Feat: Add extensive diagnostic logging for chat history transformation

To diagnose potential context loss leading to reduced AI performance (observed as 'history size: 2'), this commit introduces detailed logging throughout the chat history processing lifecycle:

1.  **ScreenCaptureService:**
    *   Logs the structure of the raw AI response received, including details of any `FunctionCallPart`s.

2.  **PhotoReasoningViewModel:**
    *   `updateAiMessage` and your message creation in `performReasoning`: Logs details of messages being added to the UI-facing `_chatState.messages`.
    *   `rebuildChatHistory`:
        *   Logs the size of the input `_chatState.messages` list.
        *   Logs how each `PhotoReasoningMessage` is processed into text components for SDK `Content`.
        *   Logs the size and detailed structure (parts and their types) of the final `List<SDK.Content>` it produces.
    *   `sendMessageWithRetry`:
        *   Logs the size and detailed structure of `chat.history` (the output of `rebuildChatHistory`) just before it's serialized with DTOs and sent to `ScreenCaptureService`.

This logging aims to provide a clear trace to identify at what stage (AI response -> UI message store -> SDK history rebuild -> pre-serialization history) any discrepancies or loss of complex `Part` information might be occurring.

Author: google-labs-jules[bot]

app/src/main/kotlin/com/google/ai/sample/ScreenCaptureService.kt

@@ -28,6 +28,7 @@ import android.widget.Toast
 import com.google.ai.client.generativeai.GenerativeModel
 import com.google.ai.client.generativeai.type.Content
 import com.google.ai.client.generativeai.type.ImagePart // For instance check
+import com.google.ai.client.generativeai.type.FunctionCallPart // For logging AI response
 import com.google.ai.sample.feature.multimodal.dtos.ContentDto
 import com.google.ai.sample.feature.multimodal.dtos.toSdk
 import kotlinx.coroutines.CoroutineScope
@@ -248,6 +249,11 @@ class ScreenCaptureService : Service() {
                         val tempChat = generativeModel.startChat(history = chatHistory) // Use the mapped SDK history
                         Log.d(TAG, "Executing AI sendMessage with history size: ${chatHistory.size}")
                         val aiResponse = tempChat.sendMessage(inputContent) // Use the mapped SDK inputContent
+                        Log.d(TAG, "Service received AI Response. Parts: ${aiResponse.parts.joinToString { it.javaClass.simpleName }}")
+                        // If a part is FunctionCallPart, log its name and args
+                        aiResponse.parts.filterIsInstance<com.google.ai.client.generativeai.type.FunctionCallPart>().forEach { fcp ->
+                            Log.d(TAG, "  AI sent FunctionCallPart: name='${fcp.name}', args='${fcp.args}'")
+                        }
                         responseText = aiResponse.text
                         Log.d(TAG, "AI call successful. Response text available: ${responseText != null}")
 

app/src/main/kotlin/com/google/ai/sample/feature/multimodal/PhotoReasoningViewModel.kt

@@ -236,7 +236,9 @@ class PhotoReasoningViewModel(
             imageUris = imageUrisForChat ?: emptyList(), // Use the new parameter here
             isPending = false
         )
+        Log.d(TAG, "performReasoning: Adding user message to _chatState. Text: \"${userMessage.text.take(100)}...\", Images: ${userMessage.imageUris.size}")
         _chatState.addMessage(userMessage)
+        Log.d(TAG, "performReasoning: _chatState now has ${_chatState.messages.size} messages.")
         _chatMessagesFlow.value = chatMessages
 
         // Add AI message with pending status
@@ -379,6 +381,20 @@ class PhotoReasoningViewModel(
             // Add @file:UseSerializers(ContentSerializer::class, PartSerializer::class etc.) if needed at top of file
             // Or create DTOs. For this subtask, we'll assume direct serialization is possible.
 
