
4532
>>> print(max_favored.value)
4532
>>> bool(max_favored)
True

Do notice that using a direct assignment to the value attribute is a more natural use.

Minimum

class zope.minmax.Minimum(value=None)[source]

Bases: zope.minmax._minmax.AbstractValue

The zope.minmax.Minimum class always resolves conflicts favoring the minimum value. Again, we instantiate an object and verify that it satisfies the interface.

>>> min_favored = zope.minmax.Minimum()
>>> zope.interface.verify.verifyObject(
...     zope.minmax.interfaces.IAbstractValue, min_favored)
True

We need a confirmation that the initial value is zero.

>>> bool(min_favored)
False
>>> print(min_favored.value)
None

Let’s populate this one too.

>>> min_favored.value = 22
>>> print(min_favored.value)
22
>>> bool(min_favored)
True

Or we can use the methods, again.

>>> min_favored.__setstate__(8796)
>>> min_favored.__getstate__()
8796
>>> print(min_favored.value)
8796
>>> bool(min_favored)
True

Please, notice, again, that using a direct assignment to the value attribute is a more natural use.

Conflict Resolution

Now, we need to exercise the conflict resolution interface. First for the zope.minmax.Maximum:

Let’s try differing values larger than the old value.

>>> max_favored._p_resolveConflict(max_favored.value, 4536, 4535)
4536
>>> max_favored._p_resolveConflict(max_favored.value, 4573, 4574)
4574

What happens when all the values are equal, including the old.

>>> max_favored._p_resolveConflict(max_favored.value, 4532, 4532)
4532

Notice that when the old value is larger than both the committed and new, it is still disregarded.

>>> max_favored._p_resolveConflict(max_favored.value, 4531, 4530)
4531

Now, the zope.minmax.Minimum:

Let’s try differing values smaller than the old value.

>>> min_favored._p_resolveConflict(min_favored.value, 8792, 8791)
8791
>>> min_favored._p_resolveConflict(min_favored.value, 8785, 8786)
8785

What happens when all the values are equal, including the old.

>>> min_favored._p_resolveConflict(min_favored.value, 8796, 8796)
8796

Notice that when the old value is smaller than both the committed and new, it is still disregarded.

>>> min_favored._p_resolveConflict(min_favored.value, 8798, 8799)
8798

How about an example that is not numerical?

>>> max_word = zope.minmax.Maximum('joy')
>>> print(max_word.value)
joy
>>> bool(max_word)
True
>>> max_word._p_resolveConflict(max_word.value, 'happiness', 'exuberance')
'happiness'
>>> max_word._p_resolveConflict(max_word.value, 'exuberance', 'happiness')
'happiness'
>>> min_word = zope.minmax.Minimum(max_word.value)
>>> print(min_word.value)
joy
>>> bool(min_word)
True
>>> min_word._p_resolveConflict(min_word.value, 'happiness', 'exuberance')
'exuberance'
>>> min_word._p_resolveConflict(min_word.value, 'exuberance', 'happiness')
'exuberance'

As indicated, we don’t need to have numbers, just homegeneous items. The homogeneous values are not really inherently required. However, it makes no sense to apply min() or max() on, say, one number and one string. Simply, the ordering relations do not work at all on heterogeneous values.

Hacking on zope.minmax

Getting the Code

The main repository for zope.minmax is in the Zope Foundation Github repository:

You can get a read-only checkout from there:

$ git clone https://github.com/zopefoundation/zope.minmax.git

or fork it and get a writeable checkout of your fork:

$ git clone git@github.com/jrandom/zope.minmax.git

The project also mirrors the trunk from the Github repository as a Bazaar branch on Launchpad:

https://code.launchpad.net/zope.minmax

You can branch the trunk from there using Bazaar:

$ bzr branch lp:zope.minmax

Working in a virtualenv

Installing

If you use the virtualenv package to create lightweight Python development environments, you can run the tests using nothing more than the python binary in a virtualenv. First, create a scratch environment:

$ /path/to/virtualenv --no-site-packages /tmp/hack-zope.minmax

Next, get this package registered as a “development egg” in the environment:

$ /tmp/hack-zope.minmax/bin/python setup.py develop

Running the tests

Run the tests using the build-in setuptools testrunner:

$ /tmp/hack-zope.minmax/bin/python setup.py test
running test
.........
----------------------------------------------------------------------
Ran 9 tests in 0.000s

OK

If you have the nose package installed in the virtualenv, you can use its testrunner too:

$ /tmp/hack-zope.minmax/bin/easy_install nose
...
$ /tmp/hack-zope.minmax/bin/nosetests
.........
----------------------------------------------------------------------
Ran 18 tests in 0.000s

OK

If you have the coverage pacakge installed in the virtualenv, you can see how well the tests cover the code:

