cellmap_flow.dashboard.app

Attributes

`logger`
`template_dir`
`static_dir`
`app`
`package_logger`
`log_handler`

Functions

`bind_dashboard`()	Bind the dashboard to a free port, and record its address as the
`create_and_run_app`([neuroglancer_url, server])	Serve the dashboard on a free port (or on `server`, bind_dashboard's),

Module Contents

cellmap_flow.dashboard.app.logger

cellmap_flow.dashboard.app.template_dir

cellmap_flow.dashboard.app.static_dir

cellmap_flow.dashboard.app.app

cellmap_flow.dashboard.app.package_logger

cellmap_flow.dashboard.app.log_handler

cellmap_flow.dashboard.app.bind_dashboard()

Bind the dashboard to a free port, and record its address as the session’s dashboard_url. Returns the server, for create_and_run_app.

Apart so the address is known before the dashboard serves: the CLIs’ startup (services.startup) builds the viewer’s prediction layers first, and each layer’s URL carries the address (Session.layer_blob).

cellmap_flow.dashboard.app.create_and_run_app(neuroglancer_url=None, server=None)

Serve the dashboard on a free port (or on server, bind_dashboard’s), until the process is stopped.

It prints its URL, and writes it where SERVICE_URL_PATH says. Its terminal lines (the URL, the models it launches) are INFO: a cellmap_flow command has configured logging by now, and a script that calls this without doing so gets the shared format at INFO. It loads the plugins, whose normalizers, postprocessors and model types its pages offer, if nothing has yet.