praisonaiagents 0.0.27__py3-none-any.whl → 0.0.29__py3-none-any.whl

praisonaiagents/memory/memory.py

@@ -0,0 +1,928 @@
+import os
+import sqlite3
+import json
+import time
+import shutil
+from typing import Any, Dict, List, Optional, Union, Literal
+import logging
+
+# Set up logger
+logger = logging.getLogger(__name__)
+
+try:
+    import chromadb
+    from chromadb.config import Settings as ChromaSettings
+    CHROMADB_AVAILABLE = True
+except ImportError:
+    CHROMADB_AVAILABLE = False
+    logger.warning("To use memory features, please run: pip install \"praisonaiagents[memory]\"")
+
+try:
+    import mem0
+    MEM0_AVAILABLE = True
+except ImportError:
+    MEM0_AVAILABLE = False
+
+try:
+    import openai
+    OPENAI_AVAILABLE = True
+except ImportError:
+    OPENAI_AVAILABLE = False
+
+
+
+
+class Memory:
+    """
+    A single-file memory manager covering:
+    - Short-term memory (STM) for ephemeral context
+    - Long-term memory (LTM) for persistent knowledge
+    - Entity memory (structured data about named entities)
+    - User memory (preferences/history for each user)
+    - Quality score logic for deciding which data to store in LTM
+    - Context building from multiple memory sources
+
+    Config example:
+    {
+      "provider": "rag" or "mem0" or "none",
+      "use_embedding": True,
+      "short_db": "short_term.db",
+      "long_db": "long_term.db",
+      "rag_db_path": "rag_db",   # optional path for local embedding store
+      "config": {
+        "api_key": "...",       # if mem0 usage
+        "org_id": "...",
+        "project_id": "...",
+        ...
+      }
+    }
+    """
+
+    def __init__(self, config: Dict[str, Any], verbose: int = 0):
+        self.cfg = config or {}
+        self.verbose = verbose
+        
+        # Set logger level based on verbose
+        if verbose >= 5:
+            logger.setLevel(logging.INFO)
+        else:
+            logger.setLevel(logging.WARNING)
+            
+        # Also set ChromaDB and OpenAI client loggers to WARNING
+        logging.getLogger('chromadb').setLevel(logging.WARNING)
+        logging.getLogger('openai').setLevel(logging.WARNING)
+        logging.getLogger('httpx').setLevel(logging.WARNING)
+        logging.getLogger('httpcore').setLevel(logging.WARNING)
+        logging.getLogger('chromadb.segment.impl.vector.local_persistent_hnsw').setLevel(logging.ERROR)
+            
+        self.provider = self.cfg.get("provider", "rag")
+        self.use_mem0 = (self.provider.lower() == "mem0") and MEM0_AVAILABLE
+        self.use_rag = (self.provider.lower() == "rag") and CHROMADB_AVAILABLE and self.cfg.get("use_embedding", False)
+
+        # Short-term DB
+        self.short_db = self.cfg.get("short_db", "short_term.db")
+        self._init_stm()
+
+        # Long-term DB
+        self.long_db = self.cfg.get("long_db", "long_term.db")
+        self._init_ltm()
+
+        # Conditionally init Mem0 or local RAG
+        if self.use_mem0:
+            self._init_mem0()
+        elif self.use_rag:
+            self._init_chroma()
+
+    def _log_verbose(self, msg: str, level: int = logging.INFO):
+        """Only log if verbose >= 5"""
+        if self.verbose >= 5:
+            logger.log(level, msg)
+
+    # -------------------------------------------------------------------------
+    #                          Initialization
+    # -------------------------------------------------------------------------
+    def _init_stm(self):
+        """Creates or verifies short-term memory table."""
+        os.makedirs(os.path.dirname(self.short_db) or ".", exist_ok=True)
+        conn = sqlite3.connect(self.short_db)
+        c = conn.cursor()
+        c.execute("""
+        CREATE TABLE IF NOT EXISTS short_mem (
+            id TEXT PRIMARY KEY,
+            content TEXT,
+            meta TEXT,
+            created_at REAL
+        )
+        """)
+        conn.commit()
+        conn.close()
+
+    def _init_ltm(self):
+        """Creates or verifies long-term memory table."""
