Debug Python code

1- Print to file

To debug Python code it is always possible to print data of variables to a file, with something like this:

f=open('C:/temp/my_debug.txt','a')
f.write('hello world\n')
f.close()

2- Python embedded in GiD

The scenery of use python in GiD is not usual:

• The ‘normal’ case is to run a python.exe process that evaluate the code of a .py python file.

In the ‘normal’ case it is easy to use the Visual Studio Code editor, install the Python extension of Microsoft, and open a .py file, set breakpoints with <F9> and start debugging with <F5> and see the value of variables, stack trace, etc when the flow reaches a break point.

• The GiD case run a gid.exe that has embedded a Tcl interpreter and at run time is loaded the tohil package that create an embedded Python interpreter and add Tcl commands to call Python to this interpreter . At run time it is also possible that this Python interpreter import the tohil module that add Python functions to call the Tcl interpreter.

Note: Running python.exe and importing tohil will create a new Tcl interpreter (not related at all with the one embedded in GiD, then GiD commands cannot be used)

It is possible in this case to do a ‘remote debugging’ with Visual Studio Code, connecting the debugger with the embedded Python of a running gid.exe in a host and port (e.g. localhost and 5678)

but before gid must start a server on this host and port calling our Tcl proc GiD_Python_StartDebuggerServer

proc GiD_Python_StartDebuggerServer { } {
    tohil::import debugpy
    #tohil::exec "debugpy.log_to('[GidUtils::GetTmp]')"
    set my_python [file join [gid_filesystem::get_folder_standard scripts] tohil/python/bin/x64/python.exe]
    #to avoid a bug of debugpy that use sys.executable and try to run another gid.exe!!. see https://github.com/microsoft/debugpy/issues/262
    tohil::exec "debugpy.configure(python='$my_python')"
    tohil::exec "debugpy.listen(5678)"
    #tohil::exec "debugpy.wait_for_client()"
    return 0
}

Note: The required Python module debugpy is pre-installed in our tohil’s Python copy

Then open the folder with the .py file in VS Code,

open the file gid_meshing.py and set an stop <F9> in the line np.set_printoptions(threshold=np.inf)

Then click in the lower state bar to set as python interpreter the one of GiD (probably initially points to another Python of our system),

e.g. select

<GiD>\scripts\tohil\python\bin\x64\python.exe

Then start GiD and call GiD_Python_StartDebuggerServer

(e.g. can write -np- GiD_Python_StartDebuggerServer, or un-comment the line of meshio.tcl #GiD_Python_StartDebuggerServer)

And the first time that run the debug in VS Code <F5> it ask the configuration way to debug: select for example attach to remote debug, configuring the host=localhost and port=5678

this information is stored in the file .vscode/launch.json, with a content like this

{
    // Use IntelliSense to learn about possible attributes.
    // Hover to view descriptions of existing attributes.
    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Attach localhostt:5678",
            "type": "python",
            "request": "attach",
            "connect": {
                "host": "localhost",
                "port": 5678
              },
            //"processId": "${command:pickProcess}",
            //"justMyCode": true
        }
    ]
}

In GiD go to menu Files->Import->Meshio

and select some mesh file, like a Kratos .mdpa file to be imported.

then the debugger will be stopped in our break point and can inspect variables, etc.