CLI Reference

FlowerPower provides a Typer-based command-line interface accessible as flowerpower. It is the fastest way to initialize projects, manage pipelines, and start the Hamilton UI.

flowerpower [OPTIONS] COMMAND [ARGS]...

Use flowerpower --help or flowerpower <COMMAND> --help to see the live options for your installed version.

Global options

Option	Description
--install-completion	Install shell completion for the current shell.
--show-completion	Print the shell completion script for copying or customizing.
--help	Show the help message and exit.

flowerpower init

Create a new FlowerPower project. This calls FlowerPowerProject.new(...) under the hood.

flowerpower init [OPTIONS]

Option	Short	Description
--name TEXT	-n	Project name. If omitted, the current directory name is used.
--base-dir TEXT	-d	Directory where the project is created. If omitted, the parent of the current directory is used.
--storage-options TEXT	-s	Storage options as a JSON/dict string.
--help		Show help.

flowerpower init --name my_project
flowerpower init --name my_project --base-dir ./projects

Note

The old FlowerPowerProject.init(...) method has been removed. Use .new(...) or the flowerpower init command instead.

flowerpower ui

Start the Hamilton UI web application.

flowerpower ui [OPTIONS]

Option	Short	Default	Description
--port INTEGER	-p	8241	Port for the UI server.
--base-dir TEXT	-d	~/.hamilton/db	Directory where the UI stores its data.
--no-migration			Skip database migrations on startup.
--no-open			Do not open the UI automatically in a browser.
--settings TEXT	-s	mini	UI settings profile (mini, dev, prod).
--config TEXT	-c		Custom UI configuration file.
--help			Show help.

flowerpower ui
flowerpower ui --port 9000 --no-open

Tip

Requires the ui extra. Install it with pip install "flowerpower[ui]" or uv pip install 'flowerpower[io,ray,ui]'.

flowerpower pipeline

Manage and execute pipelines.

flowerpower pipeline [OPTIONS] COMMAND [ARGS]...

flowerpower pipeline run

Run a pipeline synchronously. The command builds a RunConfig internally from the supplied flags and waits for the pipeline to finish.

flowerpower pipeline run [OPTIONS] NAME

Option	Short	Default	Description
--executor TEXT			Executor type (e.g. threadpool, local).
--executor-cfg TEXT			Executor configuration as a JSON/dict string.
--executor-max-workers INTEGER			Convenience flag: set max_workers.
--executor-num-cpus INTEGER			Convenience flag: set num_cpus.
--base-dir TEXT	-d		Base directory containing the project.
--inputs TEXT			Input parameters as JSON, dict string, or key=value pairs.
--final-vars TEXT, --outputs TEXT, -o			Final variables to compute, as JSON or a list.
--config TEXT			Hamilton executor configuration as JSON/dict string.
--cache TEXT			Cache configuration as JSON/dict string.
--storage-options TEXT	-s		Storage options as JSON, dict string, or key=value pairs.
--log-level TEXT			Logging level (debug, info, warning, error, critical).
--with-adapter TEXT			Adapter configuration as JSON/dict string.
--max-retries INTEGER		0	Deprecated. Use the nested retry config via Python when possible.
--retry-delay FLOAT		1.0	Deprecated. Use the nested retry config via Python when possible.
--jitter-factor FLOAT		0.1	Deprecated. Use the nested retry config via Python when possible.
--help			Show help.

flowerpower pipeline run hello
flowerpower pipeline run hello --inputs '{"message": "Hi"}'
flowerpower pipeline run hello --final-vars '["full_greeting"]' --log-level debug
flowerpower pipeline run hello --executor threadpool --executor-max-workers 4
flowerpower pipeline run hello --with-adapter '{"hamilton_tracker": true}'

Warning

--max-retries, --retry-delay, and --jitter-factor are deprecated and emit a DeprecationWarning. For production retry logic, build a RunConfig in Python and use the nested retry block (see the API reference).

Tip

The adapter key is hamilton_tracker, not tracker. The opentelemetry adapter has been removed.

flowerpower pipeline new

Create a new pipeline scaffold: a configuration file under conf/pipelines/ and a module file under pipelines/.

flowerpower pipeline new [OPTIONS] NAME

Option	Short	Default	Description
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options as JSON/dict string or key=value pairs.
--log-level TEXT			Logging level.
--overwrite / --no-overwrite		no-overwrite	Overwrite an existing pipeline with the same name.
--help			Show help.

flowerpower pipeline new hello
flowerpower pipeline new hello --overwrite

flowerpower pipeline delete

Delete a pipeline's configuration file and/or module file. If neither --cfg nor --module is passed, both are removed.

flowerpower pipeline delete [OPTIONS] NAME

Option	Short	Default	Description
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--cfg	-c		Delete only the configuration file.
--module	-m		Delete only the pipeline module.
--help			Show help.

flowerpower pipeline delete hello
flowerpower pipeline delete hello --cfg
flowerpower pipeline delete hello --module

flowerpower pipeline show-dag

Display the execution graph of a pipeline.

flowerpower pipeline show-dag [OPTIONS] NAME

Option	Short	Default	Description
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--format TEXT		png	Output format: png, svg, pdf, or raw (returns the graph object).
--help			Show help.

flowerpower pipeline show-dag hello
flowerpower pipeline show-dag hello --format svg

flowerpower pipeline save-dag

Save a pipeline DAG to a file.

flowerpower pipeline save-dag [OPTIONS] NAME

Option	Short	Default	Description
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--format TEXT		png	Output format: png, svg, or pdf.
--output-path TEXT		<name>.<format>	Custom path for the saved file.
--help			Show help.

flowerpower pipeline save-dag hello
flowerpower pipeline save-dag hello --format svg --output-path ./vis/hello.svg

flowerpower pipeline show-pipelines

List all available pipelines in the project.

flowerpower pipeline show-pipelines [OPTIONS]

Option	Short	Default	Description
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--format TEXT		table	Output format: table, json, or yaml.
--help			Show help.

flowerpower pipeline show-pipelines
flowerpower pipeline show-pipelines --format json

flowerpower pipeline show-summary

Show summary information for one or all pipelines, optionally including configuration, code, and project context.

flowerpower pipeline show-summary [OPTIONS]

Option	Short	Default	Description
--name TEXT			Specific pipeline name. All pipelines are summarized if omitted.
--cfg / --no-cfg		cfg	Include configuration details.
--code / --no-code		code	Include code/module details.
--project / --no-project		project	Include project context.
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--to-html / --no-to-html		no-to-html	Output as HTML.
--to-svg / --no-to-svg		no-to-svg	Output as SVG.
--output-file TEXT			Save the output to a file instead of printing.
--help			Show help.

flowerpower pipeline show-summary
flowerpower pipeline show-summary --name hello --cfg --code --no-project
flowerpower pipeline show-summary --to-html --output-file report.html

flowerpower pipeline add-hook

Add a hook to a pipeline configuration.

flowerpower pipeline add-hook [OPTIONS] NAME

Option	Short	Default	Description
--function TEXT	-f	(required)	Name of the hook function.
--type [mqtt-build-config]		mqtt-build-config	Hook type.
--to TEXT			Target node name or tag (required for node-specific hooks).
--base-dir TEXT	-d		Base directory for the project.
--storage-options TEXT	-s		Storage options.
--log-level TEXT			Logging level.
--help			Show help.

flowerpower pipeline add-hook hello --function log_results --type MQTT_BUILD_CONFIG

Note

Only MQTT_BUILD_CONFIG (mqtt-build-config) is currently supported as a hook type.