Under the hood

This path is for readers who want to see the calls. The ten notebooks build the same Set as Path 1, one small step per cell, through the sideman Python client:

from sideman import Live
live = Live()
live.get("live_set", "tempo")

Every code cell was executed, in order, against a fresh Set, and its output is what you see — nothing is illustrative. The last cell of each notebook asserts the checkpoint, so re-running a notebook on a Set that has drifted fails loudly instead of quietly producing a different track.

To run them yourself: install Sideman from the public repository, open Live on a fresh Set, and execute the notebooks in order with the sideman Jupyter kernel (python -m ipykernel install --user --name sideman from the project’s environment). Don’t re-run an earlier lesson on a later state; each one expects the Set exactly as the previous one left it.

Where the Live Object Model differs from what you would guess — a search that matches fourteen paths for four tracks, a parameter that reads back as a 32-bit float, a macro no script can map — the notebook shows the real result and says why. The syllabus lists the tools each lesson introduces; the tool reference and the coverage census cover the rest.

Begin with 1. Orientation.