$ /tmp/hack-zope.minmax/bin/easy_install nose coverage
...
$ /tmp/hack-zope.minmax/bin/nosetests --with coverage
running nosetests
..................
Name                        Stmts   Miss Branch BrPart  Cover   Missing
-----------------------------------------------------------------------
zope/minmax.py                  1      0      0      0   100%
zope/minmax/_minmax.py         22      0      2      0   100%
zope/minmax/interfaces.py       2      0      0      0   100%
-----------------------------------------------------------------------
TOTAL                          25      0      2      0   100%
----------------------------------------------------------------------
Ran 18 tests in 0.027s

OK

Building the documentation

zope.minmax uses the nifty Sphinx documentation system for building its docs. Using the same virtualenv you set up to run the tests, you can build the docs:

$ /tmp/hack-zope.minmax/bin/easy_install Sphinx
...
$ bin/sphinx-build -b html -d docs/_build/doctrees docs docs/_build/html
...
build succeeded.

You can also test the code snippets in the documentation:

$ bin/sphinx-build -b doctest -d docs/_build/doctrees docs docs/_build/doctest
...

Doctest summary
===============
   42 tests
    0 failures in tests
    0 failures in setup code
build succeeded.
Testing of doctests in the sources finished, look at the  \
    results in _build/doctest/output.txt.

Using zc.buildout

Setting up the buildout

zope.minmax ships with its own buildout.cfg file and bootstrap.py for setting up a development buildout:

$ /path/to/python2.7 bootstrap.py
...
Generated script '.../bin/buildout'
$ bin/buildout
Develop: '/home/jrandom/projects/Zope/zope.minmax/.'
...
Generated script '.../bin/sphinx-quickstart'.
Generated script '.../bin/sphinx-build'.

Running the tests

Run the tests:

$ bin/test --all
Running zope.testing.testrunner.layer.UnitTests tests:
  Set up zope.testing.testrunner.layer.UnitTests in 0.000 seconds.
  Ran 400 tests with 0 failures and 0 errors in 0.366 seconds.
Tearing down left over layers:
  Tear down zope.testing.testrunner.layer.UnitTests in 0.000 seconds.

Using tox

Running Tests on Multiple Python Versions

tox is a Python-based test automation tool designed to run tests against multiple Python versions. It creates a virtualenv for each configured version, installs the current package and configured dependencies into each virtualenv, and then runs the configured commands.

zope.minmax configures the following tox environments via its tox.ini file:

  • The py26, py27, py33, py34, pypy`, and ``pypy3 environments build a virtualenv with the appropriate interpreter, installs zope.minmax and dependencies, and runs the tests via python setup.py test -q.
  • The coverage environment builds a virtualenv with python2.6, installs zope.minmax, installs nose and coverage, and runs nosetests with statement coverage.
  • The docs environment builds a virtualenv with python2.6, installs zope.minmax, installs Sphinx and dependencies, and then builds the docs and exercises the doctest snippets.

This example requires that you have a working python2.6 on your path, as well as installing tox:

$ tox -e py26
GLOB sdist-make: .../zope.interface/setup.py
py26 sdist-reinst: .../zope.interface/.tox/dist/zope.interface-4.0.2dev.zip
py26 runtests: commands[0]
.........
----------------------------------------------------------------------
Ran 9 tests in 0.152s

OK
___________________________________ summary ____________________________________
py26: commands succeeded
congratulations :)

Running tox with no arguments runs all the configured environments, including building the docs and testing their snippets:

$ tox
GLOB sdist-make: .../zope.interface/setup.py
py26 sdist-reinst: .../zope.interface/.tox/dist/zope.interface-4.0.2dev.zip
py26 runtests: commands[0]
...
Doctest summary
===============
  42 tests
   0 failures in tests
   0 failures in setup code
   0 failures in cleanup code
build succeeded.
___________________________________ summary ____________________________________
py26: commands succeeded
py27: commands succeeded
py33: commands succeeded
py34: commands succeeded
pypy: commands succeeded
coverage: commands succeeded
docs: commands succeeded
congratulations :)

Contributing to zope.minmax

Submitting a Bug Report

zope.minmax tracks its bugs on Github:

Please submit bug reports and feature requests there.

Sharing Your Changes

Note

Please ensure that all tests are passing before you submit your code. If possible, your submission should include new tests for new features or bug fixes, although it is possible that you may have tested your new code by updating existing tests.

If have made a change you would like to share, the best route is to fork the Githb repository, check out your fork, make your changes on a branch in your fork, and push it. You can then submit a pull request from your branch:

If you branched the code from Launchpad using Bazaar, you have another option: you can “push” your branch to Launchpad:

$ bzr push lp:~jrandom/zope.minmax/cool_feature

After pushing your branch, you can link it to a bug report on Launchpad, or request that the maintainers merge your branch using the Launchpad “merge request” feature.

Indices and tables