Merge pull request #188 from Azure-Samples/workshop

updates to workshop
Azure-Samples · Sep 24, 2024 · 09e0b39 · 09e0b39
2 parents d7c6baf + 3aa6138
commit 09e0b39
Show file tree

Hide file tree

Showing 6 changed files with 127 additions and 108 deletions.
diff --git a/docs/workshop/WORKSHOP-README.md b/docs/workshop/WORKSHOP-README.md
@@ -2,10 +2,10 @@
 
 ## Welcome to workshop WRK551!
 
-Building Large Language Model (LLM) applications is hard! Companies want to build AI solutions but how can they do this in a reliable, reproducible and observable way?
+Building Large Language Model (LLM) applications is hard! Companies want to build AI solutions, but how can they do this in a reliable, reproducible and observable way?
 
-Challenges with AI:  
-- Getting LLM app to work with various real world inputs 
+### Challenges with AI:  
+- Getting LLM apps to work with various real world inputs 
 - Debugging (local and production) to understand failures
 - Setting up and managing production infrastructure automation
 
@@ -26,14 +26,16 @@ We'll begin by logging in to Azure Developer CLI (azd) and Azure CLI (az).
     azd auth login
     ```
 
-3.  Then sign in with Azure CLI. Again, use the `username` and `password` from the lab manual. 
+3.  Then sign in with Azure CLI. 
+    - Again, use the `username` and `password` from the lab manual. 
     - Press enter to select the default subscription when prompted.
 
     ```shell
     az login --use-device-code
     ```
 
-4.  Get the environment variables for your resource group by running the below command and **Replace `AzureEnvName`** with the value given for it in the lab manual.  
+4.  Get the environment variables for your resource group.
+    - Run the below command, **replacing `AzureEnvName`** with the value given for it in the lab manual. (It should look something like AITOUR12345678)
     - When prompted select `Y` to create the environment
     - Press enter to select the default subscription. 
     - Finally choose the location **Canada East**
@@ -53,7 +55,7 @@ We'll begin by logging in to Azure Developer CLI (azd) and Azure CLI (az).
     bash infra/hooks/roles.sh
     ```
 
-7. Run the postprovision script that will install the needed packages and deploy your app resources.
+7. Run the postprovision script 
     - Make sure to **Replace `AzureEnvName`** with the value given for it in the lab manual. 
     - This might take a while but should not take more than 5mins.  
 

diff --git a/docs/workshop/researcher/researcher-0.prompty b/docs/workshop/researcher/researcher-0.prompty
@@ -12,7 +12,7 @@ model:
     azure_deployment: gpt-35-turbo
     api_version: 2023-07-01-preview
 sample:
-  instructions: Can you generate queries to find the latest camping trends and what folks are doing in the winter? Use 'en-US' as the market code. 
+  instructions: Can you generate queries to find the latest camping trends and what folks are doing in the winter? Use 'es-ES' as the market code. 
 ---
 system:
 # Researcher Agent

diff --git a/docs/workshop/researcher/researcher-1.prompty b/docs/workshop/researcher/researcher-1.prompty
@@ -11,8 +11,12 @@ model:
     type: azure_openai
     azure_deployment: gpt-35-turbo
     api_version: 2023-07-01-preview
-sample:
-  instructions: ${file:instructions.json}
+inputs:
+  participant:
+    type: object
+  instructions:
+    type: string
+sample: ${file:instructions.json}
 ---
 system:
 # Researcher Agent

diff --git a/docs/workshop/researcher/researcher-2.prompty b/docs/workshop/researcher/researcher-2.prompty
@@ -14,7 +14,7 @@ model:
   parameters:
     tools: ${file:functions.json}
 sample:
-  instructions: Can you find the best educational material for learning Python programming
+  instructions: Can you find the best educational material for learning Python programming?
 ---
 system:
 # Researcher Agent
@@ -24,11 +24,6 @@ Given some context and writer feedback, you can use these tools to help the writ
 by formulating expert queries and providing the writer with the information they need.
 Your queries should be descriptive and match the provided instructions.
 
-# Feedback
-Use this feedback to help you refine your queries and responses - if there is any feedback:
-
-{{feedback}}
-
 # Market Codes
 The following are the market codes for the countries and regions that are supported by 
 the Microsoft Bing API and should be used when formulating your queries. Use the language 

diff --git a/docs/workshop/researcher/researcher3.py b/docs/workshop/researcher/researcher3.py
@@ -80,7 +80,7 @@ def execute_researcher_prompty(instructions: str):
 
     # Execute the researcher prompty
     function_calls: List[ToolCall] = prompty.execute(
-        "researcher-1.prompty", inputs={"instructions": instructions}
+        "researcher-2.prompty", inputs={"instructions": instructions}
     )
 
     return function_calls