feat(config): turn the Form View on by default by yangzhang75 · Pull Request #8528 · apache/texera

yangzhang75 · 2026-09-13T03:49:25Z

What changes were proposed in this PR?

Closes #8537. The closing PR of the Form View stack (parent issue #8011). Every piece of the feature has landed behind form-view-enabled = false; this flips the default to true, with the GuiConfigSpec assertion updated to match. Nothing else changes. A deployment that wants the feature off sets GUI_WORKFLOW_WORKSPACE_FORM_VIEW_ENABLED=false, which the config already reads.

On main now that #8456 (the entry points) has merged; the branch's single commit is the whole change.

Any related issues, documentation, discussions?

Closes #8537. Part of the Form View feature (parent issue #8011).

How was this PR tested?

GuiConfigSpec asserts the new default. The frontend specs that exercise the entry points already run with the flag on (the shared config mock defaults it off and each such spec enables it), so they are unaffected by the default.

Was this PR authored or co-authored using generative AI tooling?

Co-authored with Claude Code.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FVvP3ttj22f9LB4p9u2anY

github-actions · 2026-09-13T03:49:52Z

Automated Reviewer Suggestions

Based on the git blame history of the changed files, we recommend the following reviewers:

Contributors with relevant context: @Neilk1021
You can notify them by mentioning @Neilk1021 in a comment.

codecov-commenter · 2026-09-13T03:50:52Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.90%. Comparing base (5042d96) to head (a3c3639).

Additional details and impacted files

@@            Coverage Diff            @@
##               main    #8528   +/-   ##
=========================================
  Coverage     95.90%   95.90%           
  Complexity      459      459           
=========================================
  Files           589      589           
  Lines         33953    33953           
  Branches       4246     4246           
=========================================
  Hits          32564    32564           
  Misses          811      811           
  Partials        578      578

Flag	Coverage Δ	*Carryforward flag
access-control-service	`80.18% <ø> (ø)`
agent-service	`99.32% <ø> (ø)`	Carriedforward from 5042d96
amber	`90.82% <ø> (ø)`
computing-unit-managing-service	`77.14% <ø> (ø)`
config-service	`87.12% <ø> (ø)`
file-service	`83.65% <ø> (ø)`	Carriedforward from 5042d96
frontend	`96.72% <ø> (ø)`	Carriedforward from 5042d96
notebook-migration-service	`83.73% <ø> (ø)`
pyamber	`98.47% <ø> (ø)`	Carriedforward from 5042d96
workflow-compiling-service	`74.09% <ø> (ø)`

*This pull request uses carry forward flags. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

github-actions · 2026-09-13T03:55:49Z

⚠️ Benchmark changes need a look

🟢 4 better · 🔴 2 worse · ⚪ 9 noise (<±5%) · 0 without baseline

Compared against main 5042d96 benchmarked on this same runner, so the delta is largely free of cross-runner hardware noise. The "7d avg" column still reflects the gh-pages dashboard. Treat <±5% as noise unless repeated.

Dashboard · Run

	config	throughput	MB/s	latency	max Δ latest / 7d
🔴	bs=10 sw=10 sl=64	426	0.26	22,708/27,974/27,974 us	🟢 -17.7% / 🔴 +91.2%
⚪	bs=100 sw=10 sl=64	954	0.582	105,563/117,833/117,833 us	⚪ within ±5% / 🔴 +18.4%
🟢	bs=1000 sw=10 sl=64	1,092	0.666	917,404/970,561/970,561 us	🟢 -6.3% / ⚪ within ±5%

Baseline details

Latest main 5042d96 from same runner

config	metric	PR	latest main	7d avg	Δ latest	Δ 7d
bs=10 sw=10 sl=64	throughput	426 tuples/sec	448 tuples/sec	838.23 tuples/sec	-4.9%	-49.2%
bs=10 sw=10 sl=64	MB/s	0.26 MB/s	0.274 MB/s	0.512 MB/s	-5.1%	-49.2%
bs=10 sw=10 sl=64	p50	22,708 us	19,650 us	11,878 us	+15.6%	+91.2%
bs=10 sw=10 sl=64	p95	27,974 us	33,984 us	14,690 us	-17.7%	+90.4%
bs=10 sw=10 sl=64	p99	27,974 us	33,984 us	18,483 us	-17.7%	+51.4%
bs=100 sw=10 sl=64	throughput	954 tuples/sec	960 tuples/sec	1,095 tuples/sec	-0.6%	-12.9%
bs=100 sw=10 sl=64	MB/s	0.582 MB/s	0.586 MB/s	0.668 MB/s	-0.7%	-12.9%
bs=100 sw=10 sl=64	p50	105,563 us	104,682 us	92,901 us	+0.8%	+13.6%
bs=100 sw=10 sl=64	p95	117,833 us	114,751 us	99,480 us	+2.7%	+18.4%
bs=100 sw=10 sl=64	p99	117,833 us	114,751 us	112,580 us	+2.7%	+4.7%
bs=1000 sw=10 sl=64	throughput	1,092 tuples/sec	1,110 tuples/sec	1,126 tuples/sec	-1.6%	-3.0%
bs=1000 sw=10 sl=64	MB/s	0.666 MB/s	0.678 MB/s	0.687 MB/s	-1.8%	-3.1%
bs=1000 sw=10 sl=64	p50	917,404 us	891,479 us	912,010 us	+2.9%	+0.6%
bs=1000 sw=10 sl=64	p95	970,561 us	1,035,408 us	950,922 us	-6.3%	+2.1%
bs=1000 sw=10 sl=64	p99	970,561 us	1,035,408 us	983,333 us	-6.3%	-1.3%

