printobject 0.1.6

Generic pretty printer of python objects

Python version support: CPython 2.6, 2.7, 3.2, 3.3 and PyPy.

Installation

$ pip install printobject

Usage

The standard library pprint module is great at visualizing all kinds of built-in types like lists, dicts and tuples. But it does not attempt to introspect user defined types. This is where printobject comes in. It dumps the internals of any object as a dict, and pretty prints using pprint.

Key points:

• Any type of object can printed, but depending on the type the output will be more or less insightful.

• Object introspection is based on the use of dir rather than __dict__ directly.

• Object attributes only include attributes owned by the object, omitting class attributes.

• Callables are omitted when introspecting objects. The goal is to visualize the data in objects.

• The synthetic attributes ___name___ and ___type___ (yes, that’s three underscores!!!) are included in order to provide some metadata about the object being printed.

Modules

This modules defines a number of test_xxx functions at module level. They are included in a tests list and visible in the output, but not listed at top level because they are callables.

>>> import sys
>>> from printobject import pp

>>> pp(sys.modules[__name__])

{'___name___': '__main__',
 '___type___': '<module {id0}>',
 '__builtins__': <module 'builtins' (built-in)>,
 '__cached__': None,
 '__file__': '/home/user/code/printobject/printobject/demos.py',
 '__loader__': <_frozen_importlib.SourceFileLoader object at 0xb71c520c>,
 '__name__': '__main__',
 '__package__': 'printobject',
 'absolute_import': _Feature((2, 5, 0, 'alpha', 1), (3, 0, 0, 'alpha', 0), 16384),
 'defaults': ('Module',),
 're': <module 're' from '/home/user/code/printobject/.tox/py33/lib/python3.3/re.py'>,
 'sys': <module 'sys' (built-in)>,
 'tests': [<function test_module at 0xb72d23d4>,
           <function test_class at 0xb71c60bc>,
           <function test_instance at 0xb71c6104>,
           <function test_instance_collapsed at 0xb71c614c>,
           <function test_class_old at 0xb71c6194>,
           <function test_instance_old at 0xb71c61dc>,
           <function test_instance_old_collapsed at 0xb71c6224>,
           <function test_function at 0xb71c626c>,
           <function test_method at 0xb71c62b4>,
           <function test_lambda at 0xb71c62fc>,
           <function test_iterable at 0xb71c6344>,
           <function test_generator at 0xb71c638c>]}

Classes

>>> class Node(object):
...     classatt = 'hidden'
...     def __init__(self, name):
...         self.name = name

>>> from printobject import pp
>>> pp(Node)

{'___name___': 'Node',
 '___type___': '<type {id0}>',
 '__weakref__': {'___name___': '__weakref__',
                 '___type___': '<getset_descriptor {id1}>'},
 'classatt': "'hidden'"}

Instances

Object graphs often aren’t fully acyclic. Where cycles exist it usually doesn’t make sense to unroll them, so an object encountered more than once is displayed with the dup tag. Objects also get assigned id’s, so that in the case below it’s clear that dup <Node {id0}>, which appears in the refs attribute of c, is referring back to a.

>>> a, b, c, d = Node('A'), Node('B'), Node('C'), Node('D')
>>> a.refs = [b, d]
>>> b.refs = [c]
>>> c.refs = [a]
>>> d.refs = [c]

>>> from printobject import pp
>>> pp(a)

{'___type___': '<Node {id0}>',
 'name': "'A'",
 'refs': [{'___type___': '<Node {id1}>',
           'name': "'B'",
           'refs': [{'___type___': '<Node {id2}>',
                     'name': "'C'",
                     'refs': ['dup <Node {id0}>']}]},
          {'___type___': '<Node {id3}>',
           'name': "'D'",
           'refs': [{'___type___': '<Node {id2}>',
                     'name': "'C'",
                     'refs': ['dup <Node {id0}>']}]}]}

In the example above c is printed in expanded form twice, because both occurrences are found at the same level of recursion. This can make the output quite verbose if the same object is referenced numerous times, so an alternative is to expand it only the first time and emit dup entries subsequently, as shown below.

