bitmath simplifies many facets of interacting with file sizes in various units. Originally focusing on file size unit conversion, functionality now includes:

  • Converting between SI and NIST prefix units (kB to GiB)
  • Converting between units of the same type (SI to SI, or NIST to NIST)
  • Automatic human-readable prefix selection (like in hurry.filesize)
  • Basic arithmetic operations (subtracting 42KiB from 50GiB)
  • Rich comparison operations (1024 Bytes == 1KiB)
  • bitwise operations (<<, >>, &, |, ^)
  • Reading a device’s storage capacity (Linux/OS X support only)
  • argparse integration as a custom type
  • click integration as a custom parameter type
  • progressbar integration as a better file transfer speed widget
  • String parsing
  • Sorting

In addition to the conversion and math operations, bitmath provides human readable representations of values which are suitable for use in interactive shells as well as larger scripts and applications. The format produced for these representations is customizable via the functionality included in stdlibs string.format.

In discussion we will refer to the NIST units primarily. I.e., instead of “megabyte” we will refer to “mebibyte”. The former is 10^3 = 1,000,000 bytes, whereas the second is 2^20 = 1,048,576 bytes. When you see file sizes or transfer rates in your web browser, most of the time what you’re really seeing are the base-2 sizes/rates.

Don’t Forget! The source for bitmath is available on GitHub.

And did we mention there’s almost 200 unittests? Check them out for yourself.


The easiest way to install bitmath is via dnf (or yum) if you’re on a Fedora/RHEL based distribution. bitmath is available in the main Fedora repositories, as well as the EPEL6 and EPEL7 repositories. There are now dual python2.x and python3.x releases available.

Python 2.x:

$ sudo dnf install python2-bitmath

Python 3.x:

$ sudo dnf install python3-bitmath


Upgrading: If you have the old python-bitmath package installed presently, you could also run sudo dnf update python-bitmath instead


You could also install bitmath from PyPi if you like:

$ sudo pip install bitmath


pip installs need pip >= 1.1. To workaround this, download bitmath, from PyPi and then pip install bitmath-x.y.z.tar.gz. See issue #57 for more information.


Ubuntu Xenial, Wily, Vivid, Trusty, and Precise users can install bitmath from the launchpad PPA:

$ sudo add-apt-repository ppa:tbielawa/bitmath
$ sudo apt-get update
$ sudo apt-get install python-bitmath


Or, if you want to install from source:

$ sudo python ./ install

If you want the bitmath manpage installed as well:

$ sudo make install




>>> import bitmath
>>> log_size = bitmath.kB(137.4)
>>> log_zipped_size = bitmath.Byte(987)
>>> print "Compression saved %s space" % (log_size - log_zipped_size)
Compression saved 136.413kB space
>>> thumb_drive = bitmath.GiB(12)
>>> song_size = bitmath.MiB(5)
>>> songs_per_drive = thumb_drive / song_size
>>> print songs_per_drive

Convert Units

File size unit conversion:

>>> from bitmath import *
>>> dvd_size = GiB(4.7)
>>> print "DVD Size in MiB: %s" % dvd_size.to_MiB()
DVD Size in MiB: 4812.8 MiB

Select a human-readable unit

>>> small_number = kB(100)
>>> ugly_number = small_number.to_TiB()

>>> print ugly_number
9.09494701773e-08 TiB
>>> print ugly_number.best_prefix()
97.65625 KiB

Rich Comparison

>>> cd_size = MiB(700)
>>> cd_size > dvd_size
>>> cd_size < dvd_size
>>> MiB(1) == KiB(1024)
>>> MiB(1) <= KiB(1024)


>>> sizes = [KiB(7337.0), KiB(1441.0), KiB(2126.0), KiB(2178.0),
                  KiB(2326.0), KiB(4003.0), KiB(48.0), KiB(1770.0),
                  KiB(7892.0), KiB(4190.0)]

>>> print sorted(sizes)
[KiB(48.0), KiB(1441.0), KiB(1770.0), KiB(2126.0), KiB(2178.0),
KiB(2326.0), KiB(4003.0), KiB(4190.0), KiB(7337.0), KiB(7892.0)]

Custom Formatting

  • Use of the custom formatting system
  • All of the available instance properties


>>> longer_format = """Formatting attributes for %s
   ...: This instances prefix unit is {unit}, which is a {system} type unit
   ...: The unit value is {value}
   ...: This value can be truncated to just 1 digit of precision: {value:.1f}
   ...: In binary this looks like: {binary}
   ...: The prefix unit is derived from a base of {base}
   ...: Which is raised to the power {power}
   ...: There are {bytes} bytes in this instance
   ...: The instance is {bits} bits large
   ...: bytes/bits without trailing decimals: {bytes:.0f}/{bits:.0f}""" % str(ugly_number)

