monkey-brain

Small Go microservice for deterministic memory and logic reasoning using Prolog-style triples.

What it does

• Stores facts as triple(subject, relation, target)
• Supports rule-based inference (Horn clauses)
• Exposes HTTP endpoints on http://localhost:7777
• Keeps in-memory Lamport-style temporal metadata for incremental updates

Endpoints

• POST /learn - assert facts (array of triples)
• POST /forget - retract facts (single triple pattern)
• POST /pattern - define inference rules
• POST /ask - query facts/rules with limit + offset

/ask responses are deduped before pagination.

Run

go run .

Server listens on:

http://localhost:7777

Quick example

Learn facts:

curl -X POST http://localhost:7777/learn \
 -H "Content-Type: application/json" \
 -d '[
 {"subject":"alice","relation":"parent","target":"bob"},
 {"subject":"bob","relation":"parent","target":"carol"}
 ]'

Define rules:

curl -X POST http://localhost:7777/pattern \
 -H "Content-Type: application/json" \
 -d '{
 "name":"ancestor_base",
 "conditions":[{"subject":"?A","relation":"parent","target":"?B"}],
 "infers":{"subject":"?A","relation":"ancestor","target":"?B"}
 }'
curl -X POST http://localhost:7777/pattern \
 -H "Content-Type: application/json" \
 -d '{
 "name":"ancestor_recursive",
 "conditions":[
 {"subject":"?A","relation":"parent","target":"?B"},
 {"subject":"?B","relation":"ancestor","target":"?C"}
 ],
 "infers":{"subject":"?A","relation":"ancestor","target":"?C"}
 }'

Ask:

curl -X POST http://localhost:7777/ask \
 -H "Content-Type: application/json" \
 -d '{"query":{"subject":"alice","relation":"ancestor","target":"?X"},"limit":50,"offset":0}'