[geonnave]

Stacked on #343 (spin-append-placed-hint at 9ddf837, itself based on main), for the placed_from field. Review the top two commits.

Problem

Strengthening a site's spin calibration means collecting more rounds in a free frame (in the site the spins were first collected in), then moving the result onto the site already in use, through a station both calibrations solved. That move was done with a scratch script: fit the turn and shift that map the new calibration's station-0 circle centres onto the old calibration's, apply it to every station, write the file.

Change

dotbot swarm calibrate-lh2 place SOURCE --to-calibration TARGET [--station N ...] [--tag T]

• Every source circle a chosen shared station saw is mapped through that station's homography in each calibration (same raw reads, two frames); the mean of each circle's points gives one correspondence, and a least-squares rigid fit (Kabsch, no scale: both frames are metric through the spin radius) gives the move. It is applied to every source station.
• The new file lands in TARGET's site with TARGET's anchor and fence as every station's rectangle, placed_from = SOURCE, tag <source tag>-<site> unless --tag.
• It prints the turn, the shift and the residual over the circles, rms and max.
• Refused: a SOURCE not in its own free frame, a free-frame TARGET, no shared station, a --station not in both, fewer than 3 circles. With several shared stations --station is required rather than picking all of them: a weakly tied station in TARGET would pull the fit, which is usually the reason for re-collecting in the first place.

Naming: place is the word the site editor uses for moving a spin calibration onto a site (#326), --to-calibration mirrors site init --from-calibration, and --station takes an index like --drop-station. It lives under swarm calibrate-lh2 beside show and reframe, which also only read and write files, and like them refuses dotbot swarm -d. #326's place_calibration does the same move from the console with a hand-set transform; once both land, the two can share one mover.

Validation

• Unit: two synthetic stations, a spin calibration placed with site init, a second free-frame one from other robots, place through station 0: residual under 1 mm rms, a non-trivial move, and both stations then put every floor point in the overlap within 2 mm of the target's. Refusals covered too. Full suite: 1791 passed, the same 2 environment failures as on the base.
• Bench data, read-only (copies of the files): place 05aea267 --to-calibration 04707621 --station 0 gives turn 89.216 deg, shift (3105.0, 2029.6) mm, 9 circles, rms 2.99 mm, max 6.07 mm, the same move as the script that produced e0897cfb. The homographies agree with e0897cfb to 1e-12; the id differs only through those last float bits. Through both stations the rms is 9.35 mm, which is why --station is required.
• Not run on hardware. Formatting is deferred to the pass after review.

File	+	-
dotbot/calibration/conics.py	+165	-8
dotbot/tests/test_calibration_spin.py	+82	-0
dotbot/cli/swarm_lh2.py	+65	-0
Total, 3 files	+312	-8

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.56%. Comparing base (7a76840) to head (8aeadc9).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #344   +/-   ##
=======================================
  Coverage   88.56%   88.56%           
=======================================
  Files         230      230           
  Lines       37623    37623           
  Branches     2367     2368    +1     
=======================================
  Hits        33319    33319           
  Misses       4300     4300           
  Partials        4        4           

Flag	Coverage Δ
console	82.18% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
dotbot/cli/swarm_lh2.py	83.25% <ø> (ø)
dotbot/tests/test_calibration_spin.py	100.00% <ø> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.