Trace Memory

HybridAGI trace store

The Trace Memory records every step of the interpreter and indexes every action, making it possible for the agent system to recall previous actions, similar to a human's episodic memory. It also allows for the inspection of the agent's trace using graph database inspection tools, enabling efficient debugging by using Cypher to search for information.

Fig.1 - HybridAGI's trace memory indexes each past action, allowing the AI system to recall past actions between sessions.

Usage

from hybridagi import ProgramMemory, TraceMemory
from hybridagi import SentenceTransformerEmbeddings
from hybridagi.tools import (
    PastActionSearch, # Allows the system to search for past actions
)

embeddings = SentenceTransformerEmbeddings(
    dim = 384,
    model_name_or_path = "sentence-transformers/all-MiniLM-L6-v2",
)

program_memory = ProgramMemory(
    index_name = "hybrid_agi", 
    embeddings = embeddings,
)

trace_memory = TraceMemory(
    index_name = "hybrid_agi", # The global index
    embeddings = embeddings, # The embeddings to use
    graph_index = "trace_memory", # The hybridstore index (default to trace_memory)
    hostname = "localhost", # FalkorDB hostname (default to localhost)
    port = 6379, # FalkorDB port (default to 6379)
    username = "", # FalkorDB username (empty by default)
    password = "", # FalkorDB password (empty by default)
    indexed_label = "Content", # The label of the indexed nodes (default to Content)
    wipe_on_start = False, # Whether or not to wipe the hybridstore at start (default to False)
)

tools = [
    PastActionSearch(
        trace_memory = trace_memory
    ),
]

interpreter = GraphProgramInterpreter(
    program_memory = program_memory,
    trace_memory = trace_memory,
    tools = tools,
)