minor fixes

Author: MScalopez

learn-pr/wwl-data-ai/design-implement-intelligent-search-with-sql/includes/07-exercise-implement-intelligent-search.md

@@ -6,4 +6,4 @@ In this exercise, you implement different search approaches in an Azure SQL Data
 
 Launch the exercise and follow the instructions.
 
-[![Button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350413&clcid=0x409)
+[![Diagram of button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350413&clcid=0x409)

learn-pr/wwl-data-ai/design-implement-models-embeddings-with-sql/includes/06-exercise-generate-update-embeddings-sql.md

@@ -6,4 +6,4 @@ In this exercise, you create an external model, generate embeddings from text st
 
 Launch the exercise and follow the instructions.
 
-[![Button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350412&clcid=0x409)
+[![Diagram of button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350412&clcid=0x409)

learn-pr/wwl-data-ai/design-implement-rag-with-sql/includes/05-generate-process-rag-responses.md

@@ -113,11 +113,11 @@ You now know each piece of the RAG puzzle. Let's put them together into a single
 
 Here's what the procedure does:
 
-1. Convert the question into an embedding so you can compare it against your product descriptions.
-1. Find the most relevant products using vector distance to compare the question embedding against each product's precomputed description embedding.
-1. Build the prompt with a system message that tells the model to stick to the provided data, and a user message that combines the retrieved products with the original question.
-1. Send everything to Azure OpenAI.
-1. Extract the answer from the response and return it to the caller.
+1. Converts the question into an embedding so you can compare it against your product descriptions.
+1. Finds the most relevant products using vector distance to compare the question embedding against each product's precomputed description embedding.
+1. Builds the prompt with a system message that tells the model to stick to the provided data, and a user message that combines the retrieved products with the original question.
+1. Sends everything to Azure OpenAI.
+1. Extracts the answer from the response and return it to the caller.
 
 ```sql
 CREATE PROCEDURE dbo.AskProductQuestion

learn-pr/wwl-data-ai/design-implement-rag-with-sql/includes/06-exercise-implement-rag-solution.md

@@ -6,4 +6,4 @@ In this exercise, you implement a complete Retrieval Augmented Generation (RAG)
 
 Launch the exercise and follow the instructions.
 
-[![Button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350414&clcid=0x409)
+[![Diagram of button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2350414&clcid=0x409)

learn-pr/wwl-data-ai/implement-cicd-sql-database-projects/includes/04-manage-branching-pull-requests-conflict-resolution.md

@@ -4,9 +4,9 @@ Two developers add columns to the same table on the same day. One pushes to prod
 
 A simple branching strategy for SQL database projects follows three principles:
 
-1. Create a **feature branch** for every change, whether it's new tables, bug fixes, or stored procedure updates.
-2. Merge feature branches into **main** through pull requests.
-3. Always keep main in a deployable state.
+- Creates a **feature branch** for every change, whether it's new tables, bug fixes, or stored procedure updates.
+- Merges feature branches into **main** through pull requests.
+- Always keep main in a deployable state.
 
 When you start a database change, branch off `main`. Your work is isolated from everything else in progress. You modify the relevant `.sql` files, and because each object has its own file, a commit that adds a column to `Customers` touches only `Tables/Customers.sql` and nothing else.
 

learn-pr/wwl-data-ai/implement-cicd-sql-database-projects/includes/08-exercise-implement-cicd-sql-database-projects.md

@@ -6,4 +6,4 @@ In this exercise, you create a SQL database project, build it locally, push it t
 
 Launch the exercise and follow the instructions.
 
-[![Button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2351509&clcid=0x409)
+[![Diagram of button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2351509&clcid=0x409)

learn-pr/wwl-data-ai/optimize-database-performance/includes/06-identify-resolve-blocking-deadlocks.md

@@ -136,10 +136,6 @@ ORDER BY deadlock_time DESC;
 
 The deadlock graph includes three sections. The **victim-list** identifies which transaction was terminated. The **process-list** shows each process involved, including the query text, isolation level, and lock mode. The **resource-list** shows the locked resources and which process owns and waits on each one.
 
-If you save the deadlock graph XML as an `.xdl` file and open it in SQL Server Management Studio (SSMS), you get a visual representation. Processes appear as ovals and locked resources appear as rectangles. The deadlock victim has an "X" drawn across its oval.
-
-![Screenshot of an XDL file opened in SSMS. The deadlock graph is displayed graphically, with processes indicated by ovals and lock resources as rectangles.](../media/deadlock-graph.png)
-
 In Azure SQL Database, you can also configure deadlock alerts through the Azure portal to receive notifications when deadlocks occur.
 
 ### Prevent deadlocks

learn-pr/wwl-data-ai/optimize-database-performance/includes/07-exercise-optimize-query-performance.md

@@ -6,4 +6,4 @@ In this exercise, you investigate slow queries in Azure SQL Database using execu
 
 Launch the exercise and follow the instructions.
 
-[![Button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2351712&clcid=0x409)
+[![Diagram of button to launch exercise.](../media/launch-exercise.png)](https://go.microsoft.com/fwlink/?linkid=2351712&clcid=0x409)