Python: Bug: Semantic Kernel breaks FASTAPI Debugging Launch config

[sminath33]

Describe the bug
After adding the semantic kernel import to a python project running on FASTAPI from the VSCode Debugger this breaks pymeta (which is a required dependency). This prevents all debugging for our project. Seems there is a conflict between VSCode debugger and semantic kernel

To Reproduce
Steps to reproduce the behavior:

1. Add Semantic Kernel as a python dependency
2. Add 'import semantic_kernel' to a file
3. Launch FastAPI using uvicorn (-m uvicorn api-hosting-lambda.src.app:app --reload --port 8002) from a launch config (see below)
4. pymetic error

Exception has occurred: ParseError (note: full exception trace is shown but execution is paused at: _G_many_2)
(7, [('expected', 'letter or digit', None)])
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 259, in _G_many_2 (Current frame)
_G_apply_1, lastError = self._apply(self.rule_letterOrDigit, "letterOrDigit", [])
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 262, in rule_name
_G_many_3, lastError = self.many(_G_many_2)
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 604, in _G_lookahead_2
_G_apply_1, lastError = self._apply(self.rule_name, "name", [])
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 608, in rule_rule
_G_lookahead_3, lastError = self.lookahead(_G_lookahead_2)
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 641, in _G_many_1
_G_apply_1, lastError = self._apply(self.rule_rule, "rule", [])
File "/pymeta_generated_code/pymeta_grammar__Grammar.py", line 644, in rule_grammar
_G_many_2, lastError = self.many(_G_many_1)
File "/home/ubuntu/GenerativeAIGit/koch-gpt/api-hosting-lambda/src/koch_ai.py", line 8, in
import semantic_kernel
File "/home/ubuntu/GenerativeAIGit/koch-gpt/api-hosting-lambda/src/conversations_handler.py", line 26, in
from koch_ai import KochAIChatBot
File "/home/ubuntu/GenerativeAIGit/koch-gpt/api-hosting-lambda/src/app.py", line 6, in
from conversations_handler import (
File "", line 1, in
pymeta.runtime.ParseError: (7, [('expected', 'letter or digit', None)])

Expected behavior
FastAPI should start up in the debugging session within VSCode

Screenshots
VSCode debugging Launch config
{
"name": "API Hosting on FastAPI",
"type": "python",
"request": "launch",
"module": "uvicorn",
"args": [
"api-hosting-lambda.src.app:app",
"--reload",
"--port",
"8002"
],
"cwd": "${workspaceFolder}/api-hosting-lambda/src",
"env": {
"JSON_LOGS": "0",
"LOG_LEVEL": "debug"
},
"jinja": true,
"justMyCode": true
}
Platform

• OS: Ubuntu
• IDE: VSCode
• Language: Python
• Source: FastAPI project

Additional context
Starting the app with the same command 'poetry run python -m uvicorn app:app --reload --port 8002' outside of a debugger works fine.

[eavanvalkenburg]

Hi @sminath33

Not sure what is going on here, I just tried this and did not encounter issues, I have the following script, using fastapi[stardard] or fastapi[all] and SK (both just take the latest from pypi:

from typing import Union

import semantic_kernel
from fastapi import FastAPI

app = FastAPI()


kernel = semantic_kernel.Kernel()


@app.get("/")
def read_root():
    return {"Hello": "World"}


@app.get("/items/{item_id}")
def read_item(item_id: int, q: Union[str, None] = None):
    return {"item_id": item_id, "q": q}

and this starts, both with the same launch setup as you have, and with the newer debugpy launch command.

{
    "name": "Python Debugger: FastAPI",
    "type": "debugpy",
    "request": "launch",
    "module": "uvicorn",
    "args": [
        "main:app",
        "--reload"
    ],
    "jinja": true
},

[eavanvalkenburg]

Please reopen if you are still having issues!

[MahrRah]

I ran into the same issue and realized I had the User Uncaught exception option selected under Breakpoints. Once I unchecked it, I was able to debug as expected.