Installing python module within code

Question:

I need to install a package from PyPi straight within my script.
Maybe there’s some module or distutils (distribute, pip etc.) feature which allows me to just execute something like pypi.install('requests') and requests will be installed into my virtualenv.

Asked By: chuwy

||

Answers:

You define the dependent module inside the setup.py of your own package with the “install_requires” option.

If your package needs to have some console script generated then you can use the “console_scripts” entry point in order to generate a wrapper script that will be placed
within the ‘bin’ folder (e.g. of your virtualenv environment).

Answered By: Andreas Jung

This should work:

import subprocess

def install(name):
    subprocess.call(['pip', 'install', name])
Answered By: quantum

You can also use something like:

import pip

def install(package):
    if hasattr(pip, 'main'):
        pip.main(['install', package])
    else:
        pip._internal.main(['install', package])

# Example
if __name__ == '__main__':
    install('argh')
Answered By: Rikard Anglerud

If you want to use pip to install required package and import it after installation, you can use this code:

def install_and_import(package):
    import importlib
    try:
        importlib.import_module(package)
    except ImportError:
        import pip
        pip.main(['install', package])
    finally:
        globals()[package] = importlib.import_module(package)


install_and_import('transliterate')

If you installed a package as a user you can encounter the problem that you cannot just import the package. See How to refresh sys.path? for additional information.

Answered By: rominf

The officially recommended way to install packages from a script is by calling pip’s command-line interface via a subprocess. Most other answers presented here are not supported by pip. Furthermore since pip v10, all code has been moved to pip._internal precisely in order to make it clear to users that programmatic use of pip is not allowed.

Use sys.executable to ensure that you will call the same pip associated with the current runtime.

import subprocess
import sys

def install(package):
    subprocess.check_call([sys.executable, "-m", "pip", "install", package])
Answered By: Aaron de Windt

i added some exception handling to @Aaron’s answer.

import subprocess
import sys

try:
    import pandas as pd
except ImportError:
    subprocess.check_call([sys.executable, "-m", "pip", "install", 'pandas'])
finally:
    import pandas as pd
Answered By: Sohan Das

For installing multiple packages, I am using a setup.py file with the following code:

import sys
import subprocess
import pkg_resources

required  = {'numpy', 'pandas', '<etc>'} 
installed = {pkg.key for pkg in pkg_resources.working_set}
missing   = required - installed

if missing:
    # implement pip as a subprocess:
    subprocess.check_call([sys.executable, '-m', 'pip', 'install', *missing])
Answered By: Tanmay Shrivastava
import os
os.system('pip install requests')

I tried above for temporary solution instead of changing docker file. Hope these might be useful to some

Answered By: Janarthanan Ramu

Try the below. So far the best that worked for me
Install the 4 ones first and then Mention the new ones in the REQUIRED list

import pkg_resources
import subprocess
import sys
import os

REQUIRED = {
  'spacy', 'scikit-learn', 'numpy', 'pandas', 'torch', 
  'pyfunctional', 'textblob', 'seaborn', 'matplotlib'
}

installed = {pkg.key for pkg in pkg_resources.working_set}
missing = REQUIRED - installed

if missing:
    python = sys.executable
    subprocess.check_call([python, '-m', 'pip', 'install', *missing], stdout=subprocess.DEVNULL)
Answered By: Rijin

If you want a more efficient answer that expands on subprocess.check_call. You can first check if the requirement has already been met using pkg_resources.

This works for different requirment specifiers which is nice. e.g. >=, ==

import sys
import subprocess
import pkg_resources
from pkg_resources import DistributionNotFound, VersionConflict

def should_install_requirement(requirement):
    should_install = False
    try:
        pkg_resources.require(requirement)
    except (DistributionNotFound, VersionConflict):
        should_install = True
    return should_install


def install_packages(requirement_list):
    try:
        requirements = [
            requirement
            for requirement in requirement_list
            if should_install_requirement(requirement)
        ]
        if len(requirements) > 0:
            subprocess.check_call([sys.executable, "-m", "pip", "install", *requirements])
        else:
            print("Requirements already satisfied.")

    except Exception as e:
        print(e)

Example usage:

requirement_list = ['requests', 'httpx==0.18.2']
install_packages(requirement_list)

demo

Relevant Info Stackoverflow Question: 58612272

Answered By: Glen Thompson

To conditionally install multiple packages with exact version, I’ve been using this pattern basing on @Tanmay Shrivastava’s answer:

import sys
from subprocess import run, PIPE, STDOUT
import pkg_resources

def run_cmd(cmd):
    ps = run(cmd, stdout=PIPE, stderr=STDOUT, shell=True, text=True)
    print(ps.stdout)


# packages to be conditionally installed with exact version
required = {"click==8.0.1", "semver==3.0.0.dev2"}
installed = {f"{pkg.key}=={pkg.version}" for pkg in pkg_resources.working_set}
missing = required - installed

if missing:
    run_cmd(f'pip install --ignore-installed {" ".join([*missing])}')
Answered By: Glenn Mohammad
Categories: questions Tags: , , ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.