+        os.makedirs(os.path.dirname(self.long_db) or ".", exist_ok=True)
+        conn = sqlite3.connect(self.long_db)
+        c = conn.cursor()
+        c.execute("""
+        CREATE TABLE IF NOT EXISTS long_mem (
+            id TEXT PRIMARY KEY,
+            content TEXT,
+            meta TEXT,
+            created_at REAL
+        )
+        """)
+        conn.commit()
+        conn.close()
+
+    def _init_mem0(self):
+        """Initialize Mem0 client for agent or user memory."""
+        from mem0 import MemoryClient
+        mem_cfg = self.cfg.get("config", {})
+        api_key = mem_cfg.get("api_key", os.getenv("MEM0_API_KEY"))
+        org_id = mem_cfg.get("org_id")
+        proj_id = mem_cfg.get("project_id")
+        if org_id and proj_id:
+            self.mem0_client = MemoryClient(api_key=api_key, org_id=org_id, project_id=proj_id)
+        else:
+            self.mem0_client = MemoryClient(api_key=api_key)
+
+    def _init_chroma(self):
+        """Initialize a local Chroma client for embedding-based search."""
+        try:
+            # Create directory if it doesn't exist
+            rag_path = self.cfg.get("rag_db_path", "chroma_db")
+            os.makedirs(rag_path, exist_ok=True)
+
+            # Initialize ChromaDB with persistent storage
+            self.chroma_client = chromadb.PersistentClient(
+                path=rag_path,
+                settings=ChromaSettings(
+                    anonymized_telemetry=False,
+                    allow_reset=True
+                )
+            )
+
+            collection_name = "memory_store"
+            try:
+                self.chroma_col = self.chroma_client.get_collection(name=collection_name)
+                self._log_verbose("Using existing ChromaDB collection")
+            except Exception as e:
+                self._log_verbose(f"Collection '{collection_name}' not found. Creating new collection. Error: {e}")
+                self.chroma_col = self.chroma_client.create_collection(
+                    name=collection_name,
+                    metadata={"hnsw:space": "cosine"}
+                )
+                self._log_verbose("Created new ChromaDB collection")
+
+        except Exception as e:
+            self._log_verbose(f"Failed to initialize ChromaDB: {e}", logging.ERROR)
+            self.use_rag = False
+
+    # -------------------------------------------------------------------------
+    #                      Basic Quality Score Computation
+    # -------------------------------------------------------------------------
+    def compute_quality_score(
+        self,
+        completeness: float,
+        relevance: float,
+        clarity: float,
+        accuracy: float,
+        weights: Dict[str, float] = None
+    ) -> float:
+        """
+        Combine multiple sub-metrics into one final score, as an example.
+
+        Args:
+            completeness (float): 0-1
+            relevance (float): 0-1
+            clarity (float): 0-1
+            accuracy (float): 0-1
+            weights (Dict[str, float]): optional weighting like {"completeness": 0.25, "relevance": 0.3, ...}
+
+        Returns:
+            float: Weighted average 0-1
+        """
+        if not weights:
+            weights = {
+                "completeness": 0.25,
+                "relevance": 0.25,
+                "clarity": 0.25,
+                "accuracy": 0.25
+            }
+        total = (completeness * weights["completeness"]
+                 + relevance   * weights["relevance"]
+                 + clarity     * weights["clarity"]
+                 + accuracy    * weights["accuracy"]
+                )
+        return round(total, 3)  # e.g. round to 3 decimal places
+
+    # -------------------------------------------------------------------------
+    #                           Short-Term Methods
+    # -------------------------------------------------------------------------
+    def store_short_term(
+        self,
+        text: str,
+        metadata: Dict[str, Any] = None,
+        completeness: float = None,
+        relevance: float = None,
+        clarity: float = None,
+        accuracy: float = None,
+        weights: Dict[str, float] = None,
+        evaluator_quality: float = None
+    ):
+        """Store in short-term memory with optional quality metrics"""
+        logger.info(f"Storing in short-term memory: {text[:100]}...")
