feat(ui): Wave 0 foundations for AntD → ui-core-components migration

.github/playwright/impact-map.json

@@ -7,7 +7,8 @@
     "playwright/e2e/Pages/DataInsight*.spec.ts",
     "playwright/e2e/Features/KnowledgeGraph.spec.ts",
     "playwright/e2e/Features/OntologyExplorerRdf.spec.ts",
-    "playwright/e2e/Features/OntologyImportRdf.spec.ts"
+    "playwright/e2e/Features/OntologyImportRdf.spec.ts",
+    "playwright/e2e/VisualRegression/**"
   ],
   "smoke": [
     {

.github/scripts/tests/test_playwright_ci_planning.py

@@ -1435,6 +1435,67 @@ def test_dedicated_rdf_specs_are_not_selected_by_the_main_workflow():
     )
 
 
+def test_visual_regression_specs_are_not_selected_by_the_main_workflow():
+    impact_map = json.loads(
+        (SCRIPTS.parents[0] / "playwright/impact-map.json").read_text()
+    )
+
+    assert "playwright/e2e/VisualRegression/**" in impact_map["delegatedSpecs"]
+
+
+def test_changed_visual_regression_spec_is_delegated_not_selected(tmp_path, monkeypatch):
+    selector = load_script("select_playwright_tests")
+    spec_dir = tmp_path / selector.UI_ROOT / "playwright/e2e/VisualRegression"
+    spec_dir.mkdir(parents=True)
+    spec_path = spec_dir / "entityDetails.spec.ts"
+    spec_path.write_text("test('visual', () => undefined);\n")
+    impact_map = tmp_path / "impact-map.json"
+    impact_map.write_text(
+        json.dumps(
+            {
+                "smoke": [],
+                "canary": [],
+                "delegatedSpecs": ["playwright/e2e/VisualRegression/**"],
+                "sharedInfrastructure": [],
+                "mappings": [],
+            }
+        )
+    )
+    changed = tmp_path / "changed.txt"
+    changed.write_text(
+        f"{selector.UI_ROOT}playwright/e2e/VisualRegression/entityDetails.spec.ts\n"
+    )
+    output = tmp_path / "selection.json"
+    monkeypatch.chdir(tmp_path)
+    monkeypatch.delenv("GITHUB_OUTPUT", raising=False)
+    monkeypatch.setattr(
+        sys,
+        "argv",
+        [
+            "select_playwright_tests.py",
+            "--event-name",
+            "pull_request_target",
+            "--changed-files",
+            str(changed),
+            "--impact-map",
+            str(impact_map),
+            "--output",
+            str(output),
+        ],
+    )
+
+    selector.main()
+
+    selection = json.loads(output.read_text())
+    assert selection["selectors"] == []
+    assert selection["directChangedSpecs"] == []
+    assert (
+        "playwright/e2e/VisualRegression/entityDetails.spec.ts"
+        in selection["delegatedChangedSpecs"]
+    )
+    assert selection["unmappedFiles"] == []
+
+
 def test_summary_reconciles_results_and_evaluates_performance_independently():
     workflow = (
         SCRIPTS.parents[0] / "workflows/playwright-postgresql-e2e.yml"

