Prompts & Memory (maticlib.prompts & maticlib.memory)

Prompts

Use BasePromptTemplate and PromptRegistry to manage string formatting systematically.

from maticlib.prompts.registry import PromptRegistry

prompt = PromptRegistry.get("rag_qa")
formatted = prompt.format(context="Data...", question="What is it?")

Memory Buffers

Maintain conversational history cleanly without blowing up context windows. - ConversationBufferMemory: Stores all messages sequentially. - WindowBufferMemory: A sliding window of the last k messages.

from maticlib.memory.buffer import WindowBufferMemory
from maticlib.messages import HumanMessage, AIMessage

memory = WindowBufferMemory(k=10) # Keep last 10 messages
memory.add_message(HumanMessage(content="Hello!"))
memory.add_message(AIMessage(content="Hi there!"))