+        logger.info(f"Metadata: {metadata}")
+        
+        metadata = self._process_quality_metrics(
+            metadata, completeness, relevance, clarity, 
+            accuracy, weights, evaluator_quality
+        )
+        logger.info(f"Processed metadata: {metadata}")
+        
+        # Existing store logic
+        try:
+            conn = sqlite3.connect(self.short_db)
+            ident = str(time.time_ns())
+            conn.execute(
+                "INSERT INTO short_mem (id, content, meta, created_at) VALUES (?,?,?,?)",
+                (ident, text, json.dumps(metadata), time.time())
+            )
+            conn.commit()
+            conn.close()
+            logger.info(f"Successfully stored in short-term memory with ID: {ident}")
+        except Exception as e:
+            logger.error(f"Failed to store in short-term memory: {e}")
+            raise
+
+    def search_short_term(
+        self, 
+        query: str, 
+        limit: int = 5,
+        min_quality: float = 0.0,
+        relevance_cutoff: float = 0.0
+    ) -> List[Dict[str, Any]]:
+        """Search short-term memory with optional quality filter"""
+        self._log_verbose(f"Searching short memory for: {query}")
+        
+        if self.use_mem0 and hasattr(self, "mem0_client"):
+            results = self.mem0_client.search(query=query, limit=limit)
+            filtered = [r for r in results if r.get("score", 1.0) >= relevance_cutoff]
+            return filtered
+            
+        elif self.use_rag and hasattr(self, "chroma_col"):
+            try:
+                from openai import OpenAI
+                client = OpenAI()
+                
+                response = client.embeddings.create(
+                    input=query,
+                    model="text-embedding-3-small"
+                )
+                query_embedding = response.data[0].embedding
+                
+                resp = self.chroma_col.query(
+                    query_embeddings=[query_embedding],
+                    n_results=limit
+                )
+                
+                results = []
+                if resp["ids"]:
+                    for i in range(len(resp["ids"][0])):
+                        metadata = resp["metadatas"][0][i] if "metadatas" in resp else {}
+                        quality = metadata.get("quality", 0.0)
+                        score = 1.0 - (resp["distances"][0][i] if "distances" in resp else 0.0)
+                        if quality >= min_quality and score >= relevance_cutoff:
+                            results.append({
+                                "id": resp["ids"][0][i],
+                                "text": resp["documents"][0][i],
+                                "metadata": metadata,
+                                "score": score
+                            })
+                return results
+            except Exception as e:
+                self._log_verbose(f"Error searching ChromaDB: {e}", logging.ERROR)
+                return []
+        
+        else:
+            # Local fallback
+            conn = sqlite3.connect(self.short_db)
+            c = conn.cursor()
+            rows = c.execute(
+                "SELECT id, content, meta FROM short_mem WHERE content LIKE ? LIMIT ?",
+                (f"%{query}%", limit)
+            ).fetchall()
+            conn.close()
+
+            results = []
+            for row in rows:
+                meta = json.loads(row[2] or "{}")
+                quality = meta.get("quality", 0.0)
+                if quality >= min_quality:
+                    results.append({
+                        "id": row[0],
+                        "text": row[1],
+                        "metadata": meta
+                    })
+            return results
+
+    def reset_short_term(self):
+        """Completely clears short-term memory."""