.github/workflows/playwright-visual.yml

@@ -0,0 +1,127 @@
+#  Copyright 2021 Collate
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#  http://www.apache.org/licenses/LICENSE-2.0
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+# This workflow runs the Playwright visual-regression suite
+# (openmetadata-ui/src/main/resources/ui/playwright/e2e/VisualRegression) against a MySQL-backed
+# OpenMetadata server. The environment-bootstrap steps (checkout, Java/Node setup, server build +
+# start, sample-data seeding) mirror .github/workflows/playwright-mysql-e2e.yml verbatim, minus the
+# pull_request_target-only labeling/gating steps and the shard matrix, which don't apply to a single
+# fast (~18 spec) suite triggered on regular `pull_request`.
+#
+# Rendering parity with the baselines: the VisualRegression snapshots were captured using the
+# mcr.microsoft.com/playwright:v1.57.0-jammy image (pinned to the @playwright/test version in
+# package.json). The e2e workflow installs Chromium on the bare ubuntu-latest runner and runs tests
+# there directly - fine for functional assertions, but font hinting/anti-aliasing on the bare runner
+# image differs enough from the official Playwright container to make pixel-diff assertions flaky or
+# outright wrong. So only the test-execution step below runs inside that exact container (via
+# `docker run --network host`, reaching the host-run server at localhost:8585); the
+# bootstrap/build/seed steps still run directly on the runner exactly as in the e2e workflow.
+name: playwright-visual-regression
+
+on:
+  pull_request:
+    paths:
+      - "openmetadata-ui/**"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+concurrency:
+  group: playwright-visual-${{ github.event.pull_request.number || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  playwright-visual-regression:
+    name: playwright-visual-regression
+    runs-on: ubuntu-latest
+    if: ${{ !github.event.pull_request.draft }}
+    steps:
+      - name: Free Disk Space (Ubuntu)
+        uses: jlumbroso/free-disk-space@54081f138730dfa15788a46383842cd2f914a1be # v1.3.1
+        with:
+          tool-cache: false
+          android: true
+          dotnet: true
+          haskell: true
+          large-packages: false
+          swap-storage: true
+          docker-images: false
+      - name: Checkout
+        uses: actions/checkout@v7
+        with:
+          persist-credentials: false
+
+      - name: Cache Maven Dependencies
+        id: cache-output
+        uses: actions/cache@v5
+        with:
+          path: ~/.m2
+          key: ${{ runner.os }}-maven-${{ hashFiles('**/pom.xml') }}
+          restore-keys: |
+            ${{ runner.os }}-maven-
+
+      - name: Setup Openmetadata Test Environment
+        uses: ./.github/actions/setup-openmetadata-test-environment
+        with:
+          python-version: "3.10"
+          args: "-d mysql"
+          ingestion_dependency: "all"
+          install-test-dependencies: "false"
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v5
+        with:
+          node-version-file: "openmetadata-ui/src/main/resources/ui/.nvmrc"
+
+      - name: Install dependencies
+        working-directory: openmetadata-ui/src/main/resources/ui/
+        run: yarn --ignore-scripts --frozen-lockfile
+
+      # Runs inside the same Playwright image the baselines were generated with (see header
+      # comment) instead of on the bare runner. `--network host` lets the container reach the
+      # OpenMetadata server that the bootstrap step above already started on localhost:8585.
+      # `--ipc host` follows Playwright's own Docker guidance to avoid Chromium running out of
+      # shared memory. node_modules were installed on the host above and are bind-mounted in, so
+      # only `npx` (bundled with the container's Node) is needed - no dependency on yarn/corepack
+      # being set up inside the image.
+      - name: Run visual regression suite
+        working-directory: openmetadata-ui/src/main/resources/ui
+        env:
+          PLAYWRIGHT_TEST_BASE_URL: http://localhost:8585
+        run: |
+          docker run --rm \
+            --network host \
+            --ipc host \
+            -e CI=true \
+            -e PLAYWRIGHT_TEST_BASE_URL \
+            -v "$PWD:/work" \
+            -w /work \
+            mcr.microsoft.com/playwright:v1.57.0-jammy \
+            npx playwright test --project=visual-regression
+
+      - name: Upload visual diff artifacts
+        if: failure()
+        uses: actions/upload-artifact@v4
+        with:
+          name: visual-regression-diffs
+          path: |
+            openmetadata-ui/src/main/resources/ui/playwright/output/test-results/**
+            openmetadata-ui/src/main/resources/ui/playwright/e2e/VisualRegression/__snapshots__/**
+          retention-days: 14
+
+      - name: Clean Up
+        if: always()
+        continue-on-error: true
+        run: |
+          cd ./docker/development
+          docker compose down --remove-orphans
+          sudo rm -rf "${PWD}/docker-volume"

