Deploying python api and python flask application on azure function app and azure app respectively using azure devops

Question

I have a python azure functions app deployed via azure devops in an api_env virtual environment. Similarly, I have a python flask app i.e. python backend and typescript frontend. The python backend is deployed by activate a different virtual environment called backend_env. The application backend code calls an api i.e. that was deployed as azure function. However, some of the packages used such as openai, langchain etc are in requirements.txt and are in the api_env folder on the azure function app when i look at it via debug console but the azure functions is throwing module not found error for these packages. I tried various options i.e. in addition to activating the api_env via the devops pipeline, I also added the setting in the config setting but nothing seems to help. Can anyone point me in the right direction?

`YAML code steps: - script: | python3.9 -m venv api_env source api_env/bin/activate pip install --upgrade setuptools pip3.9 install -r requirements.txt workingDirectory: $(Build.SourcesDirectory)/api/python displayName: 'Install Python dependencies for api'

- task: AzureFunctionApp@1
  inputs:
    azureSubscription: 'spn'
    appType: 'functionAppLinux'
    appName: '$(knowbot-funcapp-name)'
    package: '$(System.DefaultWorkingDirectory)/api/python'
    startUpCommand: |
      source api_env/bin/activate
      func start --python
    workingDirectory: '$(System.DefaultWorkingDirectory)/api/python'
  displayName: 'Deploy Function App'`

Also tried adding the below in the function init.py sys.path.append('/home/site/wwwroot/api_env/lib/python3.9/site-packages')

I am expecting the api call to succeed.

Answer 1

I tried to deploy a Function app with virtual environment in Azure Devops using Openai module in requirements.txt and it was successful, Refer below:-

My init.py code:-

import openai
import logging
import azure.functions as func

def main(req: func.HttpRequest) -> func.HttpResponse:
    # Set up OpenAI API key
    openai.api_key = "sk-8Zg7i4icv8gh60e3IhnaT3Bxxxxx"

    # Use the OpenAI API to generate some text
    prompt = "Hello, OpenAI!"
    response = openai.Completion.create(engine="text-davinci-002", prompt=prompt)

    # Log the response
    logging.info(response)

    # Return the response as an HTTP response
    return func.HttpResponse(response.choices[0].text)

My function.json code:-

{
    "bindings": [
        {
            "authLevel": "anonymous",
            "type": "httpTrigger",
            "direction": "in",
            "name": "req",
            "methods": [
                "get"
            ]
        },
        {
            "type": "http",
            "direction": "out",
            "name": "$return"
        }
    ],
    "scriptFile": "__init__.py",
    "entryPoint": "main"
}

My host.json:-

{
  "version": "2.0",
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "excludedTypes": "Request"
      }
    }
  },
  "extensionBundle": {
    "id": "Microsoft.Azure.Functions.ExtensionBundle",
    "version": "[3.*, 4.0.0)"
  }
}

My requirements.txt :-

azure-functions
openai

My YAML Pipeline script:-

trigger:
- main

pool:
  vmImage: ubuntu-latest

steps:

- task: UsePythonVersion@0
  inputs:
    versionSpec: '3.9'
    addToPath: true
- script: |
    python3 -m venv env
    source env/bin/activate
    pip install --upgrade pip
    pip install -r requirements.txt
  displayName: 'Install Python dependencies'
- task: AzureFunctionApp@1
  inputs:
    azureSubscription: '<Subscription>'
    appType: 'functionAppLinux'
    appName: 'siliconfunc32'
    package: '$(System.DefaultWorkingDirectory)'
    startUpCommand: 'func start --python'

Check if your virtual environment api_env starts before your Function app is started. In your YAML code you're running virtual environment before your dependencies are installed try initiating your virtual environment right before your Function starts like below:-

startUpCommand: |
  source api_env/bin/activate
  func start --python

Also, Validate where your virtual environment is located you can enter this command to check your system path below:-

import sys
print(sys.path)

Add full path of your virtual environment while initiating it with the code below:-

startUpCommand: |
  source /path/to/api_env/bin/activate
  func start --python