Sim_Proc

A Windows-friendly pipeline to build SimulationCraft profiles, run chunked sims, extract metrics, and visualize results with Dash.

Features

• Chunked SimulationCraft runs with progress bar and parallel workers
• HTML/JSON parsing with rich metrics and automatic cleanup
• Talent picker web UI and 3D results dashboard
• One-command dependency setup, plus a no-Python bootstrap path
• Optional SimulationCraft build helper (installs into the parent folder)

Requirements

• Windows 10/11
• Python 3.10+ (installer provided)
• SimulationCraft simc.exe (can be built via simc_install.py)

Quick Start

1. Open PowerShell in this folder.
2. Install Python deps (creates .venv and installs requirements):

python .\setup_installer.py

• No Python? Use the bootstrap:

powershell -ExecutionPolicy Bypass -File .\no_python_installer.ps1

3. (Optional) Build SimulationCraft into the parent folder:

python .\simc_install.py --branch midnight

This places simc.exe beside this Sim_Proc folder (not inside it).

4. Run the pipeline:

python .\sim_proc.py

Follow prompts to pick talents, build reports, collect metrics, and launch the dashboard.

Key Scripts

• sim_proc.py: Orchestrates the full flow end-to-end.
• talent_dashboard.py: Web UI to pick talents and generate profile input.
• build_simc_reports.py: Splits work into chunks and runs simc.exe in parallel.
• collect_data_metrics.py: Parses HTML/JSON outputs into a combined CSV and cleans up.
• 3d_dashboard.py: Interactive 3D visualization of results.
• setup_installer.py: Creates venv and installs Python dependencies.
• no_python_installer.ps1: Downloads Python and then runs the installer.
• simc_install.py: Downloads sources and builds simc.exe into the parent directory.

Configuration (optional)

Create config.ini in this folder to override defaults.

[sim]

• iterations: e.g. 10000
• output_format: html or json
• profiles_per_sim: chunk size, e.g. 100
• workers_core_fraction: fraction of cores to use, e.g. 0.3
• simc_branch: branch to fetch when building SimulationCraft, e.g. midnight

[cleanup]

• delete_html: true|false
• delete_json: true|false

How simc.exe is found

build_simc_reports.py resolves simc.exe by checking:

1. SIMC_PATH env var (if set)
2. PATH (e.g. Program Files install)
3. Local paths: this folder, the current working directory, then the parent folder (new default install target)
4. Prompts for a path if not found

Typical Workflow

1. python sim_proc.py → follow prompts to pick talents and generate input
2. build_simc_reports.py runs sims in parallel and saves outputs under simc_data/
3. collect_data_metrics.py aggregates HTML/JSON into sim_proc_combined_results.csv and cleans up
4. 3d_dashboard.py launches the visualization

Notes

• This repo intentionally excludes SimulationCraft binaries and heavy output data. See .gitignore.
• If ports are busy (8050), close previous dashboards or change the port.
• You can re-run individual steps by running the scripts directly.

Credits

• Talent logic adapted from simc-talent-generator by Vít Šefl (MIT): https://github.com/vituscze/simc-talent-generator See THIRD_PARTY_NOTICES.md for license details.

License

MIT. See LICENSE for details.