+        conn = sqlite3.connect(self.short_db)
+        conn.execute("DELETE FROM short_mem")
+        conn.commit()
+        conn.close()
+
+    # -------------------------------------------------------------------------
+    #                           Long-Term Methods
+    # -------------------------------------------------------------------------
+    def _sanitize_metadata(self, metadata: Dict) -> Dict:
+        """Sanitize metadata for ChromaDB - convert to acceptable types"""
+        sanitized = {}
+        for k, v in metadata.items():
+            if v is None:
+                continue
+            if isinstance(v, (str, int, float, bool)):
+                sanitized[k] = v
+            elif isinstance(v, dict):
+                # Convert dict to string representation
+                sanitized[k] = str(v)
+            else:
+                # Convert other types to string
+                sanitized[k] = str(v)
+        return sanitized
+
+    def store_long_term(
+        self,
+        text: str,
+        metadata: Dict[str, Any] = None,
+        completeness: float = None,
+        relevance: float = None,
+        clarity: float = None,
+        accuracy: float = None,
+        weights: Dict[str, float] = None,
+        evaluator_quality: float = None
+    ):
+        """Store in long-term memory with optional quality metrics"""
+        logger.info(f"Storing in long-term memory: {text[:100]}...")
+        logger.info(f"Initial metadata: {metadata}")
+        
+        # Process metadata
+        metadata = metadata or {}
+        metadata = self._process_quality_metrics(
+            metadata, completeness, relevance, clarity,
+            accuracy, weights, evaluator_quality
+        )
+        logger.info(f"Processed metadata: {metadata}")
+        
+        # Generate unique ID
+        ident = str(time.time_ns())
+        created = time.time()
+
+        # Store in SQLite
+        try:
+            conn = sqlite3.connect(self.long_db)
+            conn.execute(
+                "INSERT INTO long_mem (id, content, meta, created_at) VALUES (?,?,?,?)",
+                (ident, text, json.dumps(metadata), created)
+            )
+            conn.commit()
+            conn.close()
+            logger.info(f"Successfully stored in SQLite with ID: {ident}")
+        except Exception as e:
+            logger.error(f"Error storing in SQLite: {e}")
+            return
+
+        # Store in vector database if enabled
+        if self.use_rag and hasattr(self, "chroma_col"):
+            try:
+                from openai import OpenAI
+                client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))  # Ensure API key is correctly set
+                
+                logger.info("Getting embeddings from OpenAI...")
+                logger.debug(f"Embedding input text: {text}")  # Log the input text
+                
+                response = client.embeddings.create(
+                    input=text,
+                    model="text-embedding-3-small"
+                )
+                embedding = response.data[0].embedding
+                logger.info("Successfully got embeddings")
+                logger.debug(f"Received embedding of length: {len(embedding)}")  # Log embedding details
+                
+                # Sanitize metadata for ChromaDB
+                sanitized_metadata = self._sanitize_metadata(metadata)
+                
+                # Store in ChromaDB with embedding
+                self.chroma_col.add(
+                    documents=[text],
+                    metadatas=[sanitized_metadata],
+                    ids=[ident],
+                    embeddings=[embedding]
+                )
+                logger.info(f"Successfully stored in ChromaDB with ID: {ident}")
+            except Exception as e:
+                logger.error(f"Error storing in ChromaDB: {e}")
+        
+        elif self.use_mem0 and hasattr(self, "mem0_client"):
+            try:
+                self.mem0_client.add(text, metadata=metadata)
+                logger.info("Successfully stored in Mem0")
+            except Exception as e:
+                logger.error(f"Error storing in Mem0: {e}")
+
+
+    def search_long_term(
+        self, 
+        query: str, 
+        limit: int = 5, 
+        relevance_cutoff: float = 0.0,
+        min_quality: float = 0.0
+    ) -> List[Dict[str, Any]]:
+        """Search long-term memory with optional quality filter"""
+        self._log_verbose(f"Searching long memory for: {query}")
+        self._log_verbose(f"Min quality: {min_quality}")
+
+        found = []
+
+        if self.use_mem0 and hasattr(self, "mem0_client"):
+            results = self.mem0_client.search(query=query, limit=limit)
+            # Filter by quality