docs/antd-migration/README.md

@@ -0,0 +1,24 @@
+# AntD → @openmetadata/ui-core-components migration
+
+Program design: see the Collate repo, `docs/superpowers/specs/2026-07-27-antd-untitled-migration-design.md`.
+
+## The sweep playbook (per antd component)
+
+1. **Gap check** — grep the *actually used* prop surface across openmetadata-ui,
+   collate-ui and collate-local-webserver/ui; compare against the
+   `ui-core-components` equivalent. Missing capability is added to
+   `openmetadata-ui-core-components` first, in its own PR, with unit tests.
+2. **Mapping guide** — copy `TEMPLATE.md` to `<component>.md` in this folder and
+   fill it in. The guide is the review contract for every PR of that sweep.
+3. **Codemod** (mechanical components only) — see `tooling/antd-codemods/`.
+4. **Chunked PRs** — ~20–40 files per PR, chunked by directory. Each PR also
+   deletes the dead code it orphans (grep/knip-proven only) and regenerates
+   visual baselines for intentionally changed pages.
+5. **Ledger update** — regenerate the ledger (see `tooling/antd-migration/`).
+
+## Rules encoded in every guide
+
+- Semantic Tailwind tokens only (`tw:bg-primary`), never raw palette or hex.
+- No `tw:ring-*` — borders use `border`/`outline` (see upstream `docs/colors.md` §2.3.1).
+- No string literals — `t('label.…')`, checking existing keys in both locale files first.
+- QueryBuilder files are owned by PR #29849 — do not touch.

docs/antd-migration/TEMPLATE.md

@@ -0,0 +1,31 @@
+# <AntdComponent> → <CoreComponent> mapping guide
+
+**Sweep status:** ledger row `<component>`
+**Core component:** `@openmetadata/ui-core-components` → `<import path>`
+**Codemod:** `tooling/antd-codemods/transforms/<name>.js` (or "manual — no codemod")
+
+## Import
+
+| Before | After |
+|---|---|
+| `import { X } from 'antd';` | `import { Y } from '@openmetadata/ui-core-components';` |
+
+## Prop mapping
+
+| antd prop (as used in repo) | core equivalent | notes |
+|---|---|---|
+| | | |
+
+## No direct equivalent — do this instead
+
+| antd usage | replacement pattern |
+|---|---|
+| | |
+
+## Before / after examples
+
+<!-- 2–3 real examples lifted from the codebase -->
+
+## CSS to delete with this sweep
+
+<!-- .ant-* override selectors and component .less files this sweep orphans -->

openmetadata-ui/src/main/resources/ui/knip.json

@@ -0,0 +1,13 @@
+{
+  "$schema": "https://unpkg.com/knip@5/schema.json",
+  "entry": ["src/index.tsx"],
+  "project": ["src/**/*.{ts,tsx}"],
+  "ignore": [
+    "src/generated/**",
+    "src/locale/**",
+    "**/*.test.{ts,tsx}",
+    "**/mocks/**",
+    "**/__mocks__/**"
+  ],
+  "ignoreDependencies": [".*"]
+}

openmetadata-ui/src/main/resources/ui/package.json

@@ -41,6 +41,8 @@
     "organize-imports": "yarn organize-imports:cli tsconfig[.*].json",
     "organize-imports:playwright": "yarn organize-imports:cli './playwright/tsconfig[.*].json'",
     "playwright:run": "playwright test",
+    "playwright:visual": "playwright test --project=visual-regression",
+    "playwright:visual:update": "playwright test --project=visual-regression --update-snapshots",
     "playwright:open": "playwright test --ui",
     "playwright:codegen": "playwright codegen",
     "generate:app-docs": "node generateApplicationDocs.js",

