How to "inject" context into namespace so that pylance is aware of those variables

[ya5has]

I know this is a very odd scenario. Consider the code below:

import spot_name_functions as SNF
import Enumerations_Various as EV
import delay
import logger

SNF.set_Current_FMS_VSpeed_STATUS(EV.STATUS_NORM)
vista.wait(delay.max_delay())
actual = IAS_Approach_Climb_Speed_Is_Auto.getValue()
expected = True
logger.test(actual, '==', expected)

Notice that vista and _IAS_Approach_Climb_Speed_IsAuto are clearly not imported (or defined) in this file; therefore pylance shows a squiggly line warning. However these two are special-objects that just works when the file is run in a proprietary software at my work. I like to use vscode to explore these files, debug them by attaching it to said proprietary software process and modify them if needed. I know the locations of the files in which these special-objects are defined.

Now my question is how can I inform pylance about these special objects so that I don't get the warning? (and hopefully get code-completion). I have the following constraints though:

• I do not want to globally disable reportUndefinedVariable
• The development of these files are also done with the same proprietary software therefore, I cannot add any additional code to these files just because I prefer to use vscode.

I have been ignoring these warnings but on any given file there might be over 100 occurrences of these and it stops me from knowing truly undefined variable when I edit extensively. I'm thinking of writing a custom vscode extension that runs a script when I open any python file and maybe interact with the language server and/or define these objects. I don't know if that sounds stupid. Is that even possible? If so, can you point me in the right direction on how I can achieve that? Is there a easier solution?

[erictraut]

Pylance is built on top of Microsoft's open-source type checker called Pyright. Pyright has a facility for extending the "builtins" namespace. For details, refer to this documentation. The basic idea is to create a type stub called __builtins__.pyi in your project's root directory. Within this type stub file, you can define symbols (functions, variables, classes, etc.) that should be "injected" to all modules within your project.