POC feat: move filters and sorting to backend

[Koc]

This PR is implemented on top of #1565. It contains all possible functionality added in V2. It moves table row listing, filtering, sorting and search to the backend so large tables can paginate without double-renders or stale page numbers.

I will try to extract some independent parts to separate PRs to simplify code review process.

What changed

• Backend-powered pagination, filtering, sorting and search via v2 OCS endpoints for tables, views and public shares: GET /api/2/{tables|views}/{nodeId}/rows, .../rows/count and .../rows/export, plus GET /api/2/public/{token}/rows, .../rows/count and .../rows/export. Only the requested page is loaded, and client-side sorting/filtering in the table grid is now removed completely.
• Column sort eligibility is now driven by a BACKEND_SORTABLE_TYPES allow-list in the frontend; only text-line, text-link, selection, selection-multi, number, number-stars, number-progress, datetime-date, datetime-time and datetime columns show the sorting UI.
• Table and view state (filters, sorting, search, page, rows per page and selected row IDs) is persisted in the URL query string, so links can be shared with a predefined state. Example url https://nextcloud.test/apps/tables/#/table/9?filter=[{%22columnId%22:47,%22operator%22:%22contains%22,%22value%22:%22Sa%22}]&search=en&page=2&perPage=25
• Selected rows can be shared via a rowIds URL query parameter. The selection menu has a "Share selected rows" action that copies a shareable link and temporarily shows an "Open" button. Example url https://nextcloud.test/apps/tables/#/table/9?rowIds=4025%2C3970%2C3989
• CSV export now uses a dedicated .../rows/export endpoint (private and public); the backend builds the CSV from all matching rows (with filters, sort, search and rowIds) and returns it as a file.
• Backend case-insensitive text search across raw stored cell values.
• Text filters (contains, does-not-contain, begins-with, ends-with) are currently case-sensitive; case-insensitive filter matching is implemented in Fix: Make case-insensitive search for the view filters #2816.
• Backend support for the "contains items" filter on selection columns, including multi-selection.
• New reload/refresh button in the table header with a loading spinner; the empty state is shown when no rows match.
• Pagination resets to page 1 on filter/search changes and stays in sync with the table.

Known limitations

• Backend search currently matches the raw stored cell values, so for selection, usergroup and relation columns it matches option/user/relation IDs, not the visible labels.
• Full-text search no longer highlights the matching cell in the row, since the backend does not return which cell matched.
• Backend sort is not type-cast: number and number subtypes sort lexicographically (so e.g. 9 > 100), and datetime values are sorted as raw strings.
• selection / selection-multi sort by raw option ID, not by option label.
• usergroup and relation columns sort by raw user/row IDs, not by display labels, and selection-check / text-long / text-rich are not guarded from sorting.
• Filter contains / does-not-contain on text-long / text-rich searches raw stored values (possibly with HTML for rich text) rather than visible text.
• Search and filter result cell highlighting is not wired in the new backend flow: TableRow.vue currently has 'search-result': false and 'filter-result': false with no backend signal indicating which cell matched.

What to test

• Tables and views with more rows than the page size: page size, page numbers and the row count.
• Filters of all types, especially selection "contains items".
• Search on text and number columns.
• Pagination reset when adding, changing or clearing filters and search.
• Loading spinner on the reload button and the empty state.
• Opening a table or view with a rowIds query parameter and confirming only the selected rows are loaded.
• The "Share selected rows" action: copied URL, temporary "Open" button, and reset clearing the selection.
• CSV export for selected rows and filtered rows.

🗒️ TODO

• Add Behat tests
• Add Cypress tests

🏁 Checklist

• ⛑️ Tests (unit and/or integration) are included or not needed
• 🔙 Backport requests are created or not needed: /backport to stableX.X
• 📅 Milestone is set
• 🌸 PR title is meaningful (if it should be in the changelog: is it meaningful to users?)

🤖 AI (if applicable)

• The content of this PR was partly or fully generated using AI

[Koc]

do we need to move that into separate PR?