x01dcandClaude Sonnet 5
906e9f5318
CI for csaxs_bec / test (push) Failing after 5s
Mirko wants the tomo params GUI widget available for OMNY before running
real tomo scans. That widget (Phase 2, separate follow-up) needs a working
job queue and scan progress visible outside the CLI session that started
it -- neither existed for OMNY. OMNY now inherits TomoQueueMixin
(OMNY_shared/tomo_queue_mixin.py), the same shared backend flomni/lamni
already use, giving it tomo_queue_add/delete/move/clear/show/execute/
reacquire, at_each_angle_hook registration, and tomo_scan_resume() for the
first time.
self.progress is now a global-var-backed proxy (_ProgressProxy, byte-for-
byte port of flomni's/lamni's own), not a plain in-memory dict -- the
load-bearing change the queue mixin and the eventual widget's busy-banner
both depend on. tomo_queue_mixin.py's tomo_queue_reacquire() hardcoded
tomo_type == 1; generalized via a new _TOMO_TYPE_1_LIKE class attribute
(default frozenset({1}), flomni/lamni unaffected) since OMNY has three
"equally spaced sub-tomograms" flavors (types 1/4/5), not just one.
Along the way, fixed several real bugs in tomo_scan()/write_pdf_report()
that predate this session and would have crashed on a real account/
real energy read (bec.active_account.decode() -- active_account is str
not bytes; dev.mokev doesn't exist for OMNY, only dev.ccm_energy; an
unguarded logbook send; a stale "LamNI" logo/tag copy-paste leftover) --
all direct ports of fixes flomni/lamni already needed and got. Also added
the missing unconditional omnygui_show_progress() call and heartbeat-
clearing try/finally in tomo_scan(), matching both.
Verification: 23 new tests (test_omny_tomo_queue.py, test_omny_tomo_scan.py)
covering the mixin wiring, hook dispatch/resolution, tomo_scan_resume()/
_resolve_type1_projection()/reacquire() across all three sub-tomogram
flavors, account handling, and heartbeat clearing on both normal completion
and a mid-scan exception. Full suite: 702 passed. Not yet live-tested
against a running session (the local bec-server may be in active use by
Mirko's own parallel session) -- next step is his own pull-and-test.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QLrD7sVYGLAzsQjLVJpCgt