Installation

Installation

Automatic installation (recommended on a fresh machine)

On a machine that has nothing installed (neither Python nor Ollama), the installers automate everything, checking at each step what is already present:

1. Ollama (local AI engine) — Homebrew on macOS, official script on Linux, winget/official installer on Windows, pkg install ollama on Termux when available.
2. An Ollama model suited to the machine:
   • Android/Termux: mini model llama3.2:1b (~1.3 GB) — limited RAM/storage on mobile;
   • macOS / Windows / Linux: medium model llama3:latest (~4.7 GB).
3. Python 3 — Homebrew/apt/dnf/pacman/zypper/pkg or winget.
4. Virtual environment (.venv) — created and activated automatically.
5. Python dependencies (networkx, matplotlib, requests, and psutil except on Android).
6. Launch of process_analyzer_allinone.py (installer arguments are forwarded as-is, e.g. ./install.sh --no-enrich).

A failure on Ollama or the model never interrupts the installation (the analysis works without AI); a failure on Python is fatal.

macOS / Linux / Termux:

chmod +x install.sh
./install.sh

Windows (PowerShell):

powershell -ExecutionPolicy Bypass -File install.ps1

These installers touch the system (software installation, possibly sudo/administrator rights) — read them before running them on a sensitive machine, as with any installation script fetched online.

Prerequisites (manual setup)

• Python 3.9+
• Dependencies: networkx, matplotlib, requests, and psutil (except Android/Termux). Installed automatically on first launch if missing, or:

pip install -r requirements.txt

• Ollama is optional — see Ollama Integration. Without it, the tool still works, just without AI-generated descriptions.
• An internet connection is required only for installing dependencies/Ollama, downloading a model, and displaying the 3D graph (CDN library). The analysis itself needs no external network access.

Next: Getting Started