chore: Streamline repository to focused 2-step local launcher and clean up disk

ControlAI_Colab_Demo.ipynb

@@ -1,103 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "# ControlAI: Open-Source AI Agent for Control Systems Engineering\n",
-    "\n",
-    "[![GitHub Repo](https://img.shields.io/badge/GitHub-atakankahya%2Fcontrolai--agent-blue?logo=github)](https://github.com/atakankahya/controlai-agent)\n",
-    "[![Hugging Face Model](https://img.shields.io/badge/Hugging%20Face-ControlAI--Agent-blue)](https://huggingface.co/atakankahya/ControlAI-Agent)\n",
-    "[![License](https://img.shields.io/badge/License-MIT-green.svg)](https://opensource.org/licenses/MIT)\n",
-    "\n",
-    "Welcome to the official **ControlAI** interactive notebook! You can test control engineering problems, LQR synthesis, Bode plots, and ODE simulations directly in Google Colab.\n",
-    "\n",
-    "---"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Step 1: Install Dependencies"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Clone repo & install all required dependencies\n",
-    "!git clone https://github.com/atakankahya/controlai-agent.git\n",
-    "%cd controlai-agent\n",
-    "!pip install -q -r requirements.txt pypdf\n",
-    "print(\"Dependencies installed successfully!\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Step 2: Initialize the ControlAI Agent"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import sys\n",
-    "from pathlib import Path\n",
-    "from IPython.display import display, Markdown, Image\n",
-    "import matplotlib.pyplot as plt\n",
-    "\n",
-    "from controlai_agent.orchestrator import ControlAIAgent\n",
-    "\n",
-    "print(\"Loading ControlAI Engine with Qwen 3B...\")\n",
-    "agent = ControlAIAgent(model_path=\"Qwen/Qwen2.5-3B-Instruct\")\n",
-    "print(\"ControlAI Engine ready!\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Step 3: Ask Any Control Engineering Question\n",
-    "You can edit the prompt below to test your own dynamical systems, transfer functions, or optimal control problems."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Enter your engineering query here:\n",
-    "prompt = \"Design an LQR controller for A=[[0, 1], [-2, -3]], B=[[0], [1]], Q=diag([10, 1]), R=1 and simulate the step response.\"\n",
-    "\n",
-    "print(f\"User Query: {prompt}\\n\")\n",
-    "result = agent.run(prompt)\n",
-    "\n",
-    "# Display formatted response\n",
-    "display(Markdown(result.final_response))\n",
-    "\n",
-    "# Display any generated simulation plots\n",
-    "for plot_path_rel in result.plots:\n",
-    "    clean_name = Path(plot_path_rel).name\n",
-    "    local_path = Path(\"outputs/plots\") / clean_name\n",
-    "    if local_path.exists():\n",
-    "        print(f\"\\nGenerated Simulation Plot: {clean_name}\")\n",
-    "        display(Image(filename=str(local_path)))"
-   ]
-  }
- ],
- "metadata": {
-  "language_info": {
-   "name": "python"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}

Modelfile

@@ -1,19 +0,0 @@
-# ControlAI Ollama Model Template
-FROM qwen2.5:3b
-
-# Set deterministic temperature for control engineering & exact math
-PARAMETER temperature 0.2
-PARAMETER top_p 0.95
-PARAMETER stop "<|im_end|>"
-PARAMETER stop "<|endoftext|>"
-
-# Set the safety-critical system prompt
-SYSTEM """You are ControlAI, a premier AI research scientist and expert engineering agent specialized in control systems engineering, applied mathematics, robotics, and dynamical systems.
-
-### 4-Stage Mathematical Reasoning & Proof Standard:
-When answering theoretical principles, derivations, proofs, comparisons, or limitation questions:
-1. State the state-space equations, signal spaces, and underlying assumptions.
-2. State the governing theorem with exact mathematical rigor (PBH rank test, Doyle 1978 LQG robustness counterexample, Lyapunov Invariance).
-3. Provide the full closed-form relationships in pure LaTeX.
-4. Explicitly state engineering breakdown conditions and conservation trade-offs (Bode sensitivity integral).
-"""