>>> print ugly_number.format(longer_format)
Formatting attributes for 5.96046447754 MiB
This instances prefix unit is MiB, which is a NIST type unit
The unit value is 5.96046447754
This value can be truncated to just 1 digit of precision: 6.0
In binary this looks like: 0b10111110101111000010000000
The prefix unit is derived from a base of 2
Which is raised to the power 20
There are 6250000.0 bytes in this instance
The instance is 50000000.0 bits large
bytes/bits without trailing decimals: 6250000/50000000

Utility Functions


>>> print bitmath.getsize('python-bitmath.spec')
3.7060546875 KiB


Parse a string with standard units:

>>> import bitmath
>>> a_dvd = bitmath.parse_string("4.7 GiB")
>>> print type(a_dvd)
<class 'bitmath.GiB'>
>>> print a_dvd
4.7 GiB


Parse a string with ambiguous units:

>>> import bitmath
>>> a_gig = bitmath.parse_string_unsafe("1gb")
>>> print type(a_gig)
<class 'bitmath.GB'>
>>> a_gig == bitmath.GB(1)
>>> bitmath.parse_string_unsafe('1gb') == bitmath.parse_string_unsafe('1g')


>>> import bitmath
>>> with open('/dev/sda') as fp:
...     root_disk = bitmath.query_device_capacity(fp)
...     print root_disk.best_prefix()
238.474937439 GiB


>>> for i in bitmath.listdir('./tests/', followlinks=True, relpath=True, bestprefix=True):
...     print i
('tests/', KiB(9.2900390625))
('tests/', KiB(7.1767578125))
('tests/', KiB(1.974609375))
('tests/', KiB(2.6376953125))
('tests/', KiB(3.7744140625))
('tests/', KiB(5.2568359375))
('tests/', KiB(2.03125))
('tests/', KiB(3.4580078125))
('tests/', KiB(2.2001953125))
('tests/', KiB(2.1044921875))
('tests/', KiB(3.9423828125))
('tests/', KiB(5.431640625))
('tests/', Byte(311.0))
('tests/.coverage', KiB(3.1708984375))
('tests/', KiB(5.34375))
('tests/', KiB(1.798828125))
('tests/', KiB(8.0185546875))
('tests/', KiB(4.2197265625))
('tests/listdir_symlinks/10_byte_file_link', Byte(10.0))
('tests/listdir_symlinks/depth1/depth2/10_byte_file', Byte(10.0))
('tests/listdir_nosymlinks/depth1/depth2/10_byte_file', Byte(10.0))
('tests/listdir_nosymlinks/depth1/depth2/1024_byte_file', KiB(1.0))
('tests/file_sizes/kbytes.test', KiB(1.0))
('tests/file_sizes/bytes.test', Byte(38.0))
('tests/listdir/10_byte_file', Byte(10.0))


>>> with bitmath.format(fmt_str="[{value:.3f}@{unit}]"):
...     for i in bitmath.listdir('./tests/', followlinks=True, relpath=True, bestprefix=True):
...         print i[1]

argparse Integration

Example script using bitmath.integrations.bmargparse.BitmathType as an argparser argument type:

import argparse
from bitmath.integrations.bmargparse import BitmathType
parser = argparse.ArgumentParser(
    description="Arg parser with a bitmath type argument")

results = parser.parse_args()
print "Parsed in: {PARSED}; Which looks like {TOKIB} as a Kibibit".format(

If ran as a script the results would be similar to this:

$ python ./ --block-size 100MiB
Parsed in: 100.0 MiB; Which looks like 819200.0 Kib as a Kibibit

click Integration

Example script using bitmath.integrations.bmclick.BitmathType as an click parameter type:

import click
from bitmath.integrations.bmclick import BitmathType

@click.argument('size', type=BitmathType())
def best_prefix(size):

If ran as a script the results should be similar to this:

$ python ./ "1024 KiB"
1.0 MiB

progressbar Integration

Use bitmath.integrations.bmprogressbar.BitmathFileTransferSpeed as a progressbar file transfer speed widget to monitor download speeds:

import requests
import progressbar
import bitmath
from bitmath.integrations.bmprogressbar import BitmathFileTransferSpeed

FETCH = ''
widgets = ['Bitmath Progress Bar Demo: ', ' ',
           progressbar.Bar(marker=progressbar.RotatingMarker()), ' ',

r = requests.get(FETCH, stream=True)
size = bitmath.Byte(int(r.headers['Content-Length']))
pbar = progressbar.ProgressBar(widgets=widgets, maxval=int(size),
chunk_size = 2048
with open('/dev/null', 'wb') as fd:
    for chunk in r.iter_content(chunk_size):
        if (pbar.currval + chunk_size) < pbar.maxval:
            pbar.update(pbar.currval + chunk_size)

If ran as a script the results would be similar to this:

$ python ./
Bitmath Progress Bar Demo:  ||||||||||||||||||||||||||||||||||||||||| 1.58 MiB/s