cellmap_flow.cli.main

cellmap_flow: the one command, with a subcommand for each job.

: infer <type> start a model’s inference server and open the viewer on it yaml the same for the models a YAML file lists add REF the model entry for a path, URL, repo or model name view open a dataset in the viewer; pick models in the dashboard dashboard serve the dashboard alone, for a viewer already running serve serve one model’s predictions (the launchers run this) blockwise run models over a whole volume, writing predictions to disk finetune the finetune tools: train, export-merged, build-corrections models list the model types and their arguments plugins register, unregister and list plugins envs the environments models run in: list, install, check doctor check the environment

Before 0.3.0 these were separate console scripts (cellmap_flow_yaml and the rest). Those names still work for one release: cli/aliases.py says where each went and runs it. So do this command’s own earlier subcommands, hidden from --help: cellmap_flow <type>, run (cli/infer.py), list-models, register, unregister and list-plugins.

Classes

CellMapFlowGroup

The cellmap_flow group. cellmap_flow <type>, its per-type commands

Functions

`cli`()	CellMap Flow: real-time model inference on EM data, in Neuroglancer.
`dashboard`(neuroglancer_url)	Serve the dashboard alone, for a viewer already running.
`models`()	List the model types and the arguments each takes.
`plugins_group`()	Register, unregister and list plugins.
`register_cmd`(filepath, force)	Register a custom plugin (normalizer, postprocessor, or model config).
`unregister_cmd`(name)	Unregister a previously registered plugin by name.
`list_plugins_cmd`()	List all registered plugins.
`finetune`()	The finetune tools. The dashboard's finetune tab runs train for you.
`main`([args])	The `cellmap_flow` console script: load the plugins, then run the

Module Contents

class cellmap_flow.cli.main.CellMapFlowGroup(name: str | None = None, commands: collections.abc.MutableMapping[str, Command] | collections.abc.Sequence[Command] | None = None, invoke_without_command: bool = False, no_args_is_help: bool | None = None, subcommand_metavar: str | None = None, chain: bool = False, result_callback: Callable[..., Any] | None = None, **kwargs: Any)

The cellmap_flow group. cellmap_flow <type>, its per-type commands before 0.3.0, are built on request as deprecated aliases of infer.

get_command(ctx, name): Given a context and a command name, this returns a Command object if it exists or returns None.

cellmap_flow.cli.main.cli()

CellMap Flow: real-time model inference on EM data, in Neuroglancer.

Each job is a subcommand; cellmap_flow <command> –help shows its options.

: cellmap_flow view -d /path/to/data.zarr cellmap_flow yaml config.yaml cellmap_flow infer dacapo -r my_run -i 100 -d /path/to/data cellmap_flow blockwise config.yaml

cellmap_flow.cli.main.dashboard(neuroglancer_url)

Serve the dashboard alone, for a viewer already running.

view, yaml and infer start the dashboard with the viewer they open; this is the dashboard on its own, as cellmap_flow_app served it before 0.3.0. It prints its URL, and writes it to the file named by SERVICE_URL_PATH when that is set. Ctrl+C stops it and kills the models it launched.

cellmap_flow.cli.main.models(): List the model types and the arguments each takes.

cellmap_flow.cli.main.plugins_group()

Register, unregister and list plugins.

A plugin is a .py file defining subclasses of InputNormalizer, PostProcessor or ModelConfig; registering copies it to ~/.cellmap_flow/plugins/.

cellmap_flow.cli.main.register_cmd(filepath, force)

Register a custom plugin (normalizer, postprocessor, or model config).

FILEPATH is the path to a .py file defining subclasses of InputNormalizer, PostProcessor, or ModelConfig.

Example

cellmap_flow plugins register my_normalizer.py

cellmap_flow.cli.main.unregister_cmd(name)

Unregister a previously registered plugin by name.

NAME is the plugin filename (with or without .py extension).

Example

cellmap_flow plugins unregister my_normalizer

cellmap_flow.cli.main.list_plugins_cmd(): List all registered plugins.

cellmap_flow.cli.main.finetune(): The finetune tools. The dashboard’s finetune tab runs train for you.

cellmap_flow.cli.main.main(args=None): The cellmap_flow console script: load the plugins, then run the command args (by default, the process’s arguments) names.