Skip to content

Content updates #5

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Nov 18, 2024
Merged

Content updates #5

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b6be664
intro updates
im-samz Nov 18, 2024
c116bfa
lab1 - terminology + instructions
im-samz Nov 18, 2024
90d8906
lab2 - terminology
im-samz Nov 18, 2024
9467f98
call out new binding + next steps to visit the booth
im-samz Nov 18, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 42 additions & 28 deletions docs/workshop.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ navigation_levels: 3

# Build serverless intelligent Apps with Azure Functions Flex Consumption and extension for OpenAI

Welcome to this Azure Functions Workshop. You'll be experimenting with Azure Functions service in multiple labs to achieve a real world scenario. You will use the Azure Functions Flex consumption plan for all of these labs which contains the latest features of Azure Functions. Don't worry, this is a step by step lab, you will be guided through the whole process.
Welcome to this Azure Functions Workshop. You'll be experimenting with the Azure Functions service in multiple labs to achieve a real world scenario. For all of these labs, you will use Functions' new [Flex Consumption plan](https://learn.microsoft.com/azure/azure-functions/flex-consumption-plan), which is now in general availability at Ignite. Don't worry, this is a step by step lab, and you will be guided through the whole process.

During this workshop you will have the instructions to complete each steps. The solutions are placed under the 'Toggle solution' panel.

Expand All @@ -38,14 +38,14 @@ The goal of the full lab is to upload an audio file to Azure and save the transc

![Hand's On Lab Architecture](assets/architecture-overview.png)

1. The first Azure Function (standard function) will be mainly responsible for uploading the audio file to the Storage Account.
1. The first Azure Function app will be mainly responsible for uploading the audio file to the Storage Account.
1. Whenever a blob is uploaded to the Storage Account, a `BlobCreated` event will be emitted to Event Grid
1. The Event Grid System Topic will push the event (in real time) to trigger the Azure Durable Function
1. The Azure Durable Function will start processing the audio file
1. The Azure Durable Function will use the Speech To Text service for audio transcription. It will use the Monitor pattern to check every few seconds if the transcription is done.
1. The Azure Durable Function will retrieve the transcription from the Speech to Text service
1. The Azure Durable Function will use Azure OpenAI to generate a summary of the audio file from the transcription
1. The Azure Durable Function will then store the transcription and its summary in Cosmos DB
1. The Event Grid System Topic will push the event (in real time) to trigger another function app, an Azure Durable Function
1. The durable function app will start processing the audio file
1. The durable function app will use the Speech To Text service for audio transcription. It will use the Monitor pattern to check every few seconds if the transcription is done.
1. The durable function app will retrieve the transcription from the Speech to Text service
1. The durable function app will use Azure OpenAI to generate a summary of the audio file from the transcription
1. The durable function app will then store the transcription and its summary in Cosmos DB

## Get the workshop repository

Expand Down Expand Up @@ -85,7 +85,7 @@ Open this repository with Visual Studio Code.

In this first lab you will setup the environment to make sure everything is working as expected.

The Azure Developer CLI (azd) is an open-source tool that accelerates your path from a local development environment to Azure. It provides a set of developer-friendly commands that map to key stages in your workflow (code, build, deploy, monitor).
The [Azure Developer CLI](https://learn.microsoft.com/azure/developer/azure-developer-cli/overview) (azd) is an open-source tool that accelerates your path from a local development environment to Azure. It provides a set of developer-friendly commands that map to key stages in your workflow (code, build, deploy, monitor).

In this workshop, by default, you will be provided with instructions and solutions using `azd`.

Expand Down Expand Up @@ -147,6 +147,8 @@ azd provision

## Deploy Functions to Azure

In this lab, your function apps are already provisioned in Azure for you. This section will guide you in deploying the project code, the functions, to be hosted on those function apps.