README.md

@@ -1,18 +1,37 @@
 # ControlAI: Open-Source Safety-Critical AI Agent for Control Systems Engineering
 
-[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/atakankahya/controlai-agent/blob/main/ControlAI_Colab_Demo.ipynb)
 [![GitHub stars](https://img.shields.io/github/stars/atakankahya/controlai-agent?style=social)](https://github.com/atakankahya/controlai-agent)
 [![Hugging Face Model](https://img.shields.io/badge/Hugging%20Face-ControlAI--Agent-blue)](https://huggingface.co/atakankahya/ControlAI-Agent)
 [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 [![Python 3.10+](https://img.shields.io/badge/Python-3.10%2B-blue.svg)](https://www.python.org/downloads/)
 
-**ControlAI** is an open-source, domain-specific AI agent engineered specifically for **Control Systems Engineering, Dynamical Systems, Robotics, and Applied Mathematics**.
+**ControlAI** is an open-source, domain-specific AI assistant engineered specifically for **Control Systems Engineering, Dynamical Systems, Robotics, and Applied Mathematics**.
+
+---
+
+## Quickstart: Run Locally in 2 Steps
+
+ControlAI runs locally on your workstation without cloud dependencies:
+
+```bash
+# 1. Clone the repository
+git clone https://github.com/atakankahya/controlai-agent.git
+cd controlai-agent
+
+# 2. Launch the application (Automatically opens browser at http://127.0.0.1:8000)
+./run.sh
+```
+
+To run in interactive terminal CLI mode instead:
+```bash
+./run.sh --cli
+```
 
 ---
 
 ## Motivation: Safety-Critical Verification in Control Engineering
 
-Standard large language models (LLMs) operate probabilistically without deterministic verification. When applied to physical systems—such as autonomous aerial vehicles, industrial manipulators, or power systems—general-purpose models present severe reliability challenges:
+Standard large language models (LLMs) operate probabilistically without deterministic verification. When applied to physical systems—such as autonomous aerial vehicles, industrial manipulators, or power grids—general-purpose models present severe reliability challenges:
 * **Numerical Hallucinations:** Estimating eigenvalues without characteristic polynomial evaluation, inverting singular matrices, or generating unstable feedback gains.
 * **Lack of Formal Proof Structure:** Omitting boundary conditions, PBH rank tests, or domain-specific stability limits.
 * **Physical Safety Violations:** A sign error in a state feedback gain leads directly to closed-loop instability in hardware.
@@ -20,22 +39,11 @@ Standard large language models (LLMs) operate probabilistically without determin
 **ControlAI addresses these limitations through a hybrid architecture:**
 1. **Deterministic Scientific Sandbox:** Computes continuous/discrete algebraic Riccati equations (CARE/DARE), matrix exponentials, and Bode diagrams using LAPACK, SciPy, and CVXPY.
 2. **4-Stage Mathematical Proof Standard:** Formulates system class, analytical theorems, closed-form derivations, and engineering breakdown limits.
-3. **Dynamic Simulation & Plotting:** Solves nonlinear differential equations and renders verified trajectories.
+3. **Dynamic Simulation & Plotting:** Solves nonlinear differential equations and renders verified trajectories directly in the interface.
 4. **Offline RAG Knowledge Engine:** Grounded with 68,000+ chunks indexed across classical and modern control engineering literature.
 
