made changes to memory namespace

app/agents/memory_manager_agent.py

@@ -46,7 +46,7 @@ def patch_groq_for_extractions(model: ChatGroq):
 model=ChatGroq(model="openai/gpt-oss-20b", temperature=0.2)
 model = patch_groq_for_extractions(model)
 
-namespace = ("emails", "{user_id}", "collection")
+namespace = ("emails", "{user_id}", "{sender_email_id}")
 memory_manager_agent = create_memory_store_manager(
     model, 
     schemas=[EmailMemory],

app/main.py

@@ -15,6 +15,7 @@ from sqlalchemy.orm import Session
 from app.database.connection import SessionLocal
 from fastapi import Request
 from app.database.models import User
+from app.utils.email_encode  import encode_email_for_namespace
 from app.core.auth import create_access_token,get_current_user
 import traceback
 
@@ -53,6 +54,9 @@ class SendEmailRequest(BaseModel):
     human_message: str
 # --- Helper Functions ---
 
+
+
+
 def parse_interrupt(final_state: Dict[str, Any]) -> Optional[Dict[str, Any]]:
     """Parse interrupt from graph state."""
     if "__interrupt__" not in final_state:
@@ -98,7 +102,8 @@ def process_email(request: EmailProcessRequest, db: Session = Depends(get_sessio
         config = {
             "configurable": {
                 "thread_id": thread_id,
-                "user_id": str(current_user.id)
+                "user_id": str(current_user.id),
+                "sender_email_id": encode_email_for_namespace(request.sender_email_id ),
             }
         }
         

app/prompts/context_agent_prompt.py

@@ -3,28 +3,31 @@ from langchain_core.prompts import ChatPromptTemplate
 
 context_agent_template = ChatPromptTemplate([
     ("system", """
-You are a context retrieval agent for {user_name}.
+ROLE: Context Retrieval Agent for {user_name}.
+MISSION: Retrieve only the most critical facts from memory to support a reply.
 
-Your job is to search past memory for relevant background on an incoming email and return a concise summary.
+WORKFLOW:
+1. EXTRACT: Identify 1-2 core technical entities or topics requiring verification (e.g., "backbone", "encryption key").
+2. SEARCH: Use `search_memory_tool` with short, high-entropy keywords. 
+3. SYNTHESIZE: Call `give_previous_context` with a concise summary. If no match, return: "No relevant past context found."
 
-STEPS:
-1. Identify what facts would help reply — prior commitments, open questions, shared context.
-2. Search using `search_sender_memory_tool` with specific queries. Run multiple searches if needed.
-3. Call `give_previous_context` with a brief factual summary. If nothing relevant found, pass exactly: "No relevant past context found."
+CONSTRAINTS:
+- Keep queries < 5 words.
+- Max 2 search calls to save tokens.
+- Do NOT repeat email content in queries.
 
 EXAMPLE:
-Email — Subject: "Updated proposal?" Body: "Hey, did you ever send the revised pricing proposal we discussed?"
-
-search_sender_memory_tool("pricing proposal")
-→ "User sent Alice a revised SaaS pricing proposal on June 3rd, pending her approval."
-
-give_previous_context("Alice is waiting on a revised pricing proposal sent June 3rd.")
+Email — Subject: "Model Update?" Body: "What is the CNN backbone for the NeuroAssist project?"
+- Query 1: "NeuroAssist CNN backbone"
+- Result: "Team using ResNet-50 for NeuroAssist."
+- Brief: "The NeuroAssist CNN model uses a ResNet-50 backbone."
 """),
     ("human", """
+[CONTEXT]
 Sender: {senders_email}
-Subject: {subject}
+Topic: {subject}
 Body: {body}
 
-Search memory and return any relevant past context.
+Action: Retrieve relevant context and provide a concise summary.
 """),
 ])

app/utils/email_encode.py

@@ -0,0 +1,7 @@
+
+import base64
+
+
+def encode_email_for_namespace(email: str) -> str:
+    """Encodes email to a safe string without periods."""
+    return base64.b32encode(email.lower().encode("utf-8")).decode("utf-8").replace("=", "")