>>> pp(a, collapse_duplicates=True)

{'___type___': '<Node {id0}>',
 'name': "'A'",
 'refs': [{'___type___': '<Node {id1}>',
           'name': "'B'",
           'refs': [{'___type___': '<Node {id2}>',
                     'name': "'C'",
                     'refs': ['dup <Node {id0}>']}]},
          {'___type___': '<Node {id3}>',
           'name': "'D'",
           'refs': ['dup <Node {id2}>']}]}

Old style classes (Python 2.x only)

>>> class Node():
...     classatt = 'hidden'
...     def __init__(self, name):
...         self.name = name

>>> from printobject import pp
>>> pp(Node)

{'___name___': 'Node',
 '___type___': '<classobj {id0}>',
 '__module__': "'__main__'",
 'classatt': "'hidden'"}

Old style instances (Python 2.x only)

Instances of old style classes don’t look much different from instances of new style classes. The difference is that they identity as instance type, which is visible in the ___type___ value.

>>> a, b, c, d = Node('A'), Node('B'), Node('C'), Node('D')
>>> a.refs = [b, d]
>>> b.refs = [c]
>>> c.refs = [a]
>>> d.refs = [c]

>>> from printobject import pp
>>> pp(a)

{'___type___': '<instance {id0}>',
 '__module__': "'__main__'",
 'classatt': "'hidden'",
 'name': "'A'",
 'refs': [{'___type___': '<instance {id1}>',
           '__module__': "'__main__'",
           'classatt': "'hidden'",
           'name': "'B'",
           'refs': [{'___type___': '<instance {id2}>',
                     '__module__': "'__main__'",
                     'classatt': "'hidden'",
                     'name': "'C'",
                     'refs': ['dup <instance {id0}>']}]},
          {'___type___': '<instance {id3}>',
           '__module__': "'__main__'",
           'classatt': "'hidden'",
           'name': "'D'",
           'refs': [{'___type___': '<instance {id2}>',
                     '__module__': "'__main__'",
                     'classatt': "'hidden'",
                     'name': "'C'",
                     'refs': ['dup <instance {id0}>']}]}]}

In collapsed form:

>>> pp(a, collapse_duplicates=True)

{'___type___': '<instance {id0}>',
 '__module__': "'__main__'",
 'classatt': "'hidden'",
 'name': "'A'",
 'refs': [{'___type___': '<instance {id1}>',
           '__module__': "'__main__'",
           'classatt': "'hidden'",
           'name': "'B'",
           'refs': [{'___type___': '<instance {id2}>',
                     '__module__': "'__main__'",
                     'classatt': "'hidden'",
                     'name': "'C'",
                     'refs': ['dup <instance {id0}>']}]},
        {'___type___': '<instance {id3}>',
         '__module__': "'__main__'",
         'classatt': "'hidden'",
         'name': "'D'",
         'refs': ['dup <instance {id2}>']}]}

Callables

Callables can also be printed, but they are less interesting since they have no public attributes.

Functions:

>>> from printobject import pp
>>> pp(pp)
{'___name___': 'pp', '___type___': '<function {id0}>'}

Methods:

>>> from printobject import Dumper
>>> pp(Dumper.dump)
{'___name___': 'dump', '___type___': '<instancemethod {id0}>'}

Lambdas:

>>> pp(lambda x: x)
{'___name___': '<lambda>', '___type___': '<function {id0}>'}

Iterables

Iterables are printed using their normal __repr__. In this case there are no ___type___ and ___name___ attributes synthesized in the output.

>>> it = frozenset(range(10))

>>> from printobject import pp
>>> pp(it)

['0', '1', '2', '3', '4', '5', '6', '7', '8', '9']

Generators

Generators are a special case of iterables, because the values are created dynamically. Printing generators isn’t insightful without unrolling them, so they will be materialized first. But this means that if the generator is infinite the function will never return.

>>> gen = (x for x in range(10))

>>> from printobject import pp
>>> pp(gen)

['0', '1', '2', '3', '4', '5', '6', '7', '8', '9']