Raw CSV

config_idx,batch_size,schema_width,string_len,num_batches,total_ms,total_tuples,total_bytes,tuples_per_sec,mb_per_sec,lat_p50_us,lat_p95_us,lat_p99_us
0,10,10,64,20,469.44,200,128000,426,0.260,22707.62,27973.83,27973.83
1,100,10,64,20,2095.71,2000,1280000,954,0.582,105563.11,117832.91,117832.91
2,1000,10,64,20,18317.64,20000,12800000,1092,0.666,917404.45,970561.44,970561.44

### What changes were proposed in this PR? Closes apache#8028. Part of the Form View stack (parent issue apache#8011), on main now that apache#8516 and apache#8517 have merged. The review commit is the branch's single commit. Wires the Form View entry points. The flag stays off here; the stack's closing PR, apache#8528, flips it. - The dashboard opens a workflow in its `default_view` (form or canvas), with a toggle that persists the choice; the deep link goes to the existing `/workflow/:id/form` route. Both renderers of the dashboard, the list row and the card, follow one shared rule (`default-view-landing.ts`: mark, deep link, toggle), so switching the view mode does not lose the entry point. The toggle is offered only with WRITE access, which the endpoint requires, and the handler checks the same rule rather than trusting the template; it is a proper toggle button (constant accessible name, state in `aria-pressed`, the hover title spelling out what a click does); hub links are left untouched. - The canvas menu gains the same Canvas / Form View switch the form already shows, so the two views swap in place. It saves first and hands over only once the save has completed: the switch is a full-page load, which aborts a request still in flight. Two more things the hand-over must not lose: an autosave already in flight when the switch is clicked (`WorkflowPersistService` now sends saves one at a time and in call order, at the one place every save goes through, so the switch's save lands and completes after it; each caller still gets only its own result and a failed save does not hold up the next), and an edit made while the switch's save is out (the page stays editable until the load; `workflowChanged` marks it and the hand-over saves once more before leaving). A reader, who cannot save, goes straight over. A workflow the canvas holds but has never saved (the default id) is created by that save, and the hand-over opens the id the save answered with. On the card the toggle sits in the always-visible action footer, in the same slot as on the row (right after Detail); the row's hover-revealed action group also appears while the row has the keyboard focus, so the toggle can be reached without a pointer there too. A second click while the hand-over is in progress is a no-op. A failed save keeps the user on the canvas with the error shown. Every workflow offers both views whenever the flag is on: `default_view` only decides the landing view, and neither view gates the other. - Download/upload round-trips `defaultView` as a sibling key next to the workflow content, in one shared export shape (`exportedWorkflow`) used by the dashboard download and the canvas menu's export alike; an old export without the key imports unchanged. - The computing unit the user picks is remembered per workflow (localStorage) so it survives switching between the two views; a unit selected on load (the remembered one, the last execution's, a running one) is derived rather than chosen and is not stored, or a derived unit would later outrank a fresher last execution. On load the remembered unit is honoured only once the unit list has arrived and still holds it: a unit that has since been terminated is forgotten and the last execution's unit is used instead, and a decision still pending when the workflow changes underneath it is dropped (the remembered-unit check, the last-execution lookup and its running-unit fallback alike). ### Any related issues, documentation, discussions? Closes apache#8028. Part of the Form View feature (parent issue apache#8011). ### How was this PR tested? Unit tests (vitest) cover the menu's Canvas / Form View switch through the DOM (absent with the flag off, Canvas pressed, Form View handing over, hidden while an older version is displayed), the row's and the card's default-view behavior (mark and deep link, hub link untouched, flag off leaves the dashboard as today, WRITE-only toggle in the DOM, toggle on / off / failed request / no cached row), the menu switch (navigates only once the save completes, stays on the canvas with the error when it fails, saves once more when an edit lands while its save is out, takes a reader straight over without a save, ignores a second click mid hand-over), the canvas export carrying `defaultView` next to the content and omitting it when unset, the hand-over opening the id the save assigned when the canvas held a never-saved workflow, the persist service sending saves one at a time in order with each caller getting its own result and a failure not holding up the next, the dashboard toggle handlers refusing without WRITE access, the toggle's aria-pressed following the state, the export/import round-trip including a legacy file without `defaultView`, and the computing-unit recall (waits for the first non-empty unit list, forgets a terminated unit and falls back, drops a stale decision after the workflow changed, a late last-execution answer or fallback included, positive-integer validation, storage failures, only an explicit pick remembered). Each new guard was deletion-checked (removing it turns the corresponding test red). eslint, prettier and the production (AOT) build pass; every changed line, template lines included, is statement and function covered. #### Video ##### 1. Default view on the dashboard (row and card toggle, Form View icon, deep link into the form, toggle off again, no toggle without write access) https://github.com/user-attachments/assets/04f71ba5-27ad-489d-a5be-51cdd8d811aa ##### 2. Canvas / Form View switch (save first, then the hand-over; and back) https://github.com/user-attachments/assets/39c3d395-6329-403d-9719-ca7ef179889f ##### 3. Computing unit remembered across the switch https://github.com/user-attachments/assets/6fe70b5f-d936-4333-bcf0-b3fa44f84323 ##### 4. Download / upload keeps the default view https://github.com/user-attachments/assets/0519bfcd-0677-478f-866c-6eb0350d842c ##### 5. Default view on the card view (the same toggle in the card's action row, Form View icon, deep link into the form) https://github.com/user-attachments/assets/471192b3-7935-4da9-be87-fa08a7dd6c13 ### Was this PR authored or co-authored using generative AI tooling? Yes. Generated-by: Claude Code (Claude Fable 5.1, Anthropic). Co-authored with Claude, reviewed line by line by the author before submission. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01FVvP3ttj22f9LB4p9u2anY Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>