+            filtered = [r for r in results if r.get("metadata", {}).get("quality", 0.0) >= min_quality]
+            logger.info(f"Found {len(filtered)} results in Mem0")
+            return filtered
+
+        elif self.use_rag and hasattr(self, "chroma_col"):
+            try:
+                from openai import OpenAI
+                client = OpenAI()
+                
+                # Get query embedding
+                response = client.embeddings.create(
+                    input=query,
+                    model="text-embedding-3-small"  # Using consistent model
+                )
+                query_embedding = response.data[0].embedding
+                
+                # Search ChromaDB with embedding
+                resp = self.chroma_col.query(
+                    query_embeddings=[query_embedding],
+                    n_results=limit,
+                    include=["documents", "metadatas", "distances"]
+                )
+                
+                results = []
+                if resp["ids"]:
+                    for i in range(len(resp["ids"][0])):
+                        metadata = resp["metadatas"][0][i] if "metadatas" in resp else {}
+                        text = resp["documents"][0][i]
+                        # Add memory record citation
+                        text = f"{text} (Memory record: {text})"
+                        found.append({
+                            "id": resp["ids"][0][i],
+                            "text": text,
+                            "metadata": metadata,
+                            "score": 1.0 - (resp["distances"][0][i] if "distances" in resp else 0.0)
+                        })
+                logger.info(f"Found {len(found)} results in ChromaDB")
+
+            except Exception as e:
+                self._log_verbose(f"Error searching ChromaDB: {e}", logging.ERROR)
+
+        # Always try SQLite as fallback or additional source
+        conn = sqlite3.connect(self.long_db)
+        c = conn.cursor()
+        rows = c.execute(
+            "SELECT id, content, meta, created_at FROM long_mem WHERE content LIKE ? LIMIT ?",
+            (f"%{query}%", limit)
+        ).fetchall()
+        conn.close()
+
+        for row in rows:
+            meta = json.loads(row[2] or "{}")
+            text = row[1]
+            # Add memory record citation if not already present
+            if "(Memory record:" not in text:
+                text = f"{text} (Memory record: {text})"
+            # Only add if not already found by ChromaDB/Mem0
+            if not any(f["id"] == row[0] for f in found):
+                found.append({
+                    "id": row[0],
+                    "text": text,
+                    "metadata": meta,
+                    "created_at": row[3]
+                })
+        logger.info(f"Found {len(found)} total results after SQLite")
+
+        results = found
+
+        # Filter by quality if needed
+        if min_quality > 0:
+            self._log_verbose(f"Found {len(results)} initial results")
+            results = [
+                r for r in results 
+                if r.get("metadata", {}).get("quality", 0.0) >= min_quality
+            ]
+            self._log_verbose(f"After quality filter: {len(results)} results")
+
+        # Apply relevance cutoff if specified
+        if relevance_cutoff > 0:
+            results = [r for r in results if r.get("score", 1.0) >= relevance_cutoff]
+            logger.info(f"After relevance filter: {len(results)} results")
+        
+        return results[:limit]
+
+    def reset_long_term(self):
+        """Clear local LTM DB, plus Chroma or mem0 if in use."""
+        conn = sqlite3.connect(self.long_db)
+        conn.execute("DELETE FROM long_mem")
+        conn.commit()
+        conn.close()
+
+        if self.use_mem0 and hasattr(self, "mem0_client"):
+            # Mem0 has no universal reset API. Could implement partial or no-op.
+            pass
+        if self.use_rag and hasattr(self, "chroma_client"):
+            self.chroma_client.reset()  # entire DB
+            self._init_chroma()         # re-init fresh
+
+    # -------------------------------------------------------------------------
+    #                       Entity Memory Methods
+    # -------------------------------------------------------------------------
+    def store_entity(self, name: str, type_: str, desc: str, relations: str):
+        """
+        Save entity info in LTM (or mem0/rag). 
+        We'll label the metadata type = entity for easy filtering.
+        """
+        data = f"Entity {name}({type_}): {desc} | relationships: {relations}"
+        self.store_long_term(data, metadata={"category": "entity"})
+
+    def search_entity(self, query: str, limit: int = 5) -> List[Dict[str, Any]]:
+        """
+        Filter to items that have metadata 'category=entity'.
