update to strong type agents thought payload

src/FFB/ffb-production.schema.ts

@@ -9,3 +9,19 @@ export interface FFBProduction {
   quantity: number;
   quantityUom: string;
 }
+
+export interface ThoughtPayload {
+  node: string;
+  status: 'processing' | 'completed' | 'error';
+  message?: string;    // Human-readable status (e.g., "Analyzing intent...")
+  context?: any;
+  query?: any;
+  filter?: any;
+  pipeline?: any;
+  results?: any;
+  resultsCount?: number;
+  dataContextLength?: number;
+  input?: any;         // The raw data going IN
+  output?: any;        // The data coming OUT (replacing 'result', 'results', 'pipeline')
+  metadata?: Record<string, any>; // For extra stuff like resultsCount or contextLength
+}

src/FFB/services/ffb-langchain.service.ts

@@ -6,6 +6,7 @@ import { StateGraph, START, END, Annotation } from "@langchain/langgraph";
 import { BaseMessage, HumanMessage, AIMessage } from "@langchain/core/messages";
 import { forwardRef, Inject } from '@nestjs/common';
 import { FFBGateway } from '../ffb.gateway';
+import { ThoughtPayload } from '../ffb-production.schema';
 // State Definition using Annotation
 const AgentState = Annotation.Root({
     messages: Annotation<BaseMessage[]>({
@@ -102,12 +103,14 @@ export class FFBLangChainService {
             reasoning: z.string()
         });
 
-        this.gateway.emitThought(state.socketId, {
+        let payload: ThoughtPayload = {
             node: 'router_node',
             status: 'processing',
             message: 'Analyzing user intent...',
             input: lastMessage
-        });
+        }
+
+        this.gateway.emitThought(state.socketId, payload);
 
         const routerPrompt = `
 You are an Application Router for a production database.
@@ -147,12 +150,14 @@ User Input: "${lastMessage}"
     private async clarifierNode(state: typeof AgentState.State): Promise<Partial<typeof AgentState.State>> {
         const prompt = `User mentioned ${JSON.stringify(state.entityStore)}. Ask them to clarify what they want to know (e.g., total production, specific issues, etc.).`;
 
-        this.gateway.emitThought(state.socketId, {
+        let payload: ThoughtPayload = {
             node: 'clarifier_node',
             status: 'processing',
             message: 'Asking for clarification',
             context: state.entityStore
-        });
+        }
+
+        this.gateway.emitThought(state.socketId, payload);
 
         const response = await this.model.invoke(prompt);
         return {
@@ -180,13 +185,14 @@ User Input: "${lastMessage}"
 
         const results = await this.vectorService.vectorSearch(lastMessage, 5, filter);
 
-        this.gateway.emitThought(state.socketId, {
+        let payload: ThoughtPayload = {
             node: 'vector_search_node',
             status: 'completed',
             query: lastMessage,
             filter: filter,
             resultsCount: results.length
-        });
+        }
+        this.gateway.emitThought(state.socketId, payload);
 
         return {
             actionPayload: { type: 'search', query: lastMessage, results }
@@ -239,6 +245,12 @@ User Input: "${lastMessage}"
 
         const results = await this.vectorService.aggregate(pipeline);
 
+        let payload: ThoughtPayload = {
+            node: `aggregation_node`,
+            status: 'completed',
+            pipeline: pipeline,
+            results: results
+        }
         this.gateway.emitThought(state.socketId, {
             node: 'aggregation_node',
             status: 'completed',
@@ -263,12 +275,13 @@ User Input: "${lastMessage}"
         Cite the source (e.g., "Based on aggregation results...").
         `;
 
-        this.gateway.emitThought(state.socketId, {
+        let thoughtPayload: ThoughtPayload = {
             node: 'synthesis_node',
             status: 'processing',
             message: 'Synthesizing final response',
             dataContextLength: JSON.stringify(payload).length
-        });
+        }
+        this.gateway.emitThought(state.socketId, thoughtPayload);
 
         const response = await this.model.invoke(prompt);
         return {