DIMS v1.0.1

Documentation and wording only. No analysis or payload change: a study takes
this by re-vendoring, with no rebuild.

• The README's install and dims-case new commands work as written — the
  checkout install, --visibility, and case-<name> as the directory created.
• Python 3.10 is the floor everywhere. Three pages said 3.9, which pip refuses.
• The cross-effector network is documented where it was missing: the built-in
  tab list, the scaffold's tab table and config example, and what
  include_network needs to draw anything useful.
• A dashboard reading an old asset no longer names a release the renumbering
  removed; it says the payload is older and how to rebuild.
• The agent-ready issue label is now ready.

Not on PyPI yet — install from a checkout:
git clone https://github.com/dims-network/dims && pip install -e './dims[builder]'

DIMS v1.0.0

The first public release.

See a live dashboard — a
recording, its signals, and the analyses running on them. Click the timeline and
the video and every chart follow.

What it is

DIMS turns recordings and the time series taken from them — motion tracking,
physiology, gaze codes, anything sampled over time — into a dashboard you open
in a browser. Every signal on one timeline, beside the video it came from, with
the analyses that say how two signals relate: recurrence, cross-recurrence,
cross-wavelet coherence, and a cross-effector network.

Coherence is measured against a chance level estimated by simulation. Two
unrelated signals score about 0.25, not 0, so a number without that comparison
cannot be read.

Try it

pip install "dims-network[builder]"
dims-builder

A wizard opens in your browser: point it at a folder of recordings — or load the
example study — choose the analyses, and it builds the dashboard and runs them.

Notes

• Python 3.9+. No build step, no bundler: a dashboard opens from a plain file
  server.
• A study lives in its own small repository holding its config, its data and a
  pinned copy of this core. Studies of identifiable people declare themselves
  private and get commit, push and CI guards that keep that data out of git.
• The wavelet analyses follow Torrence & Compo (1998); the constants are
  transcribed in examples/reference/ and checked against synthetic signals
  whose answers are known in advance.

Documentation: https://dims-network.github.io/