[DRAFT] Implement Vision Capabilities for Azure OpenAI Integration

src/System Application/App/AI/app.json

@@ -100,6 +100,18 @@
       "name": "Client Type Management",
       "publisher": "Microsoft",
       "version": "27.0.0.0"
+    },
+    {
+      "id": "e31ad830-3d46-472e-afeb-1d3d35247943",
+      "name": "BLOB Storage",
+      "publisher": "Microsoft",
+      "version": "27.0.0.0"
+    },
+    {
+      "id": "0846d207-5dec-4c1b-afd8-6a25e1e14b9d",
+      "name": "Base64 Convert",
+      "publisher": "Microsoft",
+      "version": "27.0.0.0"
     }
   ],
   "screenshots": [],

src/System Application/App/AI/src/Azure OpenAI/Chat Completion/AOAIChatMessages.Codeunit.al

@@ -4,6 +4,8 @@
 // ------------------------------------------------------------------------------------------------
 
 namespace System.AI;
+using System.Environment;
+using System.Utilities;
 
 /// <summary>
 /// Helper functions for the AOAI Chat Message table.
@@ -59,6 +61,70 @@ codeunit 7763 "AOAI Chat Messages"
         AOAIChatMessagesImpl.AddUserMessage(NewMessage, NewName);
     end;
 
+    /// <summary>
+    /// Adds a user message containing text and an image from a URL to the chat messages history.
+    /// </summary>
+    /// <param name="UserText">The text part of the message.</param>
+    /// <param name="ImageUrl">The URL of the image (HTTPS recommended).</param>
+    /// <param name="DetailLevel">The detail level for image processing.</param>
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; ImageUrl: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    begin
+        AOAIChatMessagesImpl.AddUserMessage(UserText, ImageUrl, DetailLevel);
+    end;
+
+    /// <summary>
+    /// Adds a user message containing text and an image from a stream to the chat messages history.
+    /// The image stream will be Base64 encoded.
+    /// </summary>
+    /// <param name="UserText">The text part of the message.</param>
+    /// <param name="ImageStream">The InStream containing the image data.</param>
+    /// <param name="FileExtension">The file extension of the image (e.g., 'png', 'jpg') used to determine the MIME type.</param>
+    /// <param name="DetailLevel">The detail level for image processing.</param>
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; var ImageStream: InStream; FileExtension: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    begin
+        AOAIChatMessagesImpl.AddUserMessage(UserText, ImageStream, FileExtension, DetailLevel);
+    end;
+
+    /// <summary>
+    /// Adds a user message containing text and images from a MediaSet to the chat messages history.
+    /// </summary>
+    /// <param name="UserText">The text part of the message.</param>
+    /// <param name="MediaSetId">The Guid of the MediaSet containing the images.</param>
+    /// <param name="DetailLevel">The detail level for image processing.</param>
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; MediaSetId: Guid; DetailLevel: Enum "AOAI Image Detail Level")
+    begin
+        AOAIChatMessagesImpl.AddUserMessage(UserText, MediaSetId, DetailLevel);
+    end;
+
+
+    /// <summary>
+    /// Adds a user message containing text and an image from a Tenant Media record to the chat messages history.
+    /// </summary>
+    /// <param name="UserText">The text part of the message.</param>
+    /// <param name="TenantMedia">The Tenant Media record containing the image.</param>
+    /// <param name="DetailLevel">The detail level for image processing.</param>
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; TenantMedia: Record "Tenant Media"; DetailLevel: Enum "AOAI Image Detail Level")
+    begin
+        AOAIChatMessagesImpl.AddUserMessage(UserText, TenantMedia, DetailLevel);
+    end;
+
+    /// <summary>
+    /// Adds a user message containing text and an image from a Temp Blob to the chat messages history.
+    /// </summary>
+    /// <param name="UserText">The text part of the message.</param>
+    /// <param name="TempBlob">The Temp Blob codeunit containing the image data.</param>
+    /// <param name="FileExtension">The file extension of the image in the Temp Blob (e.g., 'png', 'jpg').</param>
+    /// <param name="DetailLevel">The detail level for image processing.</param>
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; var TempBlob: Codeunit "Temp Blob"; FileExtension: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    begin
+        AOAIChatMessagesImpl.AddUserMessage(UserText, TempBlob, FileExtension, DetailLevel);
+    end;
+
     /// <summary>
     /// Adds a assistant message to the chat messages history.
     /// </summary>

src/System Application/App/AI/src/Azure OpenAI/Chat Completion/AOAIChatMessagesImpl.Codeunit.al

@@ -7,6 +7,7 @@ namespace System.AI;
 
 using System.Azure.KeyVault;
 using System.Environment;
+using System.Utilities;
 using System.Telemetry;
 
 codeunit 7764 "AOAI Chat Messages Impl"
@@ -69,6 +70,76 @@ codeunit 7764 "AOAI Chat Messages Impl"
         AddMessage(NewMessage, NewName, Enum::"AOAI Chat Roles"::User);
     end;
 
