Python API

The tools of the MCP server are described in Usage.

Client

Client of lys remote server (lys/localPlugins/Remote.py in lys).

This module does not depend on lys and Qt, so that it can be used from MCP server and command line. lys should be launched with –remote option, e.g. python -m lys –remote

Each lys has a label (python -m lys –remote LABEL, default is its process id). If only one lys is running, it is selected automatically. Otherwise, select it by –label or LYS_REMOTE environment variable.

lys on another computer can be used when it is launched with –port option (python -m lys –remote –port 8765). Select it by “tcp://<host>:<port>” (e.g. -l tcp://192.168.1.4:8765 or LYS_REMOTE=tcp://192.168.1.4:8765). If lys is launched with –token, give the token by –token or LYS_REMOTE_TOKEN environment variable.

Command line usage:

lys-remote --instances                    # list running lys
lys-remote "a = 1; a + 1"                 # execute code and print the result
lys-remote -l expA "a = 1; a + 1"         # execute code in lys launched with --remote expA
lys-remote -l ~/data/expA "a + 1"         # execute code in lys launched in ~/data/expA
lys-remote -f script.py                   # execute file
lys-remote --image out.png                # save image of the front canvas
lys-remote --image out.png --target g     # save image of the canvas/figure g
lys-remote --list                         # list variables in shell
lys-remote -l tcp://192.168.1.4:8765 "a"  # lys on another computer launched with --port 8765

class lys_mcp.client.LysClient(label=None, timeout=600, token=None)

Bases: object

Client to send requests to lys.

Parameters:

• label (str) – The label of lys. See resolveLabel().

• timeout (float) – Timeout in seconds.

• token (str) – The token for lys launched with –token. If None, LYS_REMOTE_TOKEN environment variable is used.

Example:

with LysClient() as c:
    res = c.exec("a = 1; a + 1")
    print(res["result"])   # "2"

close()

connect()

exec(code)

Execute code in lys shell. The value of the last expression is returned as ‘result’.

image(target=None, dpi=100)

Return response with png image (base64) of canvas, matplotlib figure or widget as ‘image’.

info()

Return label, pid, home directory and protocol version of lys.

property label

list()

Return variables in lys shell as ‘result’.

request(op, **kwargs)

Send a request and return the response as dict.

exception lys_mcp.client.LysRemoteError

Bases: Exception

lys_mcp.client.findLabel(key)

Return the label of lys whose label or home directory is key. If not found, key is returned.

lys_mcp.client.formatResponse(res)

Format response as text.

lys_mcp.client.instances()

Return the information of running lys as list of dict.

lys on another computer is included when it is given by LYS_REMOTE environment variable.

lys_mcp.client.listLabels()

Return the labels of running lys.

lys_mcp.client.main()

lys_mcp.client.resolveLabel(label=None)

Return the label of lys to be connected.

label can be the label or the home directory of lys (the directory where lys was launched), or “tcp://<host>:<port>” for lys on another computer. If label is None, LYS_REMOTE environment variable is used. If it is not set, the running lys is selected when only one lys is running.

lys_mcp.client.serverName(label)

Return the name of the local socket for label. The same rule as serverName in lys/localPlugins/Remote.py.