Skip to content

style(perf-test): filter-style setup row and one results card - #2004

Closed
mariana953 wants to merge 15 commits into
Comfy-Org:PabloWiedemann/sidebar-tabs-assetsfrom
mariana953:mar/perf-test-cards
Closed

mariana953 wants to merge 15 commits into
Comfy-Org:PabloWiedemann/sidebar-tabs-assetsfrom
mariana953:mar/perf-test-cards

Conversation

@mariana953

@mariana953 mariana953 commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

ELI5

The Performance Tests page now reads like the rest of the app. Everything you set up sits in one row of the same filter buttons used on Apps and Benchmarks (Build, Workflow, the two run counts, then Run), and the results live in one card: the median time big on the left, fastest, average and slowest on the right, and the computer details folded at the bottom.

Summary

  • Setup row. Build and Workflow use the shared filter trigger from Apps and Benchmarks: secondary 32px button, muted label, value, chevron, and the same Menu with 13px items. Warm-up runs and Measured runs are matching chips with their info tooltip and number field. Run (or Stop while a test runs) closes the row on the right. It wraps on narrow windows.
  • Workflow menu. Choose an example workflow, Import API workflow (with "or drop the .json"), and, once a file is chosen, Delete workflow in red under a divider. This replaces the loose X button. The file name and its Example or API tag show in the trigger; the full path is its tooltip. Dropping a .json on the trigger still imports it.
  • Status line. Preparing, the example download (message and progress bar) and import errors show in one line under the row, so the layout never jumps.
  • Results card. One soft card. A header names what ran ("sdxl-turbo_api.json on Stable 0.3.60") with Open folder and Export results on the right. While running: "Run 3 of 6", the phase (warm-up or measured run n of m) and a segmented bar, one segment per run with warm-ups lighter (a continuous bar above 30 runs). When done: the median in 44px with "Median run · 5 measured · 0 failed", then Fastest, Average and Slowest over a range rail with the median dot and average tick. With nothing run yet: "No results yet" and one line of guidance.
  • System information is the card's folded footer, with a one-line summary (device, VRAM, OS, PyTorch) and the full fact grid when opened.
  • Logs are folded by default with the last log line as the summary. They open on their own when a run starts and fold again when it succeeds.
  • locales/en.json and locales/zh.json: new strings for the placeholder, run label, run counter, run phase, empty state, median summary and chart legend; unused progress and sentence strings removed.

No functionality added or removed

Only layout, containers, styles and copy changed. Every handler, disabled rule, input id, error and progress state, telemetry event and IPC call is the same. Import by click and by drop, re-import over a chosen file, example download and cancel, delete, Run, Stop and export all behave as before. The workflow trigger is disabled exactly when the old import and delete buttons were (while a test launches or stops).

Before / after

Before is the base branch (569fb3a). Built app at 1320x900 (narrow at 1024x900). Screenshots use main's gray palette (#1996) applied locally: this branch is stacked on #1981, which predates it. The branch does not change the palette, so on main it renders gray. Before shots other than Idle and Build menu open are older captures, taken before the gray palette.

State Before After
No builds (empty)
Idle
Build menu open
Dragging a .json over Workflow Only the drop zone reacted
Workflow chosen
Workflow menu Separate buttons, X to remove
Example picker modal
Example downloading
Running
Stopping
Import error
Done with results
Done, scrolled
Narrow, done

Test plan

  • pnpm exec vitest run (332 files, 7669 tests) 3 runs in a row, all passed
  • pnpm run lint, pnpm run typecheck, prettier --check on changed files
  • Built app screenshots of every state above, before and after
  • Run a real test and check progress, results, System information and Export results

🤖 Generated with Claude Code

…logs

Build, workflow and measurement settings each sit in a card with an
uppercase label, and Results and Logs become cards too, as in the New
UI proposal's Performance Tests page. The run count fields are narrower
so they fit their card.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@coderabbitai

coderabbitai Bot commented Nov 10, 2026

Copy link
Copy Markdown

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

  • Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Replace the three step cards with a single setup panel split into columns,
move Run (and Stop, shown while a test is active) into the page header,
leave Results on the page with a proper empty state, and render Logs as a
console panel. BrandedPageHeader gains an optional actions slot.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@mariana953 mariana953 changed the title style(perf-test): cards for the Performance Tests steps, results and logs style(perf-test): one setup panel, Run in the header, console-style logs Nov 10, 2026
mariana953 and others added 8 commits November 10, 2026 22:40
Progress, run-duration and example-download bars use the neutral foreground
instead of yellow/blue. The setup panel, logs and system information use
fills instead of strokes, per the platform UI rules (no lines, white
accents).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Use the display title with no subtitle and the same page column as Builds
and Apps.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Use the design system card surface (bg-card) so Performance Tests matches
the other pages. The log console keeps its dark fill.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The page already follows the dashboard workspace picked in the sidebar.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Build, Workflow and Runs are three titled cards side by side. Stop
replaces Run in the header while a test is active. Results is one card
with Open folder and Export results in its title row, stat tiles and a
duration range row. System information and Logs are their own cards.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Header Run and Stop, the example workflow picker and the build select use the
32px, 13px control size of the Benchmarks toolbar. Section labels use the
subtle uppercase label colour and the Results and Logs toggles sit centred in
their header row. The picked workflow is one line with its path truncated (full
path on hover), the delete button is a square muted icon button, the results
actions are muted text buttons with 15px icons, and the empty results line is
13px muted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Build dropdown uses the design system Menu with radio items, like the
Apps filters and the Benchmarks toolbar. Options keep their source and
version line.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mariana953 pushed a commit to mariana953/Comfy-Desktop that referenced this pull request Nov 11, 2026
Build, Workflow and Runs sit in one card as three columns, so the setup reads
as one panel. The drop zone is shorter, with the icon beside the hint.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mariana953 pushed a commit to mariana953/Comfy-Desktop that referenced this pull request Nov 11, 2026
mariana953 and others added 4 commits November 11, 2026 16:14
…rkflows

