How do I get a list of Python modules installed on my computer?
32 Answers
Now, these methods I tried myself, and I got exactly what was advertised: All the modules.
Alas, really you don't care much about the stdlib, you know what you get with a python install.
Really, I want the stuff that I installed.
What actually, surprisingly, worked just fine was:
pip freeze
Which returned:
Fabric==0.9.3
apache-libcloud==0.4.0
bzr==2.3b4
distribute==0.6.14
docutils==0.7
greenlet==0.3.1
ipython==0.10.1
iterpipes==0.4
libxml2-python==2.6.21
I say "surprisingly" because the package install tool is the exact place one would expect to find this functionality, although not under the name 'freeze' but python packaging is so weird, that I am flabbergasted that this tool makes sense. Pip 0.8.2, Python 2.7.
As of pip 10, the accepted answer will no longer work. The development team has removed access to the get_installed_distributions routine. There is an alternate function in the setuptools for doing the same thing. Here is an alternate version that works with pip 10:
import pkg_resources
installed_packages = pkg_resources.working_set
installed_packages_list = sorted(["%s==%s" % (i.key, i.version)
for i in installed_packages])
print(installed_packages_list)
Please let me know if it will or won't work in previous versions of pip, too.
Works Regardless of Pip Version
Run the following in your python editor or IPython:
import pkg_resources
installed_packages = {d.project_name: d.version for d in pkg_resources.working_set}
print(installed_packages)
Read other answers and pulled together this combo, which is quickest and easiest inside Python.
Find the specific Packages
Conveniently you can then get items from your dict easily, i.e.
installed_packages['pandas'] >> '1.16.4'
Using Pip List Well
!pip list will run inside your jupyter notebook if working there, simplifying the 'quick check'
Combine with other utilities like grep(if you have installed)
pip list | grep pandas will get you your current pandas version for example
I normally use pip list to get a list of packages (with version).
This works in a virtual environment too, of course. To show what's installed in only the virtual environment (not global packages), use pip list --local.
Here's documentation showing all the available pip list options, with several good examples.
This will help
In terminal or IPython, type:
help('modules')
then
In [1]: import #import press-TAB
Display all 631 possibilities? (y or n)
ANSI audiodev markupbase
AptUrl audioop markupsafe
ArgImagePlugin avahi marshal
BaseHTTPServer axi math
Bastion base64 md5
BdfFontFile bdb mhlib
BmpImagePlugin binascii mimetools
BufrStubImagePlugin binhex mimetypes
CDDB bisect mimify
CDROM bonobo mmap
CGIHTTPServer brlapi mmkeys
Canvas bsddb modulefinder
CommandNotFound butterfly multifile
ConfigParser bz2 multiprocessing
ContainerIO cPickle musicbrainz2
Cookie cProfile mutagen
Crypto cStringIO mutex
CurImagePlugin cairo mx
DLFCN calendar netrc
DcxImagePlugin cdrom new
Dialog cgi nis
DiscID cgitb nntplib
DistUpgrade checkbox ntpath
Warning: Adam Matan discourages this use in pip > 10.0. Also, read @sinoroc's comment below
This was inspired by Adam Matan's answer (the accepted one):
import tabulate
try:
from pip import get_installed_distributions
except:
from pip._internal.utils.misc import get_installed_distributions
tabpackages = []
for _, package in sorted([('%s %s' % (i.location, i.key), i) for i in get_installed_distributions()]):
tabpackages.append([package.location, package.key, package.version])
print(tabulate.tabulate(tabpackages))
which then prints out a table in the form of
19:33 pi@rpi-v3 [iot-wifi-2] ~/python$ python installed_packages.py
------------------------------------------- -------------- ------
/home/pi/.local/lib/python2.7/site-packages enum-compat 0.0.2
/home/pi/.local/lib/python2.7/site-packages enum34 1.1.6
/home/pi/.local/lib/python2.7/site-packages pexpect 4.2.1
/home/pi/.local/lib/python2.7/site-packages ptyprocess 0.5.2
/home/pi/.local/lib/python2.7/site-packages pygatt 3.2.0
/home/pi/.local/lib/python2.7/site-packages pyserial 3.4
/usr/local/lib/python2.7/dist-packages bluepy 1.1.1
/usr/local/lib/python2.7/dist-packages click 6.7
/usr/local/lib/python2.7/dist-packages click-datetime 0.2
/usr/local/lib/python2.7/dist-packages construct 2.8.21
/usr/local/lib/python2.7/dist-packages pyaudio 0.2.11
/usr/local/lib/python2.7/dist-packages tabulate 0.8.2
------------------------------------------- -------------- ------
which lets you then easily discern which packages you installed with and without sudo.
A note aside: I've noticed that when I install a packet once via sudo and once without, one takes precedence so that the other one isn't being listed (only one location is shown). I believe that only the one in the local directory is then listed. This could be improved.
In case you have an anaconda python distribution installed, you could also use
$conda list
in addition to solutions described above.
