🧠 Preference Graph

SLM Memory Manager

Manages long-term personal state and preference graphs. Learns and adapts to user query patterns locally without cloud synchronization.

🚀 Overview & Capabilities

Manages long-term personal state and preference graphs. Learns and adapts to user query patterns locally without cloud synchronization.

Key Features

Entities and relations extraction from chat history
Builds a local knowledge graph of user preferences
Prunes older irrelevant details to fit within context limits
Auto-injects user context tags into RAG sessions

💻 Installation

Install the local CPU-optimized package using pip:

Terminal

              # Install local CPU-optimized package
pip install slm-memory
            

🐙 Checkout from GitHub

Clone only this agent's folder from the monorepo using Git sparse-checkout — no need to download the full repository:

Option 1 — Sparse Checkout (Recommended)

Terminal — Git Sparse Checkout

              # 1. Create and enter a new directory
$ mkdir slm_memory && cd slm_memory

# 2. Initialise empty git repo and add remote
$ git init
$ git remote add origin https://github.com/t00114218-stack/SLMAgents.git

# 3. Enable sparse-checkout and set target folder
$ git sparse-checkout init --cone
$ git sparse-checkout set slm_memory

# 4. Pull only that agent's source
$ git pull origin main
            

Option 2 — Full Repository Clone

Terminal — Full Clone

              $ git clone https://github.com/t00114218-stack/SLMAgents.git
$ cd SLMAgents/slm_memory
            

💡 Tip: After checkout, install the package locally with pip install -e ./slm_memory to run in editable mode without publishing to PyPI.

⚙️ Configuration API

Constructor Parameters

Instantiate SLMMemoryManager with performance options:

Parameter	Type / Default	Description
model_path	str \| None	Explicit path to ONNX model weights. If omitted, downloads standard checkpoints.
cache_dir	str \| None	Directory to store model weights offline. Defaults to `~/.cache/slm-memory/`. Also settable via SLM_MEMORY_MANAGER_CACHE_DIR.
n_threads	int \| 4	CPU thread count for ONNX inference. Optimize for CPU core count. Also settable via SLM_MEMORY_MANAGER_N_THREADS.

Methods

Method Signature	Return Type	Description
`store_fact(fact_text)`	`None`	Saves semantic user parameters offline to local storage.
`get_relevant_facts(query, system_prompt=None, user_input=None)`	`list[str]`	Queries SQLite embedding tables to retrieve contextual preference strings.

Method Parameters (Execution Customization)

All main execution methods accept optional system routing parameters:

Parameter	Type / Default	Description
system_prompt	str \| None	Optional custom system prompt instruction to override the default system template response parameters.
user_input	str \| None	Optional additional user-supplied target text variables or contextual keys.

Quick Start

from slm_memory import SLMMemoryManager

mem = SLMMemoryManager()
mem.store_fact("User prefers python code examples.")
print(mem.get_relevant_facts(
    "code preferences",
    system_prompt="Prioritize code formatting details",
    user_input="Sort by recency"
))

Environment Variables

Configure agent parameters globally using environment values:

Environment Variable	Default	Purpose
SLM_MEMORY_MANAGER_N_THREADS	4	Sets CPU inference execution threads.
SLM_MEMORY_MANAGER_CACHE_DIR	~/.cache/slm-memory/	Default directory to store downloaded ONNX weights.

CPU Performance Tuning

To run the SLMMemoryManager engine efficiently on CPU under 1.5 GB memory footprint:

Match Threads to Core Count: Set n_threads or SLM_MEMORY_MANAGER_N_THREADS to match the physical CPU core count.
Sequential Processing: Avoid concurrent processing when batch files are large.
Garbage Collection: Clear variables and run gc.collect() to release model RAM blocks after execution.

Verified Input & Output Logs

Diagnostic execution console response running locally on CPU:

→ INPUT (Store Fact):
"User prefers python code examples."

← OUTPUT (Fact Retrieval):
[
  'User prefers python code examples.'
]