1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
|
memprof
=======
``memprof`` is a memory profiler for Python.
It logs and plots the memory usage of all the variables during the
execution of the decorated methods.
Installation
------------
Stable
~~~~~~
::
sudo pip install --upgrade memprof
or ::
sudo easy_install --upgrade memprof
or (Debian testing/unstable) ::
sudo apt-get install python-memprof
Development
~~~~~~~~~~~
::
git clone git://github.com/jmdana/memprof.git
cd memprof
sudo python setup.py install
or ::
sudo pip install git+https://github.com/jmdana/memprof
Usage
-----
Using ``memprof`` is as easy as adding a decorator to the methods that
you want to profile: ::
@memprof
def foo():
And importing the module just by including the line below at the
beginning of your Python file: ::
from memprof import memprof
Now you can run as usual and logfiles with the names of your methods
will be created (e.g. ``foo.log``).
Generating plots
~~~~~~~~~~~~~~~~
The logfiles are not very interesting so you might prefer to use the
``-p``/``--plot`` flag: ::
python -m memprof --plot <python_file>
python -m memprof -p <python_file>
Which, in addition to the logfile, will generate a plot (``foo.png``):
.. figure:: examples/foo.png
:alt: Example plot
The grey bar indicates that the ``foo`` method wasn't running at that
point.
The flag may also be passed as an argument to the decorator: ::
@memprof(plot = True)
Please keep in mind that the former takes precedence over the latter.
Adjusting the threshold
~~~~~~~~~~~~~~~~~~~~~~~
You may also want to specify a ``threshold``. The value will be the
minimum size for a variable to appear in the plot (but it will always
appear in the logfile!). The default value is 1048576 (1 MB) but you can
specify a different ``threshold`` (in bytes) with the
``-t``/``--threshold`` flag: ::
python -m memprof --threshold 1024 <python_file>
python -m memprof -t 1024 <python_file>
The ``threshold`` may also be passed as an argument to the decorator: ::
@memprof(threshold = 1024)
Please keep in mind that the former takes precedence over the latter.
mp\_plot
~~~~~~~~
If, after running ``memprof``, you want to change the threshold and
generate a new plot (or you forgot to use the ``-p``/``--plot`` flag
with ``memprof``), you don't have to re-run! Just call the command: ::
mp_plot [-h] [-t THRESHOLD] logfiles [logfiles ...]
and generate the plots again doing something like: ::
mp_plot -t 128 logfile1.log logfile2.log
or: ::
mp_plot -t 1024 *.log
etc.
Contact
-------
Mailing list
~~~~~~~~~~~~
- Subscribe by sending a message to memprof+subscribe@googlegroups.com
- Once subscribed, you can send emails to memprof@googlegroups.com
- List archives at http://groups.google.com/group/memprof
--------------
Copyright 2013-2019, Jose M. Dana
|
