reflect

name	reflect
description	Use the reflect skill to persist memories, claims, and inference nodes into the TerminusDB temporal graph. Covers branch lifecycle, trusted session writes, temporal recall, and the separation between session and inference branches.

Reflect

Persist and reason over long-term knowledge in the temporal graph.

When to use

• Writing memories or claims to a persistent Terminus branch
• Creating or managing session and inference branches
• Querying the temporal graph for trusted or branch-local data
• Verifying the separation between trusted and speculative branches

Quick start

from src.terminus.adapter import TerminusMemoryRepository
from src.terminus.branch_manager import (
    BranchManager,
    session_branch_name,
    inference_branch_name,
)

repo = TerminusMemoryRepository(url="http://localhost:6363")

# Create a session branch
branch = repo.ensure_branch(session_branch_name("sess-1"))

# Write a claim to the session branch
from src.claims.models import Claim
claim = Claim(claim_text="Cert rotation caused auth failures.", claim_type="observation")
repo.write_claim(branch, claim)

# Write a memory
from src.normalization.mapper import ClaimToMemoryMapper
mapper = ClaimToMemoryMapper()
memories = mapper.map_many([claim], session_id="sess-1")
for m in memories:
    m.source_branch = branch
    repo.write_memory(branch, m)


trusted = repo.query_memories(branch=)


inference_br = repo.ensure_branch(inference_branch_name())
nodes = repo.query_inference_nodes(inference_br)