Install

pip install sillon

That is the whole toolchain: the logging client, the CLI, and the analysis library. There is no server to configure and no account to create.

Optional extras

pip install "sillon[analysis]"   # + pandas, for to_dataframe()

Figures need matplotlib, which sillon does not install for you — if you are plotting, you already have it.

Requirements

• Python 3.11 or newer
• Linux, macOS or Windows

The client and the daemon talk over a Unix domain socket where one is available, and over a loopback TCP port on Windows, where CPython does not expose AF_UNIX. You do not configure this — sillon picks the right one. Set SILLON_TRANSPORT=unix|tcp only if you want to force it (the test suite does, to exercise the Windows path on Linux).

Check it worked

sillon --version

You get two commands on your PATH:

Command	What it is
sillon	the CLI you use to explore runs
sillon-server-daemon	the background logging daemon — you never start this yourself

What gets created, and where

Nothing until you log something. The first time a script calls sillonpy, sillon creates a .sillon/ folder next to it:

.sillon/
  database.sql            SQLite: one row per run
  glob/<uuid>/glob.hdf5   heavy arrays
  artifact/<uuid>/        files you logged
  figure/<uuid>/          figures you logged
  daemon.log              the background daemon's log

That folder is your project. Copy it, move it, back it up like any other data.

Add it to your .gitignore — it holds results, not source:

.sillon/

One file lives outside the project: a registry of every project on this machine, so sillon projects can tell you where they all are. It goes in ~/.config/sillon/ (or %APPDATA%\sillon on Windows).

Install from source

git clone https://github.com/balerat/sillon
cd sillon
pip install -e ".[dev]"     # or: make install
make test

See Development setup for the layout of the repository.