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Build and Workflow use the shared filter dropdown trigger, the workflow
actions move into its menu, Run sits at the end of the setup row, and the
results, run label and system information share one card.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@mariana953 mariana953 changed the title style(perf-test): one setup panel, Run in the header, console-style logs style(perf-test): filter-style setup row and one results card Nov 11, 2026
@mariana953

Copy link
Copy Markdown
Contributor Author

Moved to #2018, based on main now that #1981 is closed. Same change.

@mariana953 mariana953 closed this Nov 11, 2026
mariana953 added a commit to mariana953/Comfy-Desktop that referenced this pull request Nov 12, 2026
Same change as Comfy-Org#2004, now based on main instead of the closed Comfy-Org#1981.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mariana953 added a commit to mariana953/Comfy-Desktop that referenced this pull request Nov 12, 2026
Same change as Comfy-Org#2004, now based on main instead of the closed Comfy-Org#1981.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
benceruleanlu added a commit to mariana953/Comfy-Desktop that referenced this pull request Nov 12, 2026
…t and failure states (Comfy-Org#2018)

* style(perf-test): filter-style setup row and one results card

Same change as Comfy-Org#2004, now based on main instead of the closed Comfy-Org#1981.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* style: page title matches the Builds title on main

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(perf-test): drag highlight, all-failed results, stopped-run card, status line

- The drop-zone highlight now clears when a dragged file leaves the
  Workflow control. dragleave.self ignored leaves that started on the
  trigger button, which fills the zone, so the outline stayed on.
- A test whose measured runs all failed has no statistics. The card now
  says "0 measured · 5 failed" instead of "Median run · ..." with no
  median.
- A run that is stopped or fails before it has results drops its
  "x on build" header, so the card goes back to its plain empty state.
- A workflow error no longer hides an example model download that is
  still running; both lines show.
- Removed the unused BrandedPageHeader actions slot, the unused
  numberLine.averageLabelAt, four orphaned performanceTest strings, and
  the stale DevPlatformWorkspaceSelector test stub.

Each new PanelApp test fails against the previous view and passes now.

* style(perf-test): bar chart per run, visible system information, matching export image

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* style(perf-test): bring back the card fill on Results and System information

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(perf-test): denser System information, results shown during cleanup, review fixes

- System information rows sit closer (12px row gap, 2px label to value); the exported image's facts are tighter too.
- Results and their bars show as soon as the run returns, even while logs are still being saved.
- System information shows when only systemInfo is present (no hardware snapshot).
- The Build trigger's accessible name includes the selected build.
- The logs summary finds the last line from the end instead of splitting the whole log.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(perf-test): full-width build and workflow rows in narrow windows

When the setup row is narrower than 36rem, Build and Workflow each take a full
row with the chevron at the end, and the run counts share the last row with
Run. Wider windows keep one row, or two rows (what to test, how to run).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor(perf-test): split the page into composables and components, fix run edge cases

The Performance Tests view was one 1,300-line component. It is now a thin page
over two composables and small components:

- usePerformanceTestRun owns the launch, submit, stop and telemetry lifecycle
  with an explicit stage (starting, running, finishing, stopping).
- usePerformanceTestWorkflow owns import, drop, delete and the example flow.
- lib/performanceTestResults holds the pure helpers (run-count clamping,
  formatting, duration rows, system facts) that the page and the PNG export
  now share, so the two can no longer drift.
- FilterButton is the shared "Label  value  v" trigger; the Build and Workflow
  menus, run-count fields, progress, summary, system facts, card and logs are
  components under views/performanceTest.

Behaviour fixes:

- A failed submission now stops the build and returns the page to Run, and
  the results card says why. Before, the build stayed up and only Stop showed.
- While the build boots the card says "Starting the build..." instead of
  claiming run 1 is executing.
- Results that arrive after the user pressed Stop are ignored.
- Run needs the chosen build to still be in the list.
- The page and the export read the saved results summary, and the export no
  longer needs a hardware snapshot.
- The export sizes text by estimated width (CJK aware), shrinks a slow median
  so it stays clear of the chart, and fits the title.
- "{workflow} on {build}" and the trigger names are single translated
  strings instead of concatenated fragments.
- The drop highlight only shows for file drags; the example download can be
  cancelled from the status line.
- Unused measuredRunCount, failedRunCount and resultContextOn strings removed.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Benjamin Lu <benjaminlu1107@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant