Format number using LaTeX notation in Python

Question

Using format strings in Python I can easily print a number in \"scientific notation\", e.g.

>> print \'%g\'%1e9
1e+09

What is the simples

Answer 1

You can write a frexp10 function:

def frexp10(x):
    exp = int(math.floor(math.log10(abs(x))))
    return x / 10**exp, exp

Formatting in LaTeX style is then:

'{0}^{{{1:+03}}}'.format(*frexp10(-1.234e9))

Answer 2

The siunitx LaTeX package solves this for you by allowing you to use the python float value directly without resorting to parsing the resulting string and turning it into valid LaTeX.

>>> print "\\num{{{0:.2g}}}".format(1e9)
\num{1e+09}

When the LaTeX document is compiled, the above code will be turned into . As andybuckley points out in the comments, the plus sign might not be accepted by siunitx (I've not tested it), so it may be necessary to do a .repace("+", "") on the result.

If using siunitx is somehow off the table, write a custom function like this:

def latex_float(f):
    float_str = "{0:.2g}".format(f)
    if "e" in float_str:
        base, exponent = float_str.split("e")
        return r"{0} \times 10^{{{1}}}".format(base, int(exponent))
    else:
        return float_str

Testing:

>>> latex_float(1e9)
'1 \\times 10^{9}'

Answer 3

Install num2tex:

pip install num2tex

and use it as so:

>>> from num2tex import num2tex
>>> '{:.0e}'.format(num2tex(1e9))
'1 \\times 10^{9}'

num2tex inherits from str so the format function can be used in the same way.

You can also change the format of the exponent by using num2tex.configure() (adding this in response to @Matt's comment).

>>>from num2tex import num2tex
>>>from num2tex import configure as num2tex_configure
>>>num2tex_configure(exp_format='cdot')
>>>num2tex(1.3489e17)
'1.3489 \cdot 10^{17}'
>>>num2tex_configure(exp_format='parentheses')
'1.3489 (10^{17})'

As of now this is undocumented in the GitHub, I'll try to change this soon!

Disclaimer: After using (and upvoting) Lauritz V. Thaulow's answer for a while (for Jupyter, Matplotlib etc.) I thought it would be better for my workflow to write a simple Python module, so I created num2tex on GitHub and registered it on PyPI. I would love to get some feedback on how to make it more useful.