Replace Percy with nitpix for visual testing

[T4rk1n]

Visual snapshots are now captured locally by dash.testing (real browser screenshots into NITPIX_SNAPSHOT_DIR, one PNG per width) and diffed in CI by plotly/nitpix: baselines live on the nitpix-baselines orphan branch, PRs get a review comment with before/after/diff images and a nitpix/visual commit status, and changes are approved by commenting "/nitpix approve" — no test-matrix re-run needed.

• dash/testing/browser.py: percy_snapshot() keeps its name and signature (public API) but saves PNGs locally instead of calling the Percy SDK. convert_canvases is now a no-op: real screenshots capture canvases.
• testing.yml: drop PERCY_* env, percy-exec wrapper and percy-finalize; snapshot-producing jobs (test-main, html-test, dcc-test, 3.12 leg only) upload nitpix-snapshots-* artifacts instead.
• nitpix.yml (new): workflow_run diff/report job in base-repo context, so fork and Dependabot PRs are fully supported — the old percy/dash placeholder-status hack in post-test-status.yml is removed.
• nitpix-approve.yml (new): handles /nitpix approve comments.
• table-visual-test still uses percy-storybook (separate token/mechanism); converting it is a follow-up.

[sonarqubecloud]

Quality Gate passed

Issues
11 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

Dash performance benchmarks

✅ all within thresholds

	scenario	metric	p90 (ms)	median	growth	baseline p90	note
✅	callback_chain	chain_ms	281.5	278.5	0.95x	499.1
✅	callback_chain	graph_ms	1.2	1.2	1.0x	2.4
✅	callback_fanout	fanout_ms	62.9	56.1	0.84x	92.5
✅	deep_nesting	render_ms	42.6	40.1	1.35x	56.8
✅	full_children_replace	replace_ms	3075.9	1144.1	17.04x	4697.1
✅	initial_render_large	render_ms	448.6	381.6	1.23x	694.4
✅	initial_render_small	render_ms	84.2	65.7	1.39x	104.0
✅	patch_append_nested	append_ms	106.0	77.4	2.15x	192.3
✅	patch_append_toplevel	append_ms	99.2	60.2	2.76x	140.2
✅	patch_scalar_update_large	update_ms	144.7	124.5	0.94x	202.9
✅	wildcard_all_resolve	wildcard_ms	175.3	159.8	0.9x	313.6
✅	wildcard_all_resolve	graph_ms	0.7	0.7	1.0x	1.3

growth = late-third / early-third per-op time; ~1 is flat, a large value means the per-op cost scales with accumulated state.

machine scale vs baseline: 0.68x - divided out of the baseline ratios so they compare like for like (the absolute warn/fail ceilings are left un-scaled); calibrated on initial_render_small.

[sonarqubecloud]

Quality Gate passed

Issues
5 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[camdecoster]

Could you add an update to the changelog? Some downstream users might be relying on Dash installing percy.

Do you think a test should be added to verify the snapshot pipeline is working? Right now, if a call to nitpix fails, there's no feedback.