Test for Python module dependencies being installed

Question

How could one test whether a set of modules is installed, given the names of the modules. E.g.

modules = set(["sys", "os", "jinja"])

for module in modules:
  # if test(module exists):
  #   do something

While it's possible to write out the tests as:

try:
  import sys
except ImportError:
  print "No sys!"

This is a bit cumbersome for what I'm doing. Is there a dynamic way to do this?

I've tried eval("import %s"%module) but that complained of a compile error.

I'm grateful for your thoughts and suggestions. Thank you.

Rick Copeland · Accepted Answer · 2009-04-20 14:43:12Z

8

You can use the __import__() function like this::

for module in modules:
    try:
        __import__(module)
    except ImportError:
        do_something()

You can also use imp.find_module to determine whether a module can be found without importing it::

import imp
for module in modules:
    try:
        imp.find_module(module)
    except ImportError:
        do_something()

Oh, and the reason you can't use eval() is because import is a statement, not an expression. You can use exec if you really want to, though I wouldn't recommend it::

for module in modules:
    try:
        exec 'import ' + module
    except ImportError:
        do_something()

answered Apr 20, 2009 at 14:43

Rick Copeland

12k5 gold badges41 silver badges38 bronze badges

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

Comments

Paolo Bergantino · Accepted Answer · 2009-04-20 14:40:41Z

6

What's wrong with this?

modules = set(["sys", "os", "jinja"])
for m in modules:
    try:
        __import__(m)
    except ImportError:
        # do something

The above uses the __import__ function. Also, it is strictly testing whether it exists or not - its not actually saving the import anywhere, but you could easily modify it to do that.

edited Apr 20, 2009 at 14:40

answered Apr 20, 2009 at 14:31

Paolo Bergantino

490k83 gold badges523 silver badges437 bronze badges

Comments

slinkp · Accepted Answer · 2010-07-14 16:05:14Z

It's worth understanding the tradeoffs between the find_module approach and the __import__ approach:

__import__('foo') will trigger any side effects of loading the module, regardless of whether you save a reference to the imported module. That might be OK, or not.
find_module('foo') won't trigger any side effects, so it should always be safe.
neither approach works well if the module is found but has other problems at load time (eg. SyntaxError, NameError, ...)
Depending on how you handle failures, simply wrapping __import__('foo') in a try/except will give you misleading results if foo is found but foo imports bar which is not found.
find_module() can't handle dotted names, so if you need to check whether 'foo.bar.bat.baz' can be imported, you have to use multiple calls to find_module and load_module as per the docs.

vartec · Accepted Answer · 2009-04-20 14:43:22Z

2

modules = set(["sys", "os", "jinja"])
try:
  for module in modules:
    __import__(module)
  doSomething()
except ImportError, e:
    print e

answered Apr 20, 2009 at 14:43

vartec

135k38 gold badges227 silver badges248 bronze badges

Collectives™ on Stack Overflow

Test for Python module dependencies being installed

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related