openmetadata-ui/src/main/resources/ui/playwright.config.ts

@@ -199,8 +199,19 @@ export default defineConfig({
         '**/IntakeForm.spec.ts',
         ...dedicatedStateTestIgnore,
         '**/DomainIsolation/**',
+        '**/VisualRegression/**',
       ],
     },
+    {
+      name: 'visual-regression',
+      testMatch: '**/VisualRegression/**/*.spec.ts',
+      dependencies: ['setup', 'entity-data-setup'],
+      use: {
+        ...devices['Desktop Chrome'],
+        viewport: { width: 1440, height: 900 },
+        storageState: 'playwright/.auth/admin.json',
+      },
+    },
     // Only register the h2 project when explicitly opted in. Always-on registration would force
     // Playwright to do discovery for it on every default run even though its spec files are
     // skipped — small cost, but pointless when the h2 server isn't running.

openmetadata-ui/src/main/resources/ui/playwright/e2e/VisualRegression/entityDetails.spec.ts

@@ -0,0 +1,71 @@
+/*
+ *  Copyright 2026 Collate.
+ *  Licensed under the Apache License, Version 2.0 (the "License");
+ *  you may not use this file except in compliance with the License.
+ *  You may obtain a copy of the License at
+ *  http://www.apache.org/licenses/LICENSE-2.0
+ *  Unless required by applicable law or agreed to in writing, software
+ *  distributed under the License is distributed on an "AS IS" BASIS,
+ *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ *  See the License for the specific language governing permissions and
+ *  limitations under the License.
+ */
+import { expect, Page, test } from '@playwright/test';
+import { TableClass } from '../../support/entity/TableClass';
+import { getDefaultAdminAPIContext } from '../../utils/common';
+import {
+  gotoForScreenshot,
+  SCREENSHOT_OPTS,
+} from '../../utils/visualRegression';
+
+/**
+ * Entity names are randomized by the support classes (TableClass generates a
+ * `pw-database-service-<uuid>` service, a `pw-table-<fullUuid>` table and
+ * `<uuid>`-suffixed column names), so mask every element that renders a
+ * name; the layout, tabs and schema table are the subject under test.
+ *
+ * - `entity-header-display-name` / `entity-header-name`: display name + name
+ *   lines in EntityHeaderTitle (both contain random uuids).
+ * - `breadcrumb`: service > database > schema names.
+ * - `column-name`: each column-name link in SchemaTable (uuid-suffixed).
+ * - `.custom-property-right-panel-container`: right-panel
+ *   CustomPropertyTable (no testid in the right-panel variant) — property
+ *   names come from seeded custom properties with random `cp-<uuid>` names.
+ */
+const NAME_MASKS = [
+  '[data-testid="entity-header-display-name"]',
+  '[data-testid="entity-header-name"]',
+  '[data-testid="breadcrumb"]',
+  '[data-testid="column-name"]',
+  '.custom-property-right-panel-container',
+];
+
+const table = new TableClass();
+
+test.beforeAll(async ({ browser }) => {
+  const { apiContext, afterAction } = await getDefaultAdminAPIContext(browser);
+  await table.create(apiContext);
+  await afterAction();
+});
+
+test.afterAll(async ({ browser }) => {
+  const { apiContext, afterAction } = await getDefaultAdminAPIContext(browser);
+  await table.delete(apiContext);
+  await afterAction();
+});
+
+const maskFor = (page: Page) =>
+  NAME_MASKS.map((selector) => page.locator(selector));
+
+test('table entity details (schema tab) matches baseline', async ({ page }) => {
+  await gotoForScreenshot(
+    page,
+    `/table/${encodeURIComponent(
+      table.entityResponseData?.fullyQualifiedName ?? ''
+    )}`
+  );
+  await expect(page).toHaveScreenshot('table-details-schema.png', {
+    ...SCREENSHOT_OPTS,
+    mask: maskFor(page),
+  });
+});