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; ImageUrl: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+        ContentAsText: Text;
+    begin
+        Initialize();
+        ContentAsText := AOAIImagesImpl.PrepareUserMessageContentFromUrl(UserText, ImageUrl, DetailLevel);
+        if ContentAsText = '' then
+            exit;
+
+        AddMessage(ContentAsText, '', Enum::"AOAI Chat Roles"::User);
+    end;
+
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; var ImageStream: InStream; FileExtension: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+        ContentAsText: Text;
+    begin
+        Initialize();
+        ContentAsText := AOAIImagesImpl.PrepareUserMessageContentFromStream(UserText, ImageStream, FileExtension, DetailLevel);
+        if ContentAsText = '' then
+            exit;
+
+        AddMessage(ContentAsText, '', Enum::"AOAI Chat Roles"::User);
+    end;
+
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; MediaSetId: Guid; DetailLevel: Enum "AOAI Image Detail Level")
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+        ContentAsText: Text;
+    begin
+        Initialize();
+        ContentAsText := AOAIImagesImpl.PrepareUserMessageContentFromMediaSet(UserText, MediaSetId, DetailLevel);
+        if ContentAsText = '' then
+            exit;
+
+        AddMessage(ContentAsText, '', Enum::"AOAI Chat Roles"::User);
+    end;
+
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; TenantMedia: Record "Tenant Media"; DetailLevel: Enum "AOAI Image Detail Level")
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+        ContentAsText: Text;
+    begin
+        Initialize();
+        ContentAsText := AOAIImagesImpl.PrepareUserMessageContentFromMediaRecord(UserText, TenantMedia, DetailLevel);
+        if ContentAsText = '' then
+            exit;
+
+        AddMessage(ContentAsText, '', Enum::"AOAI Chat Roles"::User);
+    end;
+
+    [NonDebuggable]
+    procedure AddUserMessage(UserText: Text; var TempBlob: Codeunit "Temp Blob"; FileExtension: Text; DetailLevel: Enum "AOAI Image Detail Level")
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+        ContentAsText: Text;
+    begin
+        Initialize();
+        ContentAsText := AOAIImagesImpl.PrepareUserMessageContentFromTempBlob(UserText, TempBlob, FileExtension, DetailLevel);
+        if ContentAsText = '' then
+            exit;
+
+        AddMessage(ContentAsText, '', Enum::"AOAI Chat Roles"::User);
+    end;
+
     [NonDebuggable]
     procedure AddAssistantMessage(NewMessage: Text)
     begin
@@ -239,10 +310,11 @@ codeunit 7764 "AOAI Chat Messages Impl"
                 Message := WrapUserMessages(AzureOpenAIImpl.RemoveProhibitedCharacters(Message))
             else
                 Message := AzureOpenAIImpl.RemoveProhibitedCharacters(Message);
+
             if ToolCalls.Count() > 0 then
                 MessageJsonObject.Add('tool_calls', ToolCalls)
             else
-                MessageJsonObject.Add('content', Message);
+                AddContentToMessage(Message, MessageJsonObject);
 
             if Name <> '' then
                 MessageJsonObject.Add('name', Name);
@@ -258,6 +330,38 @@ codeunit 7764 "AOAI Chat Messages Impl"
         MessagesTokenCount := AOAIToken.GetGPT4TokenCount(TotalMessages);
     end;
 
+    local procedure AddContentToMessage(Message: Text; var MessageJsonObject: JsonObject)
+    begin
+        AddTextContentToMessage(Message, MessageJsonObject);
+        AddImageContentToMessage(Message, MessageJsonObject);
+    end;
+
+    local procedure AddTextContentToMessage(Message: Text; var MessageJsonObject: JsonObject)
+    begin
+        if CheckIfImageContent(Message) then
+            exit;
+
+        MessageJsonObject.Add('content', Message);
+    end;
+
+    local procedure AddImageContentToMessage(Message: Text; var MessageJsonObject: JsonObject)
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+    begin
+        if not CheckIfImageContent(Message) then
+            exit;
+
+        MessageJsonObject.Add('content', AOAIImagesImpl.ReadImageContent(Message));
+    end;
+
+    local procedure CheckIfImageContent(Message: Text): Boolean
+    var
+        AOAIImagesImpl: Codeunit "AOAI Images Impl";
+    begin
+        exit(AOAIImagesImpl.CheckIfImageContent(Message));
+    end;
+
+
     local procedure Initialize()
     begin
         if Initialized then

src/System Application/App/AI/src/Azure OpenAI/Chat Completion/Vision/AOAIImageDetailLevel.Enum.al

@@ -0,0 +1,28 @@
+// ------------------------------------------------------------------------------------------------
+// Copyright (c) Microsoft Corporation. All rights reserved.
+// Licensed under the MIT License. See License.txt in the project root for license information.
+// ------------------------------------------------------------------------------------------------
+
+namespace System.AI;
+
+/// <summary>
+/// Specifies the detail level for image processing in vision-enabled models.
+/// </summary>
+enum 7770 "AOAI Image Detail Level"
+{
+    Extensible = true;
+    Access = Public;
+
+    value(0; low)
+    {
+        Caption = 'low';
+    }
+    value(1; high)
+    {
+        Caption = 'high';
+    }
+    value(2; auto)
+    {
+        Caption = 'auto';
+    }
+}