+        """
+        all_hits = self.search_long_term(query, limit=20)  # gather more
+        ents = []
+        for h in all_hits:
+            meta = h.get("metadata") or {}
+            if meta.get("category") == "entity":
+                ents.append(h)
+        return ents[:limit]
+
+    def reset_entity_only(self):
+        """
+        If you only want to drop entity items from LTM, you'd do a custom 
+        delete from local DB where meta LIKE '%category=entity%'. 
+        For brevity, we do a full LTM reset here.
+        """
+        self.reset_long_term()
+
+    # -------------------------------------------------------------------------
+    #                       User Memory Methods
+    # -------------------------------------------------------------------------
+    def store_user_memory(self, user_id: str, text: str, extra: Dict[str, Any] = None):
+        """
+        If mem0 is used, do user-based addition. Otherwise store in LTM with user in metadata.
+        """
+        meta = {"user_id": user_id}
+        if extra:
+            meta.update(extra)
+
+        if self.use_mem0 and hasattr(self, "mem0_client"):
+            self.mem0_client.add(text, user_id=user_id, metadata=meta)
+        else:
+            self.store_long_term(text, metadata=meta)
+
+    def search_user_memory(self, user_id: str, query: str, limit: int = 5) -> List[Dict[str, Any]]:
+        """
+        If mem0 is used, pass user_id in. Otherwise fallback to local filter on user in metadata.
+        """
+        if self.use_mem0 and hasattr(self, "mem0_client"):
+            return self.mem0_client.search(query=query, limit=limit, user_id=user_id)
+        else:
+            hits = self.search_long_term(query, limit=20)
+            filtered = []
+            for h in hits:
+                meta = h.get("metadata", {})
+                if meta.get("user_id") == user_id:
+                    filtered.append(h)
+            return filtered[:limit]
+
+    def reset_user_memory(self):
+        """
+        Clear all user-based info. For simplicity, we do a full LTM reset. 
+        Real usage might filter only metadata "user_id".
+        """
+        self.reset_long_term()
+
+    # -------------------------------------------------------------------------
+    #                 Putting it all Together: Task Finalization
+    # -------------------------------------------------------------------------
+    def finalize_task_output(
+        self,
+        content: str,
+        agent_name: str,
+        quality_score: float,
+        threshold: float = 0.7,
+        metrics: Dict[str, Any] = None,
+        task_id: str = None
+    ):
+        """Store task output in memory with appropriate metadata"""
+        logger.info(f"Finalizing task output: {content[:100]}...")
+        logger.info(f"Agent: {agent_name}, Quality: {quality_score}, Threshold: {threshold}")
+        
+        metadata = {
+            "task_id": task_id,
+            "agent": agent_name,
+            "quality": quality_score,
+            "metrics": metrics,
+            "task_type": "output",
+            "stored_at": time.time()
+        }
+        logger.info(f"Prepared metadata: {metadata}")
+        
+        # Always store in short-term memory
+        try:
+            logger.info("Storing in short-term memory...")
+            self.store_short_term(
+                text=content,
+                metadata=metadata
+            )
+            logger.info("Successfully stored in short-term memory")
+        except Exception as e:
+            logger.error(f"Failed to store in short-term memory: {e}")
+        
+        # Store in long-term memory if quality meets threshold
+        if quality_score >= threshold:
+            try:
+                logger.info(f"Quality score {quality_score} >= {threshold}, storing in long-term memory...")
+                self.store_long_term(
+                    text=content,
+                    metadata=metadata
+                )
+                logger.info("Successfully stored in long-term memory")
+            except Exception as e:
+                logger.error(f"Failed to store in long-term memory: {e}")
+        else:
+            logger.info(f"Quality score {quality_score} < {threshold}, skipping long-term storage")
+
+    # -------------------------------------------------------------------------
+    #                 Building Context (Short, Long, Entities, User)
+    # -------------------------------------------------------------------------
+    def build_context_for_task(
+        self,
+        task_descr: str,
+        user_id: Optional[str] = None,
+        additional: str = "",
+        max_items: int = 3
+    ) -> str:
+        """
+        Merges relevant short-term, long-term, entity, user memories
+        into a single text block with deduplication and clean formatting.
