pyPhix, the Python Fixed Point module

pyPhix is a python package for fixed point number representation. It is intended to support the implementation of digital signal processing systems. As such only fix-point addition, subtraction and multiplication operations are implemented.

You can find more information about the module objects and functionalities at readthedocs project page.

Features

virtually unlimited number length (depending on your RAM size)
based on NumPy
customizable rounding method (SymInf, SymZero, NonSymPos, NonSymNeg, ConvEven, ConvOdd, Floor, Ceil)
customizable wrapping method (Sat, Wrap)
support various representation formats (bin, hex, int, float)
perform single or array based operations with customizable output format (+, -, *)

License

pyPhix

pyPhix is an open source python module released under the terms of Mozilla Public License Version 2.0.

NumPy

NumPy is the fundamental package needed for scientific computing with Python and it is released under these terms.

Install

The pyPhix package is available on pypi.org. You can install it by running:

$ pip install pyphix.

Alternatively you can clone the pyPhix repository and from the folder containing the setup.py file run:

$ python setup install

This package requires Python 3.6 to work.

You can also directly download the tar.gz archive from pypi.org. The archive can be easily verified by adding the gpg public key 1E948096166391C0 to your keyring.

Usage Examples

Fix Format

This object is used to indicate the number of bits the user wants to use for value reprensetation.

Create fix-point format objets:

>>> from pyphix import fix
>>> fmt_a=fix.FixFmt(True, 2, 10)
>>> fmt_b=fix.FixFmt(False, 0, 7)

Print the maximum representable ranges and test if a value is included in the range:

>>> fmt_a.fixrange
(-4.0, 3.9990234375)

>>> fmt_b.fixrange
(0.0, 0.9921875)

>>> -10 in fmt_a
False

>>> 0 in fmt_b
True

Different representations are available:

>>> fmt_a.tuplefmt
(True, 2, 10)

>>> fmt_b.listfmt
[False, 0, 7]

>>> print(fmt_b)
(False, 0, 7)

Fix Number

This object contain fix-point number represenation.

Round methods comparison, assuming format (True, 4, 5):

Round method	Pos odd fraction	Pos even fraction	Neg odd fraction	Neg even fraction
Real value	7.296875	2.325	-1.078125	-1.08125
Mult by 2^5	233.5	74.4	-34.5	-34.6
`SymInf`	7.3125	2.3125	-1.09375	-1.09375
`SymZero`	7.28125	2.3125	-1.0625	-1.09375
`NonSymPos`	7.3125	2.3125	-1.0625	-1.09375
`NonSymNeg`	7.28125	2.3125	-1.09375	-1.09375
`ConvEven`	7.3125	2.3125	-1.0625	-1.09375
`ConvOdd`	7.28125	2.3125	-1.09375	-1.09375
`Floor`	7.28125	2.3125	-1.09375	-1.09375
`Ceil`	7.3125	2.34375	-1.0625	-1.0625

A small usage example:

>>> from pyphix import fix
>>> from pyphix.fix import ERoundMethod, EOverMethod
>>> fmt = fix.FixFmt(True, 4, 5)
>>> fix_vec = fix.FixNum(
        [7.296875,  2.325   , -1.078125, -1.08125], fmt,
        rnd=ERoundMethod.CONV_ODD, over=EOverMethod.WRAP)
>>> fix_val = fix.FixNum(
        1.16, fmt,
        rnd=ERoundMethod.CONV_ODD, over=EOverMethod.WRAP)

Perform a full resolution addition:

>>> fix_vec + fix_val
[8.4375  3.46875 0.0625  0.0625 ]

  fmt: (True, 5, 5)
  rnd: ERoundMethod.CONV_ODD
  over: EOverMethod.WRAP

Perform a multiplication and cast result to a small format:

>>> fix_val.mult(
        fix_vec, out_fmt=fix.FixFmt(False, 3, 2),
        rnd=ERoundMethod.SYM_INF, over=EOverMethod.SAT)
[7.75 2.75 0.   0.  ]

  fmt: (False, 3, 2)
  rnd: ERoundMethod.SYM_INF
  over: EOverMethod.SAT

sfavazza/pyphix