3 Commits
Author SHA1 Message Date
Erik Fröjdh 34f0d9f679 Drop detector type and quad flag from the gap pixel API
Build on RHEL9 / build (push) Successful in 2m55s
Build on RHEL8 / build (push) Successful in 3m39s
Run tests using data on local RHEL8 / build (push) Failing after 4m43s
insert_gap_pixels and gapped_shape only used the detector type to reject
anything but Jungfrau and Eiger, and the quad flag only to shrink the
module width to 512. Both Jungfrau and Eiger use 256x256 pixel chips in
1024x512 modules, and an Eiger quad is 512x512 pixels, so its only chip
boundaries are at pixel 256 and the module width never matters. The
layout is now fixed and both functions take just the image, its ROI and
the GapPixels configuration, in C++ and in Python.

RawFile keeps rejecting gap pixels for other detector types with
std::invalid_argument, now checked in the reader instead of the layout.
2026-10-13 12:35:36 +02:00
Erik Fröjdh 58da96d6f0 Insert gap pixels only between chips unless module gaps are set
Build on RHEL9 / build (push) Successful in 3m1s
Build on RHEL8 / build (push) Successful in 3m43s
Run tests using data on local RHEL8 / build (push) Failing after 4m48s
An unset GapPixels::module_gaps defaulted to the chip gap, so every
module boundary received two extra rows or columns. Two stacked
Jungfrau modules read as 1030x1030 instead of 1030x1028. Unset now
means no pixels between modules; only the chip boundaries inside a
module receive a gap.

Also drop the explicit lambda capture in split_edge_counts that clang
reported as unused for floating point pixels.
2026-10-09 21:59:07 +02:00
Erik Frojdh d1d0e5f312 missing files
Build on RHEL9 / build (push) Successful in 3m0s
Build on RHEL8 / build (push) Successful in 3m43s
Run tests using data on local RHEL8 / build (push) Failing after 4m46s
2026-10-09 08:16:59 +02:00