Command Line Interface (CLI) for Typesafe ConductR

Installation

Python 3 is required. For OS X users use brew install python3.

Install using pip

You may either install or upgrade to all users:

sudo pip3 install -U conductr-cli

... or, and if you’re not using brew (there’s a problem with user installs as of the time writing this), install to the current user (make sure to have ~/.local/bin in your PATH):

pip3 install -U --user conductr-cli

Install as a deb package

Build a docker image for building a deb package:

docker build -t debian-distribution deb_dist/

Run built docker image:

docker run -v $(pwd):/source debian-distribution

Install built deb package:

dpkg -i deb_dist/python3-conductr-cli_0.1-1_all.deb

Install required dependencies:

apt-get install -f

Autocomplete support

If you have installed argcomplete and want to activate Bash completion for the CLI, you have to execute the following, either transiently in your terminal session or more permanently in your .bashrc or .bash_profile:

eval "$(register-python-argcomplete conduct)"

Alternatively, if you have a Bash version 4.2 or later, you can activate global completion once:

activate-global-python-argcomplete --dest=/path/to/bash_completion.d

If you are running zsh, execute the following command to enable autocomplete:

autoload bashcompinit && autoload compinit && bashcompinit && compinit && eval "$(register-python-argcomplete conduct)"

Running tests

Execute the following command to run all defined tests:

python3 setup.py test

To run only a specific test case in a test suite:

python3 setup.py test -a "-- -s conductr_cli.test.test_conduct_unload:TestConductUnloadCommand.test_failure_invalid_address"

Releasing

CLI releases can be performed completely from the GitHub project page. Follow these steps to cut a release:

1. Edit conductr_cli/__init__.py file to contain the version to be released.
2. Create a new release in GitHub releases page.

After CI build is finished for the tagged commit, new version will automatically be deployed to PyPi repository.

CLI Usage

conduct

Execute conduct with any of the supported sub-commands or options, e.g.

$ conduct -h
usage: conduct.py [-h]
              {version,info,services,load,run,stop,unload,events,logs} ...

optional arguments:
  -h, --help            show this help message and exit

subcommands:
  valid subcommands

  {version,info,services,load,run,stop,unload,events,logs}
                        help for subcommands
    version             print version
    info                print bundle information
    services            print service information
    load                load a bundle
    run                 run a bundle
    stop                stop a bundle
    unload              unload a bundle
    events              show bundle events
    logs                show bundle logs

Most sub-commands connect to a ConductR instance and therefore you have to specify its IP and port; if not given, CONDUCTR_IP environment variable or 127.0.0.1 will be used for the IP and CONDUCTR_PORT or 9005 for the port. Alternatively you can specify the IP via the --ip option and the port via the --port option.

The commands provided via CLI uses version 1.0 of the ConductR API by default. When working with version 1.1 of ConductR, set the CONDUCTR_API_VERSION environment variable to 1.1. Alternatively you can specify the API version via the --api-version option.

Here’s an example for loading a bundle:

conduct load sbt-conductr-tester-1.0.0-e172570d3c0fb11f4f9dbb8de519df58dcb490799f525bab43757f291e1d104d.zip

Note that when specifying IPV6 addresses then you must surround them with square brackets e.g.:

conduct info --ip [fe80:0000:0000:0000:0cb3:e2ff:fe74:902d]

shazar

The shazar command can be used:

• for packaging a directory that has a structure of a bundle to a bundle archive;
• for packaging a bundle’s configuration to a bundle archive;

In both cases the source files are zipped and a SHA256 digest of the archive is appended to the bundle archive file name.

For pointers on command usage run shazar -h.

Information for developers

If you want to run conduct locally, i.e. without installation, cd into the project directory and execute:

python3 -m conductr_cli.conduct

Make sure to install the necessary dependencies:

pip install -e .