Files
tidewall/tools/image_stats.gd
T
claudeandClaude Opus 5.5 68635d5fba F7: Offscreen render capture harness
tools/capture.sh <scene-or-scenario> <out.png> renders the real client
in a throwaway tidewall-capture-* Docker container (debian:13, Xvfb,
Mesa software rendering), with no network, no host display, GPU or
audio, and removes the container afterwards. The tidewall-capture image
rebuilds only when its recipe changes.

Forward+ through Lavapipe is the default, matching the game;
CAPTURE_RENDERER=compatibility uses OpenGL through llvmpipe. The script
fails if Godot silently falls back to another renderer, and fails on a
blank frame, judged by ImageStats (luminance variance and colour count).

make capture-test renders the shore test scene with both renderers and
proves an empty scene is rejected. ImageStats has unit tests in
make check. CLAUDE.md now requires a capture for visual issues, and
requires that it be opened and looked at.

Closes #7

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-09 20:13:22 -05:00

46 lines
1.6 KiB
GDScript

## Measures a captured image, so a blank or broken render fails automatically.
## A capture still has to be looked at by a person or agent; this only catches the
## obvious failures.
class_name ImageStats
extends RefCounted
## Below this luminance variance (on a 0..1 scale) an image counts as blank.
const BLANK_VARIANCE: float = 0.0005
## A real scene has far more distinct colours than a cleared or broken frame.
const BLANK_MAX_COLOURS: int = 16
## Sampling stride, in pixels. Enough for a 1280x720 frame in milliseconds.
const STRIDE: int = 4
## Returns mean and variance of luminance, and the number of distinct sampled colours.
static func measure(image: Image) -> Dictionary:
var source: Image = image
if source.is_compressed():
source = image.duplicate() as Image
source.decompress()
var sum: float = 0.0
var sum_squares: float = 0.0
var samples: int = 0
var colours: Dictionary = {}
for y: int in range(0, source.get_height(), STRIDE):
for x: int in range(0, source.get_width(), STRIDE):
var colour: Color = source.get_pixel(x, y)
var luminance: float = colour.get_luminance()
sum += luminance
sum_squares += luminance * luminance
samples += 1
colours[colour.to_rgba32()] = true
if samples == 0:
return {"samples": 0, "mean": 0.0, "variance": 0.0, "colours": 0}
var mean: float = sum / samples
return {
"samples": samples,
"mean": mean,
"variance": maxf(sum_squares / samples - mean * mean, 0.0),
"colours": colours.size(),
}
static func is_blank(stats: Dictionary) -> bool:
return stats["variance"] < BLANK_VARIANCE or stats["colours"] <= BLANK_MAX_COLOURS