Refresh your azd environment using the following commands:

<div class="important" data-title="Important">
Expand Down Expand Up @@ -174,25 +176,27 @@ AZURE_ENV_NAME="ignite"
AZURE_LOCATION="<YOUR-RESOURCE-LOCATION>"
```

Now you can deploy all your Function Apps using the following command:
Now you can deploy the project code to the function apps by using the following command:

```sh
azd deploy
```

## Application deployment with VS Code

Skip this section if you have deployed the Function Apps using the [Azure Developer CLI (azd)][azd].
Skip this section if you have deployed the project code using the [Azure Developer CLI (azd)][azd].

If you encounter any issues, you can also deploy the functions manually using the [Azure Functions extension for VS Code][vscode-azure-functions-extension] in VS Code:

- Open the Azure extension in VS Code left panel
- Make sure you're signed in to your Azure account
- Open the Function App panel
- Right-click on your function app inside `src/uploader` and select `Deploy to Function App...`
- Select the Function starting with `func-std-`
- Right-click on your function app inside `src/processor` and select `Deploy to Function App...`
- Select the Function starting with `func-drbl-`
- Expand the Function App node
- Right-click on the function app starting with `func-std-` and select `Deploy to Function App`
- In the command palette, select `Browse` and navigate to the `src/uploader` directory.
- This will deploy the `uploader` project files and code to the function app.
- Right-click on the function app starting with `func-drbl-` and select `Deploy to Function App...`
- In the command palette, select `Browse` and navigate to the `src/processor` directory.
- This will deploy the `processor` project files and code to the durable function app.

![Deploy to Function App](assets/function-app-deploy.png)

Expand Down Expand Up @@ -242,11 +246,11 @@ The following sample audio files are provided in the workshop, so feel free to r
- [Azure Functions](assets/audios/AzureFunctions.wav)
- [Microsoft AI](assets/audios/MicrosoftAI.wav)

Go back to the Storage Account and check the `audios` container. You should see the files that you uploaded with your `AudioUpload` Azure Function!
Go back to the Storage Account and check the `audios` container. You should see the files that you uploaded with your `AudioUpload` function!

### Option 2: Upload with Postman

Using [Postman][postman], first go to the Azure Function and inside the function starting with `func-std-`, select `Functions` then `AudioUpload` and select the `Get Function Url` with the `default (function key)`.
Using [Postman][postman], first go to the function app starting with `func-std-`, select `Functions` then `AudioUpload` and select the `Get Function Url` with the `default (function key)`.

![Azure Function url credentials](assets/func-url-credentials.png)

Expand All @@ -261,11 +265,11 @@ Create a POST request and in the row where you set the key to `audio` for instan

![Postman](assets/func-postman.png)

Go back to the Storage Account and check the `audios` container. You should see the files that you uploaded with your `AudioUpload` Azure Function!
Go back to the Storage Account and check the `audios` container. You should see the files that you uploaded with your `AudioUpload` function!

## Managed identities and RBAC

Check the App settings of the uploader function and you should notice that there are no secrets allowing it to access the blob storage.
Check the App settings of the uploader function app and you should notice that there are no secrets allowing it to access the blob storage.

![Upload function app settings](assets/uploader-function-app-settings.png)

Expand All @@ -284,7 +288,7 @@ Using RBAC is a security best practice which helps you manage who has access to

- Open the [Azure Portal][az-portal]
- Go to your resource group inside the subscription
- Select the uploader function (the name starts with `func-std-`)
- Select the uploader function app (the name starts with `func-std-`)
- Under the blade menu on the left, select `Settings` then `Identity`
- Click on `Azure role assignments`
- Locate the role used with the resource type `Application Insights`
Expand All @@ -297,7 +301,7 @@ Using RBAC is a security best practice which helps you manage who has access to

## Lab 1 : Summary

In this Lab, you have deployed multiple Function Apps using `azd` and you have tested the uploader function by uploading an audio file to the `audios` blob storage.
In this Lab, you have deployed multiple function apps using `azd` and you have tested the uploader function by uploading an audio file to the `audios` blob storage.