 ---
 
-## Live Demo and Access
-
-* **Interactive Google Colab Demo:**  
-  Run ControlAI in a cloud environment with zero local setup:  
-  [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/atakankahya/controlai-agent/blob/main/ControlAI_Colab_Demo.ipynb)
-
-* **Hugging Face Model Repository:**  
-  Access model weights and fine-tuning configurations: [atakankahya/ControlAI-Agent](https://huggingface.co/atakankahya/ControlAI-Agent)
-
----
-
 ## Architecture Overview
 
 ```mermaid
@@ -68,9 +76,9 @@ graph TD
 
 ## Key Capabilities
 
-### 1. Web Console (`web/`)
-* **Real-Time Token Streaming (SSE):** Word-level fluid rendering with smart scroll retention.
-* **LaTeX Formula Rendering:** KaTeX integration with math delimiter protection and code block syntax highlighting.
+### 1. Modern Interactive Web Console (`web/`)
+* **Real-Time Token Streaming (SSE):** Token-by-token fluid rendering with smart scroll retention.
+* **LaTeX Formula Rendering:** KaTeX integration with math delimiter protection and syntax-highlighted code blocks.
 * **Engineering Toolbar:** One-click insertion of state matrices, transfer functions, and control parameters ($\zeta, \omega_n$).
 
 ### 2. Deterministic Mathematical Tool Suite (`controlai_agent/tools/`)
@@ -83,7 +91,7 @@ graph TD
 
 ### 3. Live Python Execution Sandbox (`python_executor.py`)
 * Executes scientific Python routines using `numpy`, `scipy.signal`, `scipy.linalg`, `control`, and `matplotlib`.
-* Automatically isolates and captures generated simulation plots to `outputs/plots/`.
+* Automatically isolates and captures generated simulation plots to `outputs/plots/` and renders them in chat.
 
 ---
 
@@ -102,44 +110,6 @@ Evaluated across **50 multi-pillar benchmark problems**:
 
 ---
 
-## Local Installation and Usage
-
-### 1. Clone and Set Up Environment
-
-```bash
-# Clone the repository
-git clone https://github.com/atakankahya/controlai-agent.git
-cd controlai-agent
-
-# Create and activate virtual environment
-python3 -m venv .venv
-source .venv/bin/activate
-
-# Install dependencies
-pip install -r requirements-corpus.txt
-pip install -r requirements-training.txt
-pip install rich
-```
-
-### 2. Launch Universal Assistant (Web & Browser Auto-Open)
-
-```bash
-./run.sh
-```
-* Starts the backend server and opens the browser interface at `http://127.0.0.1:8000`.
-
-### 3. Interactive Terminal CLI
-
-```bash
-# Launch interactive terminal session
-./run.sh --cli
-
-# Or execute a single query directly
-python cli.py "Design an LQR controller for A=[[0, 1], [-2, -3]], B=[[0], [1]], Q=diag([10, 1]), R=1"
-```
-
----
-
 ## Open-Source and Community Support
 
 ControlAI is an **open-source project** developed for control engineering researchers, robotics practitioners, and applied mathematicians.

scripts/fuse_model.py

@@ -1,60 +0,0 @@
-#!/usr/bin/env python3
-"""Merge/Fuse ControlAI LoRA fine-tuned adapter into base model to produce a standalone model."""
-
-from __future__ import annotations
-
-import argparse
-import subprocess
-import sys
-from pathlib import Path
-
-PROJECT_ROOT = Path(__file__).resolve().parent.parent
-
-
-def main() -> None:
-    parser = argparse.ArgumentParser(description="Fuse ControlAI LoRA adapter into base model")
-    parser.add_argument(
-        "--model",
-        type=str,
-        default="mlx-community/Qwen3-4B-Instruct-2507-4bit",
-        help="Base model path or Hugging Face repo ID",
-    )
-    parser.add_argument(
-        "--adapter-path",
-        type=str,
-        default="adapters/controlai_qwen3_4b_sft_v2",
-        help="Path to trained LoRA adapter",
-    )
-    parser.add_argument(
-        "--save-path",
-        type=str,
-        default="models/controlai_fused",
-        help="Target folder for standalone merged model",
-    )
-
-    args = parser.parse_args()
-
-    print(f"Fusing LoRA adapter from {args.adapter_path} into {args.model}...")
-    cmd = [
-        sys.executable,
-        "-m",
-        "mlx_lm",
-        "fuse",
-        "--model",
-        args.model,
-        "--adapter-path",
-        args.adapter_path,
-        "--save-path",
-        args.save_path,
-    ]
-
-    res = subprocess.run(cmd, cwd=PROJECT_ROOT)
-    if res.returncode == 0:
-        print(f"\nModel successfully fused into standalone checkpoint at: {args.save_path}")
-    else:
-        print(f"\nError during fusion (exit code {res.returncode})")
-        sys.exit(res.returncode)
-
-
-if __name__ == "__main__":
-    main()