minor fixes

Author: MScalopez

learn-pr/wwl-data-ai/get-started-generative-ai-azure-database-postgresql/includes/4-explore-azure-openai-schema.md

@@ -8,9 +8,9 @@ With the `azure_openai` schema, you can call the `create_embeddings()` function,
 
 The function requires you to:
 
-- specify the embedding deployment name,
-- supply the input text or array of text,
-- let the function return the embedding vector as `vector` or `vector[]`.
+- Specify the embedding deployment name.
+- Supply the input text or array of text.
+- Let the function return the embedding vector as `vector` or `vector[]`.
 
 Example:
 

learn-pr/wwl-data-ai/get-started-generative-ai-azure-database-postgresql/includes/6-explore-semantic-operators.md

@@ -1,5 +1,3 @@
-# Work with semantic operators
-
 The `azure_ai` extension includes a small set of semantic operators that allow you to work with generative AI models directly in SQL. These operators help you generate text, evaluate statements, extract structured information, and rank documents. Each operator calls a model you configured in the `azure_ai.settings` table.
 
 The semantic operators are:

learn-pr/wwl-data-ai/get-started-generative-ai-azure-database-postgresql/includes/7-examine-azure-machine-learning-schema.md

@@ -44,57 +44,57 @@ The following example shows the full flow for sending input to an Azure Machine
 
 1. Send the input payload
 
-  ```sql
-  SELECT azure_ml.invoke(
-    '{
-      "input_data": {
-        "columns": [
-          "host_is_superhost",
-          "host_has_profile_pic",
-          "host_identity_verified",
-          "neighbourhood_group_cleansed",
-          "zipcode",
-          "property_type",
-          "room_type",
-          "accommodates",
-          "bathrooms",
-          "bedrooms",
-          "beds"
-        ],
-        "index": [0],
-        "data": [
-          ["False","False","False","Central Area","98122","House","Entire home/apt",4,1.5,3,3]
-        ]
-      }
-    }'::jsonb,
-    'rental-price-deployment'
-  ) AS prediction;
-  ```
-
-  The result is returned as a `jsonb` value that contains the model’s output. You can extract fields from this result using standard PostgreSQL JSON functions.
+    ```sql
+    SELECT azure_ml.invoke(
+        '{
+          "input_data": {
+            "columns": [
+              "host_is_superhost",
+              "host_has_profile_pic",
+              "host_identity_verified",
+              "neighbourhood_group_cleansed",
+              "zipcode",
+              "property_type",
+              "room_type",
+              "accommodates",
+              "bathrooms",
+              "bedrooms",
+              "beds"
+            ],
+            "index": [0],
+            "data": [
+              ["False","False","False","Central Area","98122","House","Entire home/apt",4,1.5,3,3]
+            ]
+          }
+        }'::jsonb,
+        'rental-price-deployment'
+    ) AS prediction;
+    ```
+
+    The result is returned as a `jsonb` value that contains the model's output. You can extract fields from this result using standard PostgreSQL JSON functions.
 
 1. Working with the output
 
-  For example, if the model returns an array of predicted values:
-
-  ```sql
-  WITH result AS (
-    SELECT azure_ml.invoke(
-      '{
-        "input_data": {
-          "columns": ["neighbourhood","bedrooms","bathrooms"],
-          "index": [0],
-          "data": [["Queen Anne", 2, 1]]
-        }
-      }'::jsonb,
-      'rental-price-deployment'
-    ) AS output
-  )
-  SELECT jsonb_array_elements(output) AS predicted_value
-  FROM result;
-  ```
-
-  You can incorporate this solution into views, functions, and reporting queries as needed.
+    For example, if the model returns an array of predicted values:
+
+    ```sql
+    WITH result AS (
+        SELECT azure_ml.invoke(
+            '{
+              "input_data": {
+                "columns": ["neighbourhood","bedrooms","bathrooms"],
+                "index": [0],
+                "data": [["Queen Anne", 2, 1]]
+              }
+            }'::jsonb,
+            'rental-price-deployment'
+        ) AS output
+    )
+    SELECT jsonb_array_elements(output) AS predicted_value
+    FROM result;
+    ```
+
+    You can incorporate this solution into views, functions, and reporting queries as needed.
 
 ## Configure a connection to Azure Machine Learning