plexora connect

[user@]host to ssh into, or the name of a server saved with --save (or in Settings > Remote servers).

Optional datasource/project name to open directly.

How to invoke Plexora on the remote host (default "plexora"). Use this when it is not on a non-interactive PATH. An environment prefix is enough -- "/home/you/miniconda3/envs/plexora" -- or give a full command, e.g. "conda run --no-capture-output -n imaging plexora".

Remember this connection under NAME, so next time plexora connect NAME is the whole command. Saved servers are shared with Settings > Remote servers. No password is stored.

Treat the target as a SLURM login node and run Plexora inside a job, e.g. --srun "-p interactive -t 4:00:00 -c 16 --mem 128G" -- which is what the Settings form fills in.

With --srun, forward from the login node instead of ssh-ing into the compute node. For sites that refuse the latter.

Forward another port from the remote host, e.g. --forward 8642 for a plexora node serve running beside the viewer. Repeat for each. Only needed when your BROWSER has to reach the node directly; the viewer reaches it over there without any of this.

ssh -J jump host to reach the target through.

Extra ssh -o option; repeatable. Keys, ports and usernames are usually better placed in ~/.ssh/config.

Local port to serve the tunnel on (default: a free one).

Port to use on the remote host (default: a free-looking high one).

Seconds to wait for Plexora to answer (default 60, or 18000 with --srun, where the job may sit in a queue).

Plexora data directory on the remote host, adopted if that account has not chosen one yet. An account that has is left alone -- otherwise the viewer reads one directory while that account's own commands write to another.

Plugins to activate on the remote host.

Set up the tunnel and print the URL, but do not open a browser.

A file on the REMOTE host to serve as a data node beside the viewer, e.g. --also-serve table:cells=/scratch/cells.h5ad. Started, forwarded and registered automatically. Repeat for each.

A file on THIS machine to serve to the remote viewer, e.g. --local-serve table:cells=~/study/cells.h5ad. For the layout where the images are on the cluster and the cell table never left your laptop. Repeat for each.

Install or update Plexora on the remote host before launching it: pip install --upgrade plexora, in the environment --remote-command names. Off by default, because it writes to that machine. Saved with --save, and the same switch the Settings form calls "Install or update Plexora".

Do not run a data node on this computer. One is started for every connection by default, with nothing to serve until you pick something: it is what lets the viewer's data forms offer files from HERE, which they otherwise cannot reach at all.

What to call the data nodes this connection registers. Defaults to the saved server's name, or the remote host's.

Port for the remote data node (default: a free-looking high one).

Another browser origin the data nodes this connection starts should answer, besides this machine's own viewer address. Repeat for each.

Also print a link per data node that pairs it with the Plexora page at ORIGIN (default https://viewer.plexoraapp.com), and allow that origin on every node. Put it after the server name.