+        """
+        q = (task_descr + " " + additional).strip()
+        lines = []
+        seen_contents = set()  # Track unique contents
+
+        def normalize_content(content: str) -> str:
+            """Normalize content for deduplication"""
+            # Extract just the main content without citations for comparison
+            normalized = content.split("(Memory record:")[0].strip()
+            # Keep more characters to reduce false duplicates
+            normalized = ''.join(c.lower() for c in normalized if not c.isspace())
+            return normalized
+
+        def format_content(content: str, max_len: int = 150) -> str:
+            """Format content with clean truncation at word boundaries"""
+            if not content:
+                return ""
+            
+            # Clean up content by removing extra whitespace and newlines
+            content = ' '.join(content.split())
+            
+            # If content contains a memory citation, preserve it
+            if "(Memory record:" in content:
+                return content  # Keep original citation format
+            
+            # Regular content truncation
+            if len(content) <= max_len:
+                return content
+            
+            truncate_at = content.rfind(' ', 0, max_len - 3)
+            if truncate_at == -1:
+                truncate_at = max_len - 3
+            return content[:truncate_at] + "..."
+
+        def add_section(title: str, hits: List[Any]) -> None:
+            """Add a section of memory hits with deduplication"""
+            if not hits:
+                return
+                
+            formatted_hits = []
+            for h in hits:
+                content = h.get('text', '') if isinstance(h, dict) else str(h)
+                if not content:
+                    continue
+                    
+                # Keep original format if it has a citation
+                if "(Memory record:" in content:
+                    formatted = content
+                else:
+                    formatted = format_content(content)
+                
+                # Only add if we haven't seen this normalized content before
+                normalized = normalize_content(formatted)
+                if normalized not in seen_contents:
+                    seen_contents.add(normalized)
+                    formatted_hits.append(formatted)
+            
+            if formatted_hits:
+                # Add section header
+                if lines:
+                    lines.append("")  # Space before new section
+                lines.append(title)
+                lines.append("=" * len(title))  # Underline the title
+                lines.append("")  # Space after title
+                
+                # Add formatted content with bullet points
+                for content in formatted_hits:
+                    lines.append(f" • {content}")
+
+        # Add each section
+        # First get all results
+        short_term = self.search_short_term(q, limit=max_items)
+        long_term = self.search_long_term(q, limit=max_items)
+        entities = self.search_entity(q, limit=max_items)
+        user_mem = self.search_user_memory(user_id, q, limit=max_items) if user_id else []
+
+        # Add sections in order of priority
+        add_section("Short-term Memory Context", short_term)
+        add_section("Long-term Memory Context", long_term)
+        add_section("Entity Context", entities)
+        if user_id:
+            add_section("User Context", user_mem)
+
+        return "\n".join(lines) if lines else ""
+
+    # -------------------------------------------------------------------------
+    #                      Master Reset (Everything)
+    # -------------------------------------------------------------------------
+    def reset_all(self):
+        """
+        Fully wipes short-term, long-term, and any memory in mem0 or rag.
+        """
+        self.reset_short_term()
+        self.reset_long_term()
+        # Entities & user memory are stored in LTM or mem0, so no separate step needed.
+
+    def _process_quality_metrics(
+        self,
+        metadata: Dict[str, Any],
+        completeness: float = None,
+        relevance: float = None,
+        clarity: float = None, 
+        accuracy: float = None,
+        weights: Dict[str, float] = None,
+        evaluator_quality: float = None
+    ) -> Dict[str, Any]:
+        """Process and store quality metrics in metadata"""
+        metadata = metadata or {}
+        
+        # Handle sub-metrics if provided
+        if None not in [completeness, relevance, clarity, accuracy]:
+            metadata.update({
+                "completeness": completeness,
+                "relevance": relevance,
+                "clarity": clarity,
+                "accuracy": accuracy,
+                "quality": self.compute_quality_score(
+                    completeness, relevance, clarity, accuracy, weights
+                )
+            })
+        # Handle external evaluator quality if provided
+        elif evaluator_quality is not None:
+            metadata["quality"] = evaluator_quality
+        
+        return metadata
+
+    def calculate_quality_metrics(
+        self,
+        output: str,
+        expected_output: str,
+        llm: Optional[str] = None,
+        custom_prompt: Optional[str] = None
+    ) -> Dict[str, float]:
+        """Calculate quality metrics using LLM"""
+        logger.info("Calculating quality metrics for output")
+        logger.info(f"Output: {output[:100]}...")
