guppy3 3.1.5

Guppy 3 -- Guppy-PE ported to Python 3

Guppy 3

A Python Programming Environment & Heap analysis toolset.

This package contains the following subpackages:

• etc - Support modules. Contains especially the Glue protocol module.
• gsl - The Guppy Specification Language implementation. This can be used to create documents and tests from a common source.
• heapy - The heap analysis toolset. It can be used to find information about the objects in the heap and display the information in various ways.
• sets - Bitsets and 'nodesets' implemented in C.

Guppy 3 is a fork of Guppy-PE, created by Sverker Nilsson for Python 2.

Requirements

You should have Python 3.9, 3.10, 3.11, 3.12, or 3.13. This package is CPython only; PyPy and other Python implementations are not supported. Python 2 support can be obtained from guppy-pe by Sverker Nilsson, from which this package is forked.

To use the graphical browser, Tkinter is needed.

Free-threaded CPython is not supported, due to the excessive complexity and overhead to lock and unlock every object in the interpreter, including objects that are CPython implementation details. (Maybe _PyEval_EnableGILTransient could be a solution to this problem, but it is currently not exposed.)

Installation

Install with pip by:

pip install guppy3

Install with conda by:

conda install -c conda-forge guppy3

Usage

The following example shows

1. How to create the session context: h=hpy()
2. How to show the reachable objects in the heap: h.heap()
3. How to show the shortest paths from the root to the single largest object: h.heap().byid[0].sp
4. How to create and show a set of objects: h.iso(1,[],{})

>>> from guppy import hpy; h=hpy()
>>> h.heap()
Partition of a set of 30976 objects. Total size = 3544220 bytes.
 Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)
     0   8292  27   739022  21    739022  21 str
     1   7834  25   625624  18   1364646  39 tuple
     2   2079   7   300624   8   1665270  47 types.CodeType
     3    400   1   297088   8   1962358  55 type
     4   4168  13   279278   8   2241636  63 bytes
     5   1869   6   269136   8   2510772  71 function
     6    400   1   228464   6   2739236  77 dict of type
     7     79   0   139704   4   2878940  81 dict of module
     8   1061   3    93368   3   2972308  84 types.WrapperDescriptorType
     9    172   1    81712   2   3054020  86 dict (no owner)
<89 more rows. Type e.g. '_.more' to view.>
>>> h.heap().byid[0].sp
 0: h.Root.i0_modules['os'].__dict__
>>> h.iso(1,[],{})
Partition of a set of 3 objects. Total size = 348 bytes.
 Index  Count   %     Size   % Cumulative  % Kind (class / dict of class)
     0      1  33      248  71       248  71 dict (no owner)
     1      1  33       72  21       320  92 list
     2      1  33       28   8       348 100 int
>>>

People have written awesome posts on how to use this toolset, including:

• How to use guppy/heapy for tracking down memory usage
• Debugging Django memory leak with TrackRefs and Guppy
• Diagnosing Memory "Leaks" in Python
• Digging into python memory issues in ckan with heapy
• Optimizing memory usage in Python: a case study
• Memory profiling in python

Formal and API documentation are also available.

Contributing

Issues and pull requests are welcome. You may also ask for help on using this toolset; however, in such cases, we will only provide guidance, and not profile your code for you.

Please make sure to update tests as appropriate.

Testing

To test if the heapy build and installation was ok, you can do:

>>> from guppy import hpy
>>> hpy().test()
Testing sets
Test #0
Test #1
Test #2
...

There will be several more tests. Some tests may take a while.

License

Copyright (C) 2005-2013 Sverker Nilsson, S. Nilsson Computer System AB
Copyright (C) 2019-2021 YiFei Zhu
Copyright (C) 2021-2025 YiFei Zhu, Google LLC

The right is granted to copy, use, modify and redistribute this code according to the rules in what is commonly referred to as an MIT license.

This is not an official Google product.

*** USE AT YOUR OWN RISK AND BE AWARE THAT THIS IS AN EARLY RELEASE ***