AdaptiveSolvePlanner

AdaptiveSolvePlanner is a production-oriented adaptive tournament planner and solver. It supports sequential and process-parallel search over retention/decay parameters to produce "nice" tournament round plans (groups, qualifiers, wildcards) while honoring configurable limits.

Quickstart

To install from PyPI:

python -m pip install adaptivesolveplanner

To install latest from GitHub:

pip install git+https://github.com/Smoki-Cool/adaptivesolveplanner.git

Example:

from adaptive_solve_planner import plan

p = plan(start_guess=100, finals_target=12, rounds_desired=5, strictness="generous",
         prefer_wc=True, allow_wc_down=True, top_k_per_round=12, max_nodes=300000, time_limit=12.0)
print(p["_meta"])
for r in p["rows"]:
    print(r)

API

Two primary entry points:

• plan(...) — sequential single-process planner
• parallel_plan(...) — process-parallel planner with auto-tuning

Both return a dict with:

• rows — per-round dicts
• final_total — final number of players
• _meta — diagnostic metadata
• _score — score used for selecting best plan

See examples/ for runnable samples.

Configuration

PlannerConfig centralizes defaults and can be overridden from environment variables using the ASP_ prefix (e.g. ASP_WC_PCT=0.12). See adaptive_solve_planner/config.py.

CLI

If you installed the package with the console script, run:

adaptivesolveplanner --help

Development

Create a virtualenv, install dev requirements, run tests:

python -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install build twine pytest
python -m pytest

Contributing

See CONTRIBUTING.md.

License

This project is licensed under the GNU General Public License v3.0 — see the LICENSE file.