+            Log.d(TAG, "sendMessageWithRetry: Preparing to serialize chat.history. Current chat.history size: ${chat.history.size}")
+            chat.history.forEachIndexed { index, content ->
+                Log.d(TAG, "  ViewModel chat.history Content[$index]: role=${content.role}, parts=${content.parts.joinToString { it.javaClass.simpleName } { part ->
+                    when(part) {
+                        is com.google.ai.client.generativeai.type.TextPart -> "Text(\"${part.text.take(50)}...\")"
+                        is com.google.ai.client.generativeai.type.ImagePart -> "Image"
+                        is com.google.ai.client.generativeai.type.BlobPart -> "Blob(${part.mimeType})"
+                        is com.google.ai.client.generativeai.type.FunctionCallPart -> "FunctionCall(${part.name}, args=${part.args})"
+                        is com.google.ai.client.generativeai.type.FunctionResponsePart -> "FunctionResponse(${part.name})"
+                        else -> "UnknownPart"
+                    }
+                }}")
+            }
+
             Log.d(TAG, "sendMessageWithRetry: Logging original Bitmap properties before DTO conversion and saving:")
 
             // Log properties for inputContent's images
@@ -455,6 +471,7 @@ class PhotoReasoningViewModel(
      * Update the AI message in chat history
      */
     private fun updateAiMessage(text: String, isPending: Boolean = false) {
+        Log.d(TAG, "updateAiMessage: Adding to _chatState. Text: \"${text.take(100)}...\", Participant: MODEL, IsPending: $isPending")
         // Find the last AI message and update it or add a new one if no suitable message exists
         val messages = _chatState.messages.toMutableList()
         val lastAiMessageIndex = messages.indexOfLast { it.participant == PhotoParticipant.MODEL }
@@ -481,6 +498,7 @@ class PhotoReasoningViewModel(
 
         // Update the flow
         _chatMessagesFlow.value = chatMessages
+        Log.d(TAG, "updateAiMessage: _chatState now has ${_chatState.messages.size} messages.")
 
         // Save chat history after updating message
         // Only save if the operation wasn't stopped, or if it's a deliberate update after stopping
@@ -634,6 +652,7 @@ class PhotoReasoningViewModel(
      * Rebuild the chat history for the AI based on the current messages
      */
     private fun rebuildChatHistory(context: Context) { // Added context parameter
+        Log.d(TAG, "rebuildChatHistory: Starting. Input _chatState.messages size: ${_chatState.messages.size}")
         // Convert the current chat messages to Content objects for the chat history
         val history = mutableListOf<Content>()
 
@@ -687,6 +706,7 @@ class PhotoReasoningViewModel(
                     continue
                 }
             }
+            Log.d(TAG, "  Processed PhotoReasoningMessage: role=${message.participant}, text collected for current SDK Content part: \"${ (if (message.participant == PhotoParticipant.USER) currentUserContent else currentModelContent).take(100) }...\"")
         }
 
         // Add any remaining content
@@ -697,6 +717,20 @@ class PhotoReasoningViewModel(
             history.add(content(role = "model") { text(currentModelContent) })
         }
 
+        Log.d(TAG, "rebuildChatHistory: Finished processing. Generated SDK history size: ${history.size}")
+        history.forEachIndexed { index, content ->
+            Log.d(TAG, "  Generated SDK.Content[$index]: role=${content.role}, parts=${content.parts.joinToString { it.javaClass.simpleName } { part ->
+                when(part) {
+                    is com.google.ai.client.generativeai.type.TextPart -> "Text(\"${part.text.take(50)}...\")"
+                    is com.google.ai.client.generativeai.type.ImagePart -> "Image" // Bitmaps are not easily loggable
+                    is com.google.ai.client.generativeai.type.BlobPart -> "Blob(${part.mimeType})"
+                    is com.google.ai.client.generativeai.type.FunctionCallPart -> "FunctionCall(${part.name}, args=${part.args})"
+                    is com.google.ai.client.generativeai.type.FunctionResponsePart -> "FunctionResponse(${part.name})"
+                    else -> "UnknownPart"
+                }
+            }}")
+        }
+
         // Create a new chat with the rebuilt history
         if (history.isNotEmpty()) {
             chat = generativeModel.startChat(