Merge branch 'feature/agent-chat-history-fixes' into 'develop'

Feature/agent chat history fixes

See merge request genaiic-reusable-assets/engagement-artifacts/genaiic-idp-accelerator!414

Author: rstrahan

CHANGELOG.md

@@ -5,6 +5,22 @@ SPDX-License-Identifier: MIT-0
 
 ## [Unreleased]
 
+## [0.4.2]
+
+### Changed
+
+- **Agent Companion Chat - Chat History Feature**
+  - Added chat history feature from Agent Analysis back into Agent Companion Chat
+  - Users can now load and view previous chat sessions with full conversation context
+  - Chat history dropdown displays recent sessions with timestamps and message counts
+
+### Fixed
+
+- **Agent Companion Chat - Session Persistence and input control**
+  - Agent Companion Chat in-session memory now persists even when user changes pages
+  - Prompt input is disabled during active streaming responses to prevent concurrent requests
+  - Fixed issue where charts in loaded chat history were not displaying
+
 ## [0.4.1]
 
 ### Changed

lib/idp_common_pkg/idp_common/agents/analytics/agent.py

@@ -255,6 +255,8 @@ def create_analytics_agent(
     Remember, DO NOT attempt to execute multiple tools in parallel. The input of some tools depend on the output of others. Only ever execute one tool at a time.
     
     Also remember, DO NOT EVER GENERATE SYNTHETIC DATA. Only answer questions or generate plots based on REAL DATA retrieved from databases. If no data can be retrieved, or if there are gaps in the data, do not make up fake data. It is better to show an empty plot or explain you are unable to answer than to make up data.
+
+    If a tool or several tools result in error after 2 times of retry, reply by mentioning the error that has occurred and stop retrying the tool(s). 
     
     Your final response should be directly parsable as json with no additional text before or after. The json should conform to the result format description shown above, with top level key "responseType" being one of "plotData", "table", or "text". You may have to clean up the output of the python code if, for example, it contains extra strings from logging or otherwise. Return only directly parsable json in your final response.
     """

lib/idp_common_pkg/idp_common/agents/orchestrator/agent.py

@@ -274,6 +274,8 @@ async def tool_func(query: str) -> AsyncIterator:
 - For table/plot responses: Return ONLY the JSON with zero additional text
 - Synthesize information from multiple agents when needed
 - Keep responses clear and user-friendly
+- If a subagent or several subagents result in error after 2 times of retry, reply gracefully by mentioning the error that has occurred and STOP retrying the agents. 
+
 """
 
     # Get model ID using configuration helper (checks env vars, config table, then defaults)

src/ui/src/components/agent-chat/AgentChatLayout.jsx

@@ -360,7 +360,7 @@ const AgentChatLayout = ({
                 onChange={handleInputChange}
                 onKeyDown={handleKeyDown}
                 placeholder={placeholder}
-                disabled={isLoading || isLoadingSession}
+                disabled={isLoading || isLoadingSession || waitingForResponse}
                 actionButtonIconName="send"
                 onAction={handlePromptSubmit}
                 minRows={3}

src/ui/src/hooks/use-agent-chat.js

@@ -495,16 +495,32 @@ const useAgentChat = (config = {}) => {
       }
 
       // Convert messages to the format expected by the UI
-      const formattedMessages = messagesToLoad.map((msg, index) => ({
-        role: msg.role,
-        content: msg.content,
-        messageType: 'text',
-        toolUseData: null,
-        isProcessing: false, // Historical messages are never processing
-        sessionId: msg.sessionId,
-        timestamp: msg.timestamp,
-        id: `${msg.timestamp}-${index}`,
-      }));
+      const formattedMessages = messagesToLoad.map((msg, index) => {
+        const baseMessage = {
+          role: msg.role,
+          content: msg.content,
+          messageType: 'text',
+          toolUseData: null,
+          isProcessing: false, // Historical messages are never processing
+          sessionId: msg.sessionId,
+          timestamp: msg.timestamp,
+          id: `${msg.timestamp}-${index}`,
+        };
+
+        // For assistant messages, parse content to extract structured data (charts, tables, etc.)
+        if (msg.role === 'assistant' && msg.content) {
+          const parsedData = parseResponseData(msg.content);
+
+          if (parsedData) {
+            // If we found structured data, add it to the message
+            baseMessage.parsedData = parsedData;
+            // Update content to show only the text portion (without the JSON)
+            baseMessage.content = parsedData.textContent || msg.content;
+          }
+        }
+
+        return baseMessage;
+      });
 
       // Update context with loaded session
       updateAgentChatState({

template.yaml

@@ -249,17 +249,15 @@ Parameters:
 
   ChatCompanionModelId:
     Type: String
-    Default: "us.anthropic.claude-3-7-sonnet-20250219-v1:0"
+    Default: "us.anthropic.claude-sonnet-4-20250514-v1:0"
     AllowedValues:
-      - "us.amazon.nova-lite-v1:0"
       - "us.amazon.nova-pro-v1:0"
       - "us.amazon.nova-premier-v1:0"
       - "us.anthropic.claude-3-haiku-20240307-v1:0"
       - "us.anthropic.claude-3-5-haiku-20241022-v1:0"
       - "us.anthropic.claude-3-5-sonnet-20241022-v2:0"
       - "us.anthropic.claude-3-7-sonnet-20250219-v1:0"
       - "us.anthropic.claude-sonnet-4-20250514-v1:0"
-      - "eu.amazon.nova-lite-v1:0"
       - "eu.amazon.nova-pro-v1:0"
       - "eu.anthropic.claude-3-haiku-20240307-v1:0"
       - "eu.anthropic.claude-3-5-sonnet-20241022-v2:0"
@@ -532,7 +530,7 @@ Metadata:
           - KnowledgeBaseVectorStore
           - KnowledgeBaseModelId
       - Label:
-          default: "Agentic Analysis"
+          default: "Agentic Companion Chat"
         Parameters:
           - ChatCompanionModelId
       - Label: