Command-line interface

giskard

Giskard Command Line

giskard [OPTIONS] COMMAND [ARGS]...

Options

--version

Show the version and exit.

hub

Giskard UI management

giskard hub [OPTIONS] COMMAND [ARGS]...

clean

Delete Docker container, container (and possibly a volume) associated with the current version of Giskard Hub
giskard hub clean [OPTIONS]

Options

--data

Delete user data (giskard-home volume)

-v, --verbose

Enable verbose logging

diagnose

Save hub logs to a local archive (Useful for support).
giskard hub diagnose [OPTIONS]

Options

-o, --out_path <local_dir>

Destination directory to save diagnose archive to

-v, --verbose

Enable verbose logging

expose

Expose your local Giskard Hub to the outside world using ngrok to use in notebooks like Google Colab
giskard hub expose [OPTIONS]

Options

--ngrok-token <token>

Required In case you have an ngrok account, you can use a token generated from https://dashboard.ngrok.com/get-started/your-authtoken

-v, --verbose

Enable verbose logging

logs

Prints logs of hub services
giskard hub logs [OPTIONS] [[backend|frontend|worker|db]]

Options

-l, --lines <nb_lines>

Number of log lines to show

-f, --follow

Follow the logs stream

-v, --verbose

Enable verbose logging

Arguments

SERVICE

Optional argument

restart

Restart Giskard Hub.

Stops any running Giskard hub and starts it again.

giskard hub restart [OPTIONS] [[backend|frontend|worker|db]]

Options

--hard

Hard restart. Restarts the whole container

-v, --verbose

Enable verbose logging

Arguments

SERVICE

Optional argument

start

Start Giskard Hub.

By default, the server starts detached and will run in the background. You can attach to it by using -a

giskard hub start [OPTIONS]

Options

-a, --attach

Starts the server and attaches to it, displaying logs in console.

-s, --skip-version-check

Force the server to start with a different version of the giskard python library.

-e, --env <environment>

Set environment variables to the server.

--env-file <env_file>

Read in a file of environment variables.

--version <version>

Version of Giskard hub to start

-v, --verbose

Enable verbose logging

status

Check if server container is running and status of each internal service
giskard hub status [OPTIONS]

Options

-v, --verbose

Enable verbose logging

stop

Stop Giskard Hub.

Stops a running Giskard hub. Does nothing if Giskard hub is not running.

giskard hub stop [OPTIONS]

Options

-v, --verbose

Enable verbose logging

upgrade

Update Giskard Hub. Uses the latest available version if not specified.
giskard hub upgrade [OPTIONS] [VERSION]

Options

-v, --verbose

Enable verbose logging

Arguments

VERSION

Optional argument

worker

ML Worker management

giskard worker [OPTIONS] COMMAND [ARGS]...

logs

giskard worker logs [OPTIONS]

Options

-v, --verbose

Enable verbose logging

-n, --lines <lines>

Output the last N lines of the log file, 10 lines are displayed by default

-f, --follow

Output appended data as new logs are being generated

restart

Restart ML Worker

giskard worker restart [OPTIONS]

Options

-v, --verbose

Enable verbose logging

-u, --url <url>

Remote Giskard hub url

--name <worker_name>

Name/id of the worker starting

-k, --key <api_key>

Giskard hub API key

--hf-token <hf_token>

Access token for Giskard hosted in a private Hugging Face Spaces

-s, --server

Server mode. Used to control Giskard managed ML Worker

--parallelism <nb_workers>

Number of processes to use for parallelism (None for number of cpu)

-d, --daemon

Should ML Worker be started as a Daemon in a background

--log-level <log_level>

Global log level

Options:

DEBUG | INFO | WARNING | ERROR

Environment variables

GSK_API_KEY

Provide a default for --key

GSK_HF_TOKEN

Provide a default for --hf-token

start

Start ML Worker.

ML Worker can be started in 2 modes:

  • server: Giskard starts an ML Worker managed by itself on the server.

  • client: ML Worker acts as a client and should connect to a running Giskard instance

    by specifying this instance’s host and port.

giskard worker start [OPTIONS]

Options

-v, --verbose

Enable verbose logging

-u, --url <url>

Remote Giskard hub url

--name <worker_name>

Name/id of the worker starting

-k, --key <api_key>

Giskard hub API key

--hf-token <hf_token>

Access token for Giskard hosted in a private Hugging Face Spaces

-s, --server

Server mode. Used to control Giskard managed ML Worker

--parallelism <nb_workers>

Number of processes to use for parallelism (None for number of cpu)

-d, --daemon

Should ML Worker be started as a Daemon in a background

--log-level <log_level>

Global log level

Options:

DEBUG | INFO | WARNING | ERROR

Environment variables

GSK_API_KEY

Provide a default for --key

GSK_HF_TOKEN

Provide a default for --hf-token

stop

Stop running ML Workers

giskard worker stop [OPTIONS]

Options

-v, --verbose

Enable verbose logging

-u, --url <url>

Remote Giskard hub url

--name <worker_name>

Name/id of the worker starting

-k, --key <api_key>

Giskard hub API key

--hf-token <hf_token>

Access token for Giskard hosted in a private Hugging Face Spaces

-s, --server

Server mode. Used to control Giskard managed ML Worker

-a, --all

Stop all running ML Workers

Environment variables

GSK_API_KEY

Provide a default for --key

GSK_HF_TOKEN

Provide a default for --hf-token