+        logger.info(f"Expected: {expected_output[:100]}...")
+        
+        # Default evaluation prompt
+        default_prompt = f"""
+        Evaluate the following output against expected output.
+        Score each metric from 0.0 to 1.0:
+        - Completeness: Does it address all requirements?
+        - Relevance: Does it match expected output?
+        - Clarity: Is it clear and well-structured?
+        - Accuracy: Is it factually correct?
+
+        Expected: {expected_output}
+        Actual: {output}
+
+        Return ONLY a JSON with these keys: completeness, relevance, clarity, accuracy
+        Example: {{"completeness": 0.95, "relevance": 0.8, "clarity": 0.9, "accuracy": 0.85}}
+        """
+
+        try:
+            # Use OpenAI client from main.py
+            from ..main import client
+            
+            response = client.chat.completions.create(
+                model=llm or "gpt-4o",
+                messages=[{
+                    "role": "user", 
+                    "content": custom_prompt or default_prompt
+                }],
+                response_format={"type": "json_object"},
+                temperature=0.3
+            )
+            
+            metrics = json.loads(response.choices[0].message.content)
+            
+            # Validate metrics
+            required = ["completeness", "relevance", "clarity", "accuracy"]
+            if not all(k in metrics for k in required):
+                raise ValueError("Missing required metrics in LLM response")
+            
+            logger.info(f"Calculated metrics: {metrics}")
+            return metrics
+            
+        except Exception as e:
+            logger.error(f"Error calculating metrics: {e}")
+            return {
+                "completeness": 0.0,
+                "relevance": 0.0,
+                "clarity": 0.0,
+                "accuracy": 0.0
+            }
+
+    def store_quality(
+        self,
+        text: str,
+        quality_score: float,
+        task_id: Optional[str] = None,
+        iteration: Optional[int] = None,
+        metrics: Optional[Dict[str, float]] = None,
+        memory_type: Literal["short", "long"] = "long"
+    ) -> None:
+        """Store quality metrics in memory"""
+        logger.info(f"Attempting to store in {memory_type} memory: {text[:100]}...")
+        
+        metadata = {
+            "quality": quality_score,
+            "task_id": task_id,
+            "iteration": iteration
+        }
+        
+        if metrics:
+            metadata.update({
+                k: v for k, v in metrics.items()  # Remove metric_ prefix
+            })
+            
+        logger.info(f"With metadata: {metadata}")
+        
+        try:
+            if memory_type == "short":
+                self.store_short_term(text, metadata=metadata)
+                logger.info("Successfully stored in short-term memory")
+            else:
+                self.store_long_term(text, metadata=metadata)
+                logger.info("Successfully stored in long-term memory")
+        except Exception as e:
+            logger.error(f"Failed to store in memory: {e}")
+
+    def search_with_quality(
+        self,
+        query: str,
+        min_quality: float = 0.0,
+        memory_type: Literal["short", "long"] = "long",
+        limit: int = 5
+    ) -> List[Dict[str, Any]]:
+        """Search with quality filter"""
+        logger.info(f"Searching {memory_type} memory for: {query}")
+        logger.info(f"Min quality: {min_quality}")
+        
+        search_func = (
+            self.search_short_term if memory_type == "short" 
+            else self.search_long_term
+        )
+        
+        results = search_func(query, limit=limit)
+        logger.info(f"Found {len(results)} initial results")
+        
+        filtered = [
+            r for r in results 
+            if r.get("metadata", {}).get("quality", 0.0) >= min_quality
+        ]
+        logger.info(f"After quality filter: {len(filtered)} results")
+        
+        return filtered