cabloy 5.1.61 → 5.1.63

package/cabloy-docs/frontend/table-resource-crud-cookbook.md

@@ -0,0 +1,496 @@
+# Table + Resource CRUD Cookbook
+
+This cookbook explains how Zova Table fits into Cabloy Basic’s resource-driven CRUD list pages.
+
+Use this page when your next question is practical rather than framework-neutral:
+
+- how should I assemble a standard resource list page?
+- where do filter, toolbar, table, and pager blocks come from?
+- how does `ModelResource` feed schema and data into `ZTable`?
+- where should list-page customization happen without breaking the existing runtime?
+
+Use this page together with:
+
+- [Table Guide](/frontend/table-guide)
+- [TableCell Authoring Cookbook](/frontend/table-cell-cookbook)
+- [Zova Table Under the Hood](/frontend/zova-table-under-the-hood)
+- [Model Resource Owner Pattern](/frontend/model-resource-owner-pattern)
+- [Rest Resource Under the Hood](/frontend/rest-resource-under-the-hood)
+- [Rest Resource Source Reading Map](/frontend/rest-resource-source-reading-map)
+- [Backend Metadata to Frontend Table Actions](/fullstack/backend-metadata-to-frontend-table-actions)
+- [Tutorial 2: Create Your First CRUD](/fullstack/tutorial-2-first-crud)
+- [Tutorial 3: Frontend Metadata Sharing](/fullstack/tutorial-3-frontend-metadata-sharing)
+- [Tutorial 5: Backend Contract Sharing](/fullstack/tutorial-5-backend-contract-sharing)
+
+## What this cookbook is for
+
+If your next question is how the resource list page runtime is assembled under the hood, continue with [Resource List Page Deep Dive](/frontend/resource-list-page-deep-dive). If your next question is why row or bulk actions are visible or hidden, continue with [Table Action Visibility and Permission Flow Guide](/frontend/table-action-visibility-permission-flow-guide).
+
+A Cabloy Basic CRUD list page is usually not authored as:
+
+- one page-local table component
+- one page-local fetch call
+- one page-local pager state
+- one page-local filter form
+
+The more typical Zova-native shape is:
+
+- one resource-owner model
+- one page block that owns query and permissions
+- several reusable blocks for filter, bulk toolbar, table, and pager
+- schema and metadata driving the visible UI
+
+That means the practical CRUD question becomes less:
+
+- “how do I hand-build one list page?”
+
+and more:
+
+- “how do I plug my resource into the existing resource-page runtime?”
+
+## The shortest correct mental model
+
+If you only remember one idea, remember this one:
+
+> In Cabloy Basic CRUD list pages, `ZTable` is usually one block inside a larger resource-page runtime owned by `basic-page:blockPage` and backed by `ModelResource`.
+
+That larger runtime usually owns:
+
+- resource bootstrap
+- list query state
+- filter state
+- permissions
+- row schema
+- paged response state
+- coordination among filter, table, and pager
+
+## One running example through this guide: Student list page
+
+To keep the guide concrete, the examples below all use the same teaching thread:
+
+- resource: `training-student:student`
+- list-page concerns: filter, row actions, create action, pagination
+- metadata sources: backend row DTO and row-action metadata
+
+This is the same general shape already used in the built-in specimen:
+
+- `vona/src/suite-vendor/a-test/modules/test-rest/src/dto/productSelectResItem.tsx`
+
+## Step 1: Start from the generated CRUD contract, not from page-local UI
+
+For a standard CRUD list page, begin with the generated backend thread.
+
+The first useful entrypoint is still the CRUD generator:
+
+```bash
+npm run vona :tools:crud student -- --module=training-student
+```
+
+That generated thread already gives you the important contract anchors such as:
+
+- backend entity
+- select request DTO
+- select response DTO
+- select row-item DTO
+
+A practical rule is:
+
+- treat the generated resource contract as the list-page foundation
+- refine schema and metadata before hand-building frontend list code
+
+For the first generated CRUD walkthrough, see [Tutorial 2: Create Your First CRUD](/fullstack/tutorial-2-first-crud).