There are many way to skin a cat.
The most simple way is to use the
pydocfunction directly from the shell with:
pydoc modulesBut for more information use the tool called pip-date that also tell you the installation dates.
pip install pip-date
There are many ideas, initially I am pondering on these two:
pip
cons: not always installed
help('modules')
cons: output to console; with broken modules (see ubuntu...) can segfault
I need an easy approach, using basic libraries and compatible with old python 2.x
And I see the light: listmodules.py
Hidden in the documentation source directory in 2.5 is a small script that lists all available modules for a Python installation.
Pros:
uses only imp, sys, os, re, time
designed to run on Python 1.5.2 and newer
the source code is really compact, so you can easy tinkering with it, for example to pass an exception list of buggy modules (don't try to import them)
I needed to find the specific version of packages available by default in AWS Lambda. I did so with a mashup of ideas from this page. I'm sharing it for posterity.
import pkgutil
__version__ = '0.1.1'
def get_ver(name):
try:
return str(__import__(name).__version__)
except:
return None
def lambda_handler(event, context):
return {
'statusCode': 200,
'body': [{
'path': m.module_finder.path,
'name': m.name,
'version': get_ver(m.name),
} for m in list(pkgutil.iter_modules())
#if m.module_finder.path == "/var/runtime" # Uncomment this if you only care about a certain path
],
}
What I discovered is that the provided boto3 library was way out of date and it wasn't my fault that my code was failing. I just needed to add boto3 and botocore to my project. But without this I would have been banging my head thinking my code was bad.
{
"statusCode": 200,
"body": [
{
"path": "/var/task",
"name": "lambda_function",
"version": "0.1.1"
},
{
"path": "/var/runtime",
"name": "bootstrap",
"version": null
},
{
"path": "/var/runtime",
"name": "boto3",
"version": "1.9.42"
},
{
"path": "/var/runtime",
"name": "botocore",
"version": "1.12.42"
},
{
"path": "/var/runtime",
"name": "dateutil",
"version": "2.7.5"
},
{
"path": "/var/runtime",
"name": "docutils",
"version": "0.14"
},
{
"path": "/var/runtime",
"name": "jmespath",
"version": "0.9.3"
},
{
"path": "/var/runtime",
"name": "lambda_runtime_client",
"version": null
},
{
"path": "/var/runtime",
"name": "lambda_runtime_exception",
"version": null
},
{
"path": "/var/runtime",
"name": "lambda_runtime_marshaller",
"version": null
},
{
"path": "/var/runtime",
"name": "s3transfer",
"version": "0.1.13"
},
{
"path": "/var/runtime",
"name": "six",
"version": "1.11.0"
},
{
"path": "/var/runtime",
"name": "test_bootstrap",
"version": null
},
{
"path": "/var/runtime",
"name": "test_lambda_runtime_client",
"version": null
},
{
"path": "/var/runtime",
"name": "test_lambda_runtime_marshaller",
"version": null
},
{
"path": "/var/runtime",
"name": "urllib3",
"version": "1.24.1"
},
{
"path": "/var/lang/lib/python3.7",
"name": "__future__",
"version": null
},
...
What I discovered was also different from what they officially publish. At the time of writing this:
- Operating system – Amazon Linux
- AMI – amzn-ami-hvm-2017.03.1.20170812-x86_64-gp2
- Linux kernel – 4.14.77-70.59.amzn1.x86_64
- AWS SDK for JavaScript – 2.290.0\
- SDK for Python (Boto 3) – 3-1.7.74 botocore-1.10.74
Here is a python code solution that will return a list of modules installed. One can easily modify the code to include version numbers.
import subprocess
import sys
from pprint import pprint
installed_packages = reqs = subprocess.check_output([sys.executable, '-m', 'pip', 'freeze']).decode('utf-8')
installed_packages = installed_packages.split('\r\n')
installed_packages = [pkg.split('==')[0] for pkg in installed_packages if pkg != '']
pprint(installed_packages)
Installation
pip install pkgutil
Code
import pkgutil
for i in pkgutil.iter_modules(None): # returns a tuple (path, package_name, ispkg_flag)
print(i[1]) #or you can append it to a list
Sample Output:
multiprocessing
netrc
nntplib
ntpath
nturl2path
numbers
opcode
pickle
pickletools
pipes
pkgutil
If none of the above seem to help, in my environment was broken from a system upgrade and I could not upgrade pip. While it won't give you an accurate list you can get an idea of which libraries were installed simply by looking inside your env>lib>python(version here)>site-packages> . Here you will get a good indication of modules installed.
pip3 freeze
run this command globally - will show you all the global installations run this command on the virtual environment - will show you all the local installations
For anyone wondering how to call pip list from a Python program you can use the following:
import pip
pip.main(['list]) # this will print all the packages
From the shell
ls site-packages
If that's not helpful, you can do this.
import sys
import os
for p in sys.path:
print os.listdir( p )
And see what that produces.
Related