[hanmarsan]

Summary

This PR sets the imaging boundary: everything specific to the 2P imaging experiments moves to the private rig repository, reiserlab/scanimage-2026-bergamo, under 2p-rig-nodes/. That directory is a pixi workspace in the rig-nodes layout (package twop_rig_nodes), built on the public rig_nodes package. Michael decided this on 2026-10-28 (LAB-303, LAB-319).

The PR has two commits: the removal, and the gateway half of #261 / #266, which stays public.

1. Removal (32264be)

• Deleted. Each file now lives in the private repo; 2p-rig-nodes/PROVENANCE.md there gives each file's source commit:
  • tiff2mcap (rig: tiff2mcap --runlog drives the timeline and the clock (LAB-303) #273): rig/tools/tiff_run_to_mcap.py, check_tiff2mcap.py, rig/tests/test_tiff2mcap_runlog.py, and the three tasks
  • tools/led_sweep_si.py, tools/led_sweep_quicklook.py
  • the records: rig/SCANIMAGE-DAY.md, rig/scanimage-day-2026-10-16.md, docs/development/2p-scanimage-signal-logging.md, the superseded 2p-line-sync-rig-checklist.md
• The ScanImage node leaves the arena composition: rig.py loses --scanimage, --si-mock and --scanimage-args, and the scanimage-node, scanimage-cmd, si-mock tasks go. The harness no longer starts the ScanImage stand-in and node, and drops its 5 ScanImage checks; rig/WINDOWS-TEST.md has the new totals.
  • rig-nodes drops the node for 0.4.0 (0.4.0: the ScanImage JSON connector moves to the private 2P rig repo (draft, after #3) rig-nodes#4). Re-pinning rig/ to 0.4.0 must not come before this PR.
• Kept: rig/tools/patfile.py. It is arena code, bound for pyDisplayTools per §3.3; the private repo has a copy.
• Docs:
  • rig-software-repos-and-runners.md gets a dated §3.3 amendment recording the decision and the boundary.
  • The orientation skill's rig-nodes row and CLAUDE.md's rig line are updated.
  • Comments in the Bergamo rig profile and three 2P protocols point to the moved notes.
  • tests/test-flasher-catalog.js loses its exception for the moved checklist.

2. Gateway half of #261 / #266 (e25914e), ported from #266's head a4442ad (which includes #261):

• bridge_node:
  • the sipm connector: /sipm/cmd|reply|status in CONNECTOR_TOPICS, every bag's topic map, and hello_ack;
  • STREAM_TOPICS for the ScanImage frame stream: /scanimage/frame_meta, frame_stats, preview/meta, and preview/ch1-4 as sensor_msgs/CompressedImage. They are recorded on arrival and never derived into the run log.
• mcap_convert: /sipm/reply|status → {"type":"connector","src":"sipm",…}.
• rig/tests/test_gateway_registration.py (pixi run test-gateway, 5 checks): the gateway checks from LAB-319 / Block S: read-only SiPM controller node (bias, enable, offsets, temperature into the run log) #266's test_sipm (D1–D3) and LAB-303 / Block R: ScanImage ZMQ frame-stream wrapper node, latency report, off-rig harness #261's test_scanimage_frames (§G).

What stays public (the boundary):

• the connector mechanism: dispatch, CONNECTOR_TOPICS (including scanimage), mcap_convert's connector derivation, the /<src>/cmd|reply|status contract, ConnectorReply
• the Studio's ScanImagePlugin, and the emulator's --scanimage-cmds option
• for now: the 2P protocols (g6_2p_*, g6_2x10_2p_*) and configs/rigs/bergamo_g6_2x10_2p.yaml

History: the moved files were copied, not rewritten in history, so earlier versions stay readable here. #261 and #266 get closed with a pointer after the release.

Provenance: commits in the private repo, main:

• e3e9ba8: 2p-rig-nodes/ with the ScanImage, frames and SiPM nodes, the records and PROVENANCE
• d235f9e: tiff2mcap and the LED-sweep tools
• 8ceddf6: SETUP.md, the Bergamo rows from rig-nodes
• dea41ed: re-pinned to rig_nodes 0.3.0
• 694348b: the patfile.py note

Off-rig on the Bergamo PC, the private tests pass: test-sipm 25/25, test-frames 39/39, test-scanimage 11/11, test-tiff2mcap.

Tier

next candidate (2026-10-28 rc2). It merges after #281, #269 and #282. It is based on main 8d5b952, not rebased onto the candidate. Expect conflicts in rig/nodes/rig.py and rig/pixi.toml, resolved in the candidate build. rig/pixi.lock is byte-identical to main; #281 carries the re-lock.

Release notes

none (rig-side only)

Test plan

Run on the Bergamo PC (Windows 11, pixi 0.82), on rig_nodes 0.2.1 as main pins it.

• pixi run test (root): 72/72 files
• cd rig && pixi run test: 54/54 (main: 59, minus the 5 ScanImage checks)
• cd rig && pixi run -e typed test: 57/57 (main: 62)
• cd rig && pixi run test-gateway: 5/5; on the gateway commit alone, the harness is 59/59, the same as main
• the two commits merge cleanly
• Note: pixi 0.82 rejects main's rig/pixi.lock with --locked (rig/pixi.lock: re-lock with pixi >= 0.80 and pin requires-pixi (plain install on a rig rewrites the lock) #280). For these runs the environments were installed unlocked, and the lock was then restored.

🤖 Generated with Claude Code