使用 vscode 和 conda 在本地调试 azure function

Question

there has been multiple articles online on how to run somewhat similar scenario.

I have tried following and nothing works completely.

1. Used conda at machine level
2. used conda inside VScode
3. used venv (not conda).

issue is how launch.json calls "host start". this command changes into whatever is the value of "azureFunctions.pythonVenv" and then append " /Scripts/activate ".

now the issue is.

1. if i am using conda inside vscode, which creates ".conda" folder inside my workspace. this particular conda is missing "activate" file so command fails.

2. if i use conda at machine level, it works, but it starts conda with environment as "base". there is no way i can change the environment inside "task.json", which type is "func"

3. i can change type of task from "func" to "shell", and then i can activate required env in conda, and azure function runs too. but it doesn't go into debug mode. it just gives connection error for:9091.

After burning multiple days on this issue. I can say that VSCODE + conda (not base) + AzureFunction + Debug doesnt work.

this is a pain. once you have multiple programs on your local machine, you lost any benefit conda provides.

Currently i am using option 2, as it atleast let me debug, which is a big deal for me.

and again, everything works fine, if I just want to run AzureFunction, multiple ways to do that. But issue comes when i have to use conda (particular environment).

Answer 1

config that worked, with using base conda.

tasks.json

{
    "version": "2.0.0",
    "tasks": [
        {
            "label": "pipInstall",
            "type": "shell",
            "windows": {
                "command": [
                    "conda.bat activate && pip install -r requirements.txt"
                ]
            },
            "problemMatcher": []
        },
        {
            "type": "func",
            "label": "func: host start",
            "command": "host start",
            "problemMatcher": "$func-python-watch",
            "isBackground": true,
            "dependsOn": ["pipInstall"]
        }
    ]
  }

settings.json (both settings may be a overkill, but its ok.

"python.defaultInterpreterPath": "C:XXXXXXXXX\\Miniconda3\\python.exe",
"azureFunctions.pythonVenv": "C:XXXXXXXXXXX\\Miniconda3",