[claude]

Tests-only: strengthens tests/test_view.py for rupture_generator/view.py.

What the new tests pin down

• _packed puts channels into packed 0xRRGGBBAA uint32 values, rounds each channel and sets alpha to 0xFF. The test checks exact values, including fractional inputs that have to round.
• VIRIDIS and HOT each have 256 shades, with exact endpoints (0x440154FF/0xFDE725FF and black/white). _colours clips values outside [low, high] to the ends of the map and puts a midpoint on shade 127.
• _isochrones returns an empty (0, 2, 3) array when no edge crosses the level. _onset_contours returns empty lines and anchors and no labels for a rupture whose onset is the same everywhere.
• _positive_seconds accepts a valid --step, and main records with a non-default step.
• main exits with status 2 and prints a rupture-view: message for a missing file and for a file that isn't an SRF, and writes no recording in either case. When rerun-sdk can't be imported, it exits with status 2 and says to install rupture-generator[vis].

Coverage of view.py

• Before: 96% (uncovered lines 281, 501, 535–545, 556)
• After: 99% (only line 556 is left: the if __name__ == "__main__" guard)

Mutants

The new _packed exact-value test and the colour-map endpoint test kill all 13 surviving _packed mutants:

• mutmut_1, mutmut_2 and mutmut_3 replace the conversion with None. Each one either raises or skips the uint32 cast.
• mutmut_4, mutmut_5 and mutmut_6 turn a | into &, which drops channels or alpha.
• mutmut_7, mutmut_9 and mutmut_11 flip << to >>, which drops a channel.
• mutmut_8, mutmut_10 and mutmut_12 shift a channel by one bit too many.
• mutmut_13 sets alpha to 256 instead of 0xFF.

I judged none of the survivors equivalent.

No issues filed: none of the tests showed the code doing something wrong.

🤖 Generated with Claude Code

[github-actions]

Benchmarks

faa8d5e against base 26b166a, on AMD EPYC 9V45 96-Core Processor. Times are medians over the rounds. The median change is 1.4%; changes past 5% are in bold.

Kernels

kernel	base	head	change	head IQR	rounds	padded cells
embed[Mw6.5]	27.8 ms	27.4 ms	-1.3%	324 µs	35	448,000 (+0.0%)
circulant_draw[Mw6.5]	2.58 ms	2.61 ms	+1.4%	312 µs	258
embed[Mw7.5]	86.6 ms	87.6 ms	+1.2%	10 ms	12	1,204,224 (+0.0%)
circulant_draw[Mw7.5]	6.02 ms	5.99 ms	-0.5%	370 µs	124
eikonal_solve	945 ms	930 ms	-1.6%	12.2 ms	5
synthesise_pulses[impulse]	381 µs	365 µs	-4.3%	17.8 µs	1253
synthesise_pulses[liu]	331 ms	337 ms	+1.9%	8.64 ms	5

End to end

Each round is the command line in a fresh interpreter, start-up included. Peak RSS is the largest the child process reached over the rounds.

example	base time	head time	change	base peak RSS	head peak RSS	change
alpine_hope	6.67 s	6.6 s	-1.0%	294 MiB	295 MiB	+0.3%
beavan	2.45 s	2.45 s	-0.2%	232 MiB	232 MiB	+0.0%
colombia	6.37 s	6.15 s	-3.5%	3,171 MiB	3,172 MiB	+0.0%
crustal	1.04 s	1.04 s	-0.4%	246 MiB	246 MiB	-0.1%
landers	2.03 s	1.99 s	-1.9%	385 MiB	385 MiB	-0.1%
northridge	1.03 s	1.02 s	-0.3%	290 MiB	289 MiB	-0.5%
two_faults	1.74 s	1.76 s	+1.0%	354 MiB	354 MiB	+0.2%
two_faults_segments	1.03 s	1.01 s	-1.9%	192 MiB	191 MiB	-0.2%
venezuela	12.5 s	12.2 s	-2.1%	5,769 MiB	5,766 MiB	-0.1%