M365 Developer: Using the Microsoft Graph Python SDK and Semantic Kernel

This application reads the emails from a user, extracts actions items and creates a To-Do list.

Preparation

1. Sign up for a sandbox tenant in the M365 Developer Program.
2. Log into the Azure portal with your M365 sandbox tenant.
3. Go to Azure Active Directory and register your application in the portal.

Create a virtual environment

To get started, create a folder somewhere in your computer or a GitHub repository to host your project and open the project in Visual Studio Code.

1. Inside your project, create an app.py file.
2. Press Ctrl + Shift + P and select Create environment -> Venv -> {python version you are using}.
3. Once the environment has been created, go to /.venv/Scripts/Activate.ps1 in the VS code terminal and run this file. For example: & c:/Users/isvargas/Documents/GitHub/m365dev-python/.venv/Scripts/Activate.ps1. This will activate your environment. You'll know your environment is active if there is a green (.env) at the beginning of your command line in the terminal. Example: (.venv) PS C:\Users\isvargas\Documents\GitHub\m365dev-python>
4. Inside this file, scroll down to the Add the venv to the PATH section and add the authentication variables. It should look like this:

# Add the venv to the PATHCopy-Item-Path Env:PATH -Destination Env:_OLD_VIRTUAL_PATH
$Env:PATH="$VenvExecDir$([System.IO.Path]::PathSeparator)$Env:PATH"$Env:AZURE_CLIENT_ID="client_id_from_app_registration"$Env:AZURE_CLIENT_SECRET="client_secret_from_app_registration"$Env:AZURE_TENANT_ID="tenant_id_from_app_registration"$Env:OPENAI_API_KEY=""$Env:OPENAI_ORG_ID=""$Env:AZURE_OPENAI_DEPLOYMENT_NAME=""$Env:AZURE_OPENAI_ENDPOINT=""$Env:AZURE_OPENAI_API_KEY=""

Replace the value_id_from_app_registration with the actual IDs you received from the app registration.

Install the Microsoft Graph Python SDK

Now we can start installing the libraries we'll use. In the VS Code terminal run:

pipinstallmsgraph-sdk

Note: Enable long paths in your environment if you receive a Could not install packages due to an OSError. For details, see Enable Long Paths in Windows 10, Version 1607, and Later.

Install the Semantic Kernel

The Semantic Kernel is an SDK that integrates Large Language Models (LLMs) like OpenAI, Azure OpenAI, and Hugging Face with conventional programming languages like C#, Python, and Java. To install the Semantic Kernel in your project run the following command in the VS Code command line:

pip install semantic-kernel

Explore the Semantic Kernel repo and the Getting started documentation.

Set up authentication and initialize the Microsoft Graph Python SDK

The first piece of this application is initializing the Microsoft Graph service.

1. In your project, create an app.py file.
2. Import the ClientSecretCredential library and the AzureIdentityAuthenticationProvider. We'll use these libraries to manage authentication against the Microsoft Graph API. To explore other ways to authenticate, visit the Microsoft Graph Python SDK repo.
3. Import the GraphRequestAdapter and the GraphServiceClient. We'll use this to manage the communications with the Microsoft Graph API.
4. Set up your authentication by initializing an instance of the ClientSecretCredential library and an instance of the GraphServiceClient.

importasynciofromazure.identity.aioimportClientSecretCredentialfromkiota_authentication_azure.azure_identity_authentication_providerimportAzureIdentityAuthenticationProviderfrommsgraphimportGraphRequestAdapterfrommsgraphimportGraphServiceClientasyncio.set_event_loop_policy(asyncio.WindowsSelectorEventLoopPolicy())
# Create auth proviver object. Used to authenticate requestcredential=ClientSecretCredential("tenantID",
"clientID",
"cientSecret")
scopes= ['https://graph.microsoft.com/.default']
auth_provider=AzureIdentityAuthenticationProvider(credential, scopes=scopes)
# Initialize a request adapter. Handles the HTTP concernsrequest_adapter=GraphRequestAdapter(auth_provider)
# Get a service clientclient=GraphServiceClient(request_adapter)

Get the emails of a user

Now we are ready to read the emails of the selected user.

importasynciofromazure.identity.aioimportClientSecretCredentialfromkiota_authentication_azure.azure_identity_authentication_providerimportAzureIdentityAuthenticationProviderfrommsgraphimportGraphRequestAdapterfrommsgraphimportGraphServiceClientasyncio.set_event_loop_policy(asyncio.WindowsSelectorEventLoopPolicy())
# Create auth proviver object. Used to authenticate requestcredential=ClientSecretCredential("tenantID",
"clientID",
"clientSecret")
scopes= ['https://graph.microsoft.com/.default']
auth_provider=AzureIdentityAuthenticationProvider(credential, scopes=scopes)
# Initialize a request adapter. Handles the HTTP concernsrequest_adapter=GraphRequestAdapter(auth_provider)
# Get a service clientclient=GraphServiceClient(request_adapter)
# GET emails from userasyncdefget_user_messages():
"""Getting the messages of a user"""try:
messages=awaitclient.users_by_id("AlexW@contoso.com").messages.get()
formsginmessages.value:
print(
msg.subject,
msg.id,
)
exceptExceptionase_rr:
print(f'Error: {e_rr.error.message}')
asyncio.run(get_user_messages())

Extract action items from emails

Placeholder.

Create a new ToDo list

With the action items ready, we can create a new ToDo list to add our tasks.

# Create ToDo listfrommsgraph.generated.models.todo_task_listimportTodoTaskListasyncdefcreate_todo_list():
"""Create a ToDo list"""try:
request_body=TodoTaskList()
request_body.display_name='Action items from emails'result=awaitclient.users_by_id("AlexW@contoso.com").todo.lists.post(request_body)
exceptExceptionase_rr:
print(f'Error: {e_rr.error.message}')
asyncio.run(create_todo_list())

Add tasks to the list

We'll add the action items extracted from the emails as tasks in the new ToDo list.

# Create new tasksfrommsgraph.generated.models.todo_taskimportTodoTaskasyncdefcreate_new_task(item):
"""Create a new task"""try:
request_body=TodoTask()
request_body.title=itemresult=awaitclient.users_by_id("AlexW@M365x86781558.OnMicrosoft.com").todo.lists_by_id(my_list_id).tasks.post(request_body)
exceptExceptionase_rr:
print(f'Error: {e_rr.error.message}')
# placeholder action itemsmy_action_items= [
"Create email", "Review presentation",
"Schedule meeting"
]
foriteminmy_action_items:
asyncio.run(create_new_task(item))