WebImport

Import python libraries over HTTP. This package allows you to load python modules on a system where these packages may not be available. It uses a HTTP connection to fetch these packages remotely.

This package is written for Python3.4 or newer.

Use cases

• Running dynamic code from a read-only filesystem
• Using non-standard packages on a machine without leaving traces
• Testing code on devices you cannot install packages on

How to use

If WebImporter is available on a given system, you can import it as shown below:

    import webimport
    webimport.register(location='some.http.server', port=8080, override=False)

The register command takes 3 arguments.

• location (optional): The host to connect to. Default: localhost
• port (optional): The port to connect to. Default: 8080
• override: (optional): Try to import from remote first. Default: False

Preparing the server

To create the webserver to serve the libraries, you should create a single directory containing all these packages.

To do this, an easy method to do this would be:

mkdir my_package_dir
pip install <package> --target my_package_dir

Example uses:

Running scripts with shared modules over ssh:

First Terminal

cd /path/to/python/libs
python3 -m "http.server" 8000

Second Terminal

cat webimport/webimport.py script.py | ssh -R 8000:localhost:8000 [email protected] 'cat -|python3'

Running scripts without SSH tunnel:

Set up the webserver with the webimport-server command (requires pip installation of webimport):

webimport-server my_package_dir

From the other device, you can run any of the following:

Open a terminal and pre-load the webimport code

python3 -i -c "$(curl server-url:8080/_hook)"