All communications were secured using managed identities and RBAC.

Expand Down Expand Up @@ -329,7 +333,7 @@ Processing the audio file involves the following actions:

To ensure the execution of all these steps and to orchestrate all of this process, you will need a Durable Function which is already created for you in `src/processor`.

Durable Function is an extension of Azure Functions that lets you write stateful functions in a serverless environment. This extension manages state, checkpoints, and restarts for you.
[Durable Functions](https://learn.microsoft.com/azure/azure-functions/durable/durable-functions-overview) are an extension of Azure Functions that lets you write stateful functions in a serverless environment. This extension manages state, checkpoints, and restarts for you.

## Detect a file upload event

Expand Down Expand Up @@ -424,7 +428,7 @@ azd deploy processor

By now you should have a solution that invoke the execution of an Azure Durable Function responsible for retrieving the audio transcription thanks to a Speech to Text batch processing call. You can try to delete and upload once again the audio file in the storage `audios` container of your Storage Account. You will see the different Activity Functions be called in the Azure Functions logs.

After a few minutes, you should see the transcription of the audio file in the logs of the Azure Function:
After a few minutes, you should see the transcription of the audio file in the logs of the durable function app:

For the `StartTranscription` Activity Function:

Expand Down Expand Up @@ -458,7 +462,9 @@ You have now a connection setup between your Azure Durable Function and the Spee

In this lab you will use Azure Functions to call the Azure OpenAI service to analyze the transcription of the audio file and add some information to the Cosmos DB entry.

You will go back to the Azure Durable Function you did in the previous lab and add a connection to Azure OpenAI to be able to summarize the transcription you saved.
You will go back to the durable function app you worked on in the previous lab and add a connection to Azure OpenAI to be able to summarize the transcription you saved.

This connection to Azure OpenAI will preview the new [Azure Functions OpenAI text completion binding](https://learn.microsoft.com/azure/azure-functions/functions-bindings-openai-textcompletion-input?tabs=isolated-process&pivots=programming-language-csharp), which makes developing with Azure OpenAI much easier.

So the scope of the lab is this one:

Expand Down Expand Up @@ -519,15 +525,15 @@ public static AudioTranscription EnrichTranscription(

## Deploy to Azure

You can now redeploy your `processor` function and upload an audio file to see if the transcription is correctly running and check the logs of your Azure Function to see the different steps of the orchestration running.
You can now redeploy your `processor` function and upload an audio file to see if the transcription is correctly running and check the logs of your durable function app to see the different steps of the orchestration running.

```sh
azd deploy processor
```

## Test the scenario

You can try to delete and upload once again the audio file in the storage `audios` container of your Storage Account. You will see the `EnrichTranscription` Activity Functions be called in the Azure Functions logs:
You can try to delete and upload once again the audio file in the storage `audios` container of your Storage Account. You will see the `EnrichTranscription` Activity functions be called in the function app's logs:

![Enrich Transcription activity function](assets/func-enrich-transcription.png)

Expand All @@ -539,12 +545,20 @@ You should also see a new item created in your Cosmos DB container called **audi

## Lab 3 : Summary

By now you should have a solution that invoke Azure OpenAI to create a summary of the transcription.
By now you should have a solution that invokes Azure OpenAI to create a summary of the transcription.

Congratulations!

Well done on completing the lab at Ignite! Your dedication and hard work have truly paid off. Keep up the great work!

Continue on for a bonus lab to integrate your functions with Azure API Management. If not, visit us at the Serverless booth to learn more about the technologies you tried out today:

- Azure Functions Flex Consumption plan
- Azure Durable Functions
- Azure OpenAI text completion binding
- Azure AI Speech to text
- Azure Functions' triggers and bindings

![](assets/congrats.jpeg)

---
Expand Down