Logos Core: The Semantic Overlay

Version: 0.5 (Bicameral Mind) Status: Operational Prototype Architecture: Hybrid Deterministic/Semantic

1. Project Overview

Logos is a Semantic Operating System overlay for Windows. It introduces a "Meaning Layer" between the user and the file system. Instead of managing files based solely on their location or extension, Logos manages them based on their Context and Intent.

The system now operates with a Bicameral Mind:

1. Reflex (Left Hemisphere): Instant, rule-based execution (e.g., "If file contains 'invoice', move to Finance").
2. Intuition (Right Hemisphere): Vector-based conceptual matching (e.g., "If file feels like a payment confirmation, move to Finance").

---

2. Architecture

/logos-core
│
├── /cortex          (The Brain)
│   ├── interpreter.py     # Parses .logos manifests
│   ├── rules_engine.py    # Strict Logic (Reflex)
│   ├── semantic_engine.py # Fuzzy Logic (Intuition/Vectors)
│   └── actuator.py        # The Hands (File Ops, Logging)
│
├── /sensory         (The Senses)
│   └── watcher.py         # File System Monitor (Watchdog)
│
├── /memory          (The Hippocampus)
│   └── vector_memory.py   # Local Vector Store (all-MiniLM-L6-v2)
│
├── /workspace       (The Membrane)
│   └── project.logos      # The Configuration Manifest
│
├── main.py          # Entry Point
└── requirements.txt # Dependencies

---

3. Installation & Usage

Prerequisites

• Python 3.10+
• Windows OS (tested on win32)

Setup

1. Clone the repository:

   git clone https://github.com/BruinGrowly/Logos-Core
   cd Logos-Core

2. Install dependencies:

   pip install -r requirements.txt

   Note: This includes sentence-transformers (~80MB model download on first run).

3. Run the System:

   python main.py

4. Interact:

   • The system monitors the ./workspace folder.
   • Drop files into this folder to trigger rules defined in project.logos.

---

4. Configuration: The .logos Syntax

The system is controlled by project.logos files placed in monitored directories.

Example Configuration

Manifest: Project_Alpha
Type: Entity

Rules:
  # --- REFLEX RULES (Strict/Fast) ---
  
  # Rule 1: Organization by Keyword
  - Trigger:
      Type: "File_Event"
      Condition: "Filename"
      Contains: ["invoice", "receipt", "bill"]
    Action:
      Type: "Move_File"
      Destination: "./Financial_Docs"

  # Rule 2: Security by Extension
  - Trigger:
      Type: "File_Event"
      Condition: "Extension"
      Is: [".secret", ".key"]
    Action:
      Type: "Log_Event"
      Message: "Secure file detected. Verify encryption."

  # --- INTUITION RULES (Fuzzy/Smart) ---

  # Rule 3: Semantic Matching
  # Matches files like "payment_doc_001.txt" even if they don't say "invoice"
  - Trigger:
      Type: "Semantic_Match"
      Concept: "Payment Confirmation"  # The system calculates the vector for this
      Threshold: 0.4                   # 0.0 to 1.0 similarity score
    Action:
      Type: "Move_File"
      Destination: "./Financial_Docs"

---

5. How It Works (The Pipeline)

When a file enters the workspace:

1. Sensation: watcher.py detects the file creation/modification.
2. Interpretation: interpreter.py reads the local project.logos to understand the rules.
3. Reflex Check: rules_engine.py checks for strict matches (Filename/Extension).
   • If Match: The Actuator executes the action immediately.
4. Intuition Check: If no reflex matches, semantic_engine.py activates.
   • It encodes the filename into a vector.
   • It compares it against the "Concepts" defined in your rules.
   • If Similarity > Threshold: The Actuator executes the action.

---

6. Development Roadmap

• Phase 1: Interpreter (YAML Parsing)
• Phase 2: Watcher (File System Events)
• Phase 3: Reflex Engine (Strict Logic)
• Phase 4: Actuator (File Operations)
• Phase 5: Semantic Brain (Local Vector Embeddings)
• Phase 6: Deep Content Reading (OCR/PDF Parsing)
• Phase 7: Feedback Loop (Reinforcement Learning)

---

Logos Core — Code with Meaning.