+
+## Step 2: Understand the block chain for a standard list page
+
+A standard resource list page is usually declared through backend-owned block metadata like this:
+
+```tsx
+ZovaRender.block('basic-page:blockPage', {
+  blocks: [
+    ZovaRender.block('basic-page:blockFilter'),
+    ZovaRender.block('basic-page:blockToolbarBulk', {
+      actions: [ZovaRender.tableActionBulk('basic-table:actionCreate')],
+    }),
+    ZovaRender.block('basic-page:blockTable'),
+    ZovaRender.block('basic-page:blockPager'),
+  ],
+});
+```
+
+That declaration is important because it shows the intended authoring direction clearly:
+
+- backend or shared contract metadata chooses the block composition
+- frontend block implementations provide the runtime behavior
+- the page is assembled from existing building blocks rather than from ad hoc local wiring
+
+A practical reading takeaway is:
+
+- **the block list is the public page composition surface**
+- **the frontend block controllers are the runtime owners**
+
+## Step 3: Know what `basic-page:blockPage` owns
+
+The main runtime owner for a resource list page is:
+
+- `basic-page:blockPage`
+
+Its controller is responsible for:
+
+- resolving the `ModelResource` selector for the current resource
+- preparing page-level JSX/CEL support
+- creating query state
+- loading schema and paged data
+- exposing `data`, `schemaRow`, `schemaFilter`, `permissions`, and `paged`
+- refreshing table metadata when permissions change
+
+This is one of the most important architecture facts about CRUD pages.
+
+The page block does not only render layout. It owns the list-page resource state.
+
+A practical rule is:
+
+- if the concern is about resource query, schema, or permissions, start debugging in `blockPage`
+- if the concern is about one visual block, continue to the corresponding block controller next
+
+## Step 4: Let `blockFilter` own filter-form behavior
+
+The standard filter block is:
+
+- `basic-page:blockFilter`
+
+This block uses `ZForm` with:
+
+- `schema={$$page.schemaFilter}`
+- `schemaScene="filter"`
+- inline layout
+- page-owned filter data
+
+Its job is not to duplicate the list query logic.
+
+Its job is to:
+
+- render the filter form from resource filter schema
+- normalize submitted filter data
+- call `$$page.onFilter(...)`
+
+A practical rule is:
+
+- if you need to refine which filter fields exist, start from backend filter-side metadata and schema
+- if you need to refine how filter submission affects the list, inspect `blockFilter` and `blockPage.onFilter(...)`
+
+## Step 5: Let `blockToolbarBulk` own bulk-action display
+
+The standard bulk toolbar block is:
+
+- `basic-page:blockToolbarBulk`
+
+This block is intentionally thin.
+
+Its main job is to:
+
+- inspect the configured bulk actions
+- filter them by permission
+- render each action through the page JSX runtime
+
+That means the toolbar block usually should not become a second action-semantics layer.
+
+A practical rule is:
+
+- if the page only needs standard create or other bulk actions, keep the existing block and adjust the metadata
+- if an action is reusable, prefer a reusable action render resource rather than page-local ad hoc code
+
+## Step 6: Let `blockTable` bridge page state into `ZTable`
+
+The standard table block is:
+
+- `basic-page:blockTable`
+
+Its bridge role is very clear:
+
+- render `ZTable`
+- pass page data into `data`
+- pass row schema into `schema`
+- pass page CEL scope into `tableScope`
+- capture the table controller through `controllerRef`
+
+That means `blockTable` does not own list fetching or pagination logic.
+
+It owns the handoff from page resource state to the reusable table runtime.
+
+A practical rule is:
+
+- if the table shows the wrong rows or wrong schema, inspect `blockPage` first
+- if the handoff from page to table looks wrong, inspect `blockTable`
+- if one column or cell renders incorrectly, continue into `a-table` and `tableCell` logic
+
+## Step 7: Let `blockPager` own page navigation UI
+
+The standard pager block is:
+
+- `basic-page:blockPager`
+
+This block reads the paged response from `$$page.paged` and renders:
+
+- total items
+- total pages
+- previous-page button
+- current page indicator
+- next-page button
+
+It delegates actual page movement back to the page block through `gotoPage(...)`.