Every piece of the Form View (parent issue apache#8011) has landed behind form-view-enabled = false. Flip the default to true, with the GuiConfigSpec assertion updated to match. GUI_WORKFLOW_WORKSPACE_FORM_VIEW_ENABLED=false turns it off for a deployment. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FVvP3ttj22f9LB4p9u2anY

The eyeserver deployment is the Form View: a biologist opens a workflow's form, picks their data and runs it. Upstream ships the flag off until the rollout PR (apache#8528) flips it, so this branch turns it on and the config spec asserts that. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FVvP3ttj22f9LB4p9u2anY

github-actions Bot assigned yangzhang75 Sep 13, 2026

github-actions Bot added frontend Changes related to the frontend GUI common labels Sep 13, 2026

yangzhang75 force-pushed the formview-pr17 branch from f14616d to 859b50f Compare September 13, 2026 06:29

yangzhang75 mentioned this pull request Sep 13, 2026

feat(gui): wire the Form View entry points #8456

Merged

yangzhang75 force-pushed the formview-pr17 branch 2 times, most recently from 32f93aa to e1a0c0a Compare September 13, 2026 23:36

This was referenced Sep 13, 2026

feat(workflow-form): add the Form View entry points and remember the computing unit #8028

Closed

Form View: a form-based view for running a workflow #8011

Open

yangzhang75 force-pushed the formview-pr17 branch from e1a0c0a to a3c3639 Compare September 14, 2026 00:20

yangzhang75 marked this pull request as ready for review September 14, 2026 00:20

github-actions Bot removed the frontend Changes related to the frontend GUI label Sep 14, 2026

yangzhang75 mentioned this pull request Sep 14, 2026

feat(config): turn the Form View on by default #8537

Open

6 tasks

Navigation Menu

Uh oh!

feat(config): turn the Form View on by default - #8528

feat(config): turn the Form View on by default#8528
yangzhang75 wants to merge 1 commit into
apache:mainfrom
yangzhang75:formview-pr17

yangzhang75 commented Sep 13, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Sep 13, 2026 •

edited

Loading

Uh oh!

codecov-commenter commented Sep 13, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Sep 13, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

yangzhang75 commented Sep 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What changes were proposed in this PR?

Any related issues, documentation, discussions?

How was this PR tested?

Was this PR authored or co-authored using generative AI tooling?

Uh oh!

github-actions Bot commented Sep 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Automated Reviewer Suggestions

Uh oh!

codecov-commenter commented Sep 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

github-actions Bot commented Sep 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

⚠️ Benchmark changes need a look

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

yangzhang75 commented Sep 13, 2026 •

edited

Loading

github-actions Bot commented Sep 13, 2026 •

edited

Loading

codecov-commenter commented Sep 13, 2026 •

edited

Loading

github-actions Bot commented Sep 13, 2026 •

edited

Loading