3.2: Customize To App¶

ITERATION STEP 0: Let's fix the error and customize the defaults!

1. Create `chat-0.prompty`¶

Our final Contoso Chat application is called chat.prompty.

Let's reflect that by copying over the basic.prompty to a starter asset called chat-0.prompty. We can then bump up the version number with each step to get an intuitive sense of the iterative nature of prompt engineering.

Make sure you are in the sandbox/ folder - then run this command!

cp basic.prompty chat-0.prompty

Open or reload the chat-0.prompty file in your editor!

2. Before → Identify Fixes¶

Look at the file contents (below). Let's identify the fixes to make in this iteration:

Update App Metadata: Lines 2-5 should describe the application and developer details.
Update Model Config: Line 11 should replace <your-deployment> with a model name.
Remove Context Lines 17-22 provide sample context that is irrelevant to our app.
Remove Usage Also remove Lines 35-37 since the context variable no longer exists.

chat-0.prompty (before)

---
name: ExamplePrompt
description: A prompt that uses context to ground an incoming question
authors:
- Seth Juarez
model:
api: chat
configuration:
    type: azure_openai
    azure_endpoint: ${env:AZURE_OPENAI_ENDPOINT}
    azure_deployment: <your-deployment>
    api_version: 2024-07-01-preview
parameters:
    max_tokens: 3000
sample:
firstName: Seth
context: >
    The Alpine Explorer Tent boasts a detachable divider for privacy, 
    numerous mesh windows and adjustable vents for ventilation, and 
    a waterproof design. It even has a built-in gear loft for storing 
    your outdoor essentials. In short, it's a blend of privacy, comfort, 
    and convenience, making it your second home in the heart of nature!
question: What can you tell me about your tents?
---

system:
You are an AI assistant who helps people find information. As the assistant, 
you answer questions briefly, succinctly, and in a personable manner using 
markdown and even add some personal flair with appropriate emojis.

# Customer
You are helping {{firstName}} to find answers to their questions.
Use their name to address them in your responses.

# Context
Use the following context to provide a more personalized response to {{firstName}}:
{{context}}

user:
{{question}}

3. After → Apply Fixes¶

For convenience, we've got staged versions that have these fixes completed. Let's copy that over.
```
cp ../docs/workshop/src/1-build/chat-0.prompty .
```

This is what our first iteration looks like. Let's run this next to validate the fixes.

chat-0.prompty (after)

---
name: Contoso Chat Prompt
description: A retail assistant for Contoso Outdoors products retailer.
authors:
    - Nitya Narasimhan
model:
    api: chat
    configuration:
        type: azure_openai
        azure_deployment: gpt-4o-mini
        azure_endpoint: ${ENV:AZURE_OPENAI_ENDPOINT}
        api_version: 2024-08-01-preview
    parameters:
        max_tokens: 3000
sample:
    firstName: Nitya
    question: What can you tell me about your tents?
---

system:
You are an AI assistant who helps people find information. As the assistant, 
you answer questions briefly, succinctly, and in a personable manner using 
markdown and even add some personal flair with appropriate emojis.

# Customer
You are helping {{firstName}} to find answers to their questions.
Use their name to address them in your responses.

# user
{{question}}

4. Run The Prompty¶

Reload chat-0.prompty in the VS Code editor to refresh it.

Run the refreshed prompty (using the play icon or by clicking F5)

You should see a valid response (like this) in the Visual Studio Code terminal

Hello Nitya! 😊

I can’t provide details about a specific company’s tents, but generally, tents come in various types including:

- **Dome Tents:** Great for easy setup and stability.
- **Backpacking Tents:** Lightweight and compact for hiking.
- **Family Tents:** Spacious with room for multiple people.
- **Pop-Up Tents:** Quick to pitch, perfect for casual outings.

They vary by material, weather resistance, and features like vestibules or built-in storage. If you’re looking for something specific, let me know! 🏕️✨

Note: Generative AI models are stochastic, meaning that the same query will not always generate the same response. Your output may look different - but this validates that the previous error is fixed.

CONGRATULATIONS. You have a functioning prompt asset!

3.2: Customize To App¶

1. Create chat-0.prompty¶

2. Before → Identify Fixes¶

3. After → Apply Fixes¶

4. Run The Prompty¶

1. Create `chat-0.prompty`¶