+
+That means the pager block owns the visible pager UI, while the page block still owns pagination state.
+
+A practical rule is:
+
+- if the pager UI looks wrong, inspect `blockPager`
+- if paging requests or totals are wrong, inspect `blockPage` and the resource model/query side
+
+## Step 8: Know where the table columns and row actions really come from
+
+In a standard CRUD list page, the table contents are usually not first defined in the block controller.
+
+They come from the backend resource contract, especially row DTO metadata.
+
+A representative example is:
+
+```tsx
+@Api.field(
+  v.title($locale('Operations')),
+  ZovaRender.order(1, 'max'),
+  ZovaRender.cell('basic-table:actionOperationsRow', {
+    actions: [
+      ZovaRender.tableActionRow('basic-table:actionUpdate'),
+      ZovaRender.tableActionRow('basic-table:actionDelete'),
+    ],
+  }),
+)
+_operationsRow?: unknown;
+```
+
+This is the practical reverse-sharing model:
+
+- the backend row contract chooses the table cell resource identity
+- the frontend table runtime resolves that resource
+- the page block and table block do not need page-local hard-coded row-action wiring
+
+For the built-in metadata-sharing teaching path, see [Tutorial 3: Frontend Metadata Sharing](/fullstack/tutorial-3-frontend-metadata-sharing).
+
+For the forward-chain row-action teaching path, see [Tutorial 5: Backend Contract Sharing](/fullstack/tutorial-5-backend-contract-sharing).
+
+## Step 9: The common extension points
+
+For standard CRUD list pages, most customization should happen in one of these places.
+
+### Extension point A: backend row schema metadata
+
+Use this when:
+
+- a field should change order
+- a column should be visible or hidden
+- a field should use a different built-in or custom `tableCell`
+
+This is usually the first and best extension point.
+
+### Extension point B: backend block composition metadata
+
+Use this when:
+
+- the page should add or remove a standard block
+- the page should include bulk actions
+- the page should change list-page composition order
+
+### Extension point C: custom `tableCell` resources
+
+Use this when:
+
+- a field needs business-specific table rendering
+- one action cell should be reusable across pages
+- one operations row should orchestrate several child actions
+
+For concrete patterns, continue with [TableCell Authoring Cookbook](/frontend/table-cell-cookbook).
+
+### Extension point D: page-level block implementation
+
+Use this only when:
+
+- the existing block runtime is structurally insufficient
+- the behavior really belongs to the block runtime rather than to metadata
+- you have already confirmed that schema, metadata, and cell resources are not enough
+
+A practical rule is:
+
+- prefer metadata and cell resources before changing block controllers
+
+## Step 10: Distinguish list page from entry page
+
+A common source-reading mistake is to mix `basic-page` list runtime with `basic-pageentry` form runtime.
+
+They are related but they do not own the same concerns.
+
+### `basic-page` list runtime
+
+Typical responsibilities:
+
+- filter form
+- bulk toolbar
+- table
+- pager
+- list query
+- row schema and permissions
+
+### `basic-pageentry` entry runtime
+
+Typical responsibilities:
+
+- view/create/edit form scene
+- form schema and form data
+- form submission
+- page dirty state and page title
+
+A practical rule is:
+
+- if the page is about rows in a list, stay in `basic-page`
+- if the page is about one entry form, stay in `basic-pageentry`
+
+This distinction prevents a lot of source-reading confusion.
+
+## Pattern 1: The standard generated CRUD list page
+
+Use this pattern when:
+
+- the CRUD generator already created the resource thread
+- built-in page blocks are enough
+- backend metadata can express the list behavior
+
+Recommended path:
+
+1. generate the CRUD thread
+2. inspect row DTO metadata
+3. keep the standard block chain
+4. verify the list page in Admin
+
+This should be the default choice for first CRUD pages.
