Dynamically import modules for use in lambda expression

Question

I need to dynamically create a function from a lambda expression that is provided as a string (it will be read from a configuration file). I want to allow the user to also specify a list of modules that can then be used within the lambda expression. For this I need to dynamically load the modules and make them available so that they can be referred to by name within the lambda expression.

The following is a possible implementation of what I want to do:

import importlib

def create_function(lambda_expression, modules=[]):
    for module in modules:
        globals()[module] = importlib.import_module(module)
    function = eval('lambda ' + lambda_expression)
    return function

It would be used like this:

f = create_function('x: numpy.clip(x, 0, 1)', ['numpy'])

However, the use of globals() does not seem like a nice solution. Is there a more elegant way to achieve this?

can you setup your configuration file differently, like as actual code, and then import the module? — Chris_Rands
– Chris_Rands, Commented Sep 20, 2019 at 13:57

soulmerge · Accepted Answer · 2019-09-20 14:39:12Z

1

Python's eval function accepts a globals argument. So you can do something like this:

function = eval('lambda ' + lambda_expression, lambda_globals)

answered Sep 20, 2019 at 14:39

soulmerge

76.2k20 gold badges121 silver badges160 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Emil Over a year ago

This works well and looks like a perfect solution for my task. Thanks.

Collectives™ on Stack Overflow

Dynamically import modules for use in lambda expression

1 Answer 1

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Related