+
+## Pattern 2: Add a filter field through contract refinement
+
+Use this pattern when:
+
+- the list page needs one more search field
+- the field belongs in the resource contract
+- the filter form should stay schema-driven
+
+Recommended path:
+
+1. refine backend field and query DTO metadata
+2. verify `schemaFilter`
+3. reuse `basic-page:blockFilter`
+4. avoid page-local manual filter UI unless really necessary
+
+## Pattern 3: Add row actions through backend metadata
+
+Use this pattern when:
+
+- the row should expose update/delete/summary/deleteForce-like actions
+- the action identity belongs in the contract
+- the frontend should reuse `tableCell` resources
+
+Recommended path:
+
+1. define or refine backend row-action metadata
+2. point the action cell to built-in or custom table-cell resources
+3. if needed, add generated API and thin model helpers
+4. verify the list page row actions in Admin
+
+## Pattern 4: Add a business-specific cell for one field
+
+Use this pattern when:
+
+- one field such as `level`, `status`, or `score` needs module-owned UI behavior
+- the list page should remain resource-driven overall
+
+Recommended path:
+
+1. create a custom `tableCell` bean
+2. point backend field metadata to it
+3. keep `basic-page:blockTable` and `ZTable` unchanged
+
+This is the preferred path when the customization belongs to a field rather than to page structure.
+
+## Common mistakes to avoid
+
+### Mistake 1: Rebuilding list-page fetch state inside the table block
+
+`blockTable` is only the page-to-table bridge. Query ownership lives in `blockPage` and `ModelResource`.
+
+### Mistake 2: Hand-writing table columns before checking row schema metadata
+
+For most CRUD pages, the row schema is the first source of truth.
+
+### Mistake 3: Mixing list-page and entry-page concerns
+
+`basic-page` and `basic-pageentry` are related, but they own different runtime responsibilities.
+
+### Mistake 4: Modifying block controllers when metadata or `tableCell` resources are enough
+
+Most business customization should happen before block-controller changes.
+
+### Mistake 5: Treating row actions as only frontend-local UI
+
+In Cabloy, row actions often belong to a broader contract chain involving backend metadata, generated API, model helpers, and table-cell resources.
+
+## A practical authoring order
+
+If you want the shortest path to a real CRUD list page, use this order:
+
+1. generate or confirm the backend CRUD contract thread
+2. confirm the resource-owner model already exposes the required schemas and permissions
+3. keep the standard `blockPage -> blockFilter -> blockToolbarBulk -> blockTable -> blockPager` chain
+4. refine backend row metadata for visible columns and row actions
+5. reuse built-in `tableCell` resources first
+6. add custom `tableCell` resources only where the UI becomes business-specific
+7. change block controllers only when the existing runtime is structurally insufficient
+8. continue with [Table Guide](/frontend/table-guide) for the public `ZTable` surface
+9. continue with [TableCell Authoring Cookbook](/frontend/table-cell-cookbook) for custom cell patterns
+10. continue with [Backend Metadata to Frontend Table Actions](/fullstack/backend-metadata-to-frontend-table-actions) when the visible row actions belong to a larger contract-loop chain
+11. continue with [Zova Table Under the Hood](/frontend/zova-table-under-the-hood) when you want the controller-level runtime explanation
+12. continue with [Zova Table Source Reading Map](/frontend/zova-table-source-reading-map) when you need targeted file-order guidance for the table runtime
+
+## Verification checklist
+
+When authoring or documenting a resource CRUD list page, verify in this order:
+
+1. confirm the generated or existing backend CRUD contract files are present
+2. confirm the block composition still matches the intended standard page chain
+3. confirm row schema metadata points to the intended built-in or custom render resources
+4. make sure the local dev workflow is running:
+
+   ```bash
+   npm run dev
+   ```
+
+5. open `http://localhost:7102/admin/`
+6. enter the target list page and verify:
+   - filter works
+   - bulk actions render correctly
+   - table columns and row actions match metadata
+   - pager updates the list correctly
+7. if you changed docs, build the docs site:
+
+   ```bash
+   npm run docs:build
+   ```
+
+## Final takeaway
+
+A good Cabloy Basic CRUD list page is usually not a custom page-local table stack.
+
+It is a resource-driven composition of:
+
+- one resource-owner model
+- one page block
+- one schema-driven table runtime
+- several reusable page blocks
+- optional custom `tableCell` resources
+
+That is the path that keeps CRUD list pages consistent, extensible, and aligned with the Zova-native architecture.