Feat/Fix : Dataframe element fix and improvements

[Allaoua9]

This PR was implemented in collaboration with @tonca, many thanks to him !

• Fixed issue with horizontal overflow in the pagination footer
• Improved the Dataframe element by adding filters, sorting, column selection and ability to select the number of lines to display
• New features are optional and disabled by default

---

Summary by cubic

Adds per-column filters, a column visibility menu, a page-size selector, expandable row details, and a clear-filters action to the Dataframe element for easier exploration. Also fixes horizontal overflow in the pagination footer and guards against None data on the backend.

• New Features

  • Per-column filters and sorting in headers.
  • Column visibility dropdown with select/deselect all; controllable from Python via Dataframe(show_column_visibility=..., show_column_filters=...) and exposed as showColumnVisibility/showColumnFilters in @chainlit/react-client.
  • Page size selector (5/10/20/50) and a compact rows count indicator.
  • Expandable rows with a full details view on click.
  • Clear filters button and a localized “no results” state, with new i18n strings for the Dataframe UI in all locales.

• Bug Fixes

  • Fixed horizontal overflow in the pagination footer.
  • Backend: handle data=None in Dataframe.__post_init__ to avoid errors when no data is provided.

^{Written for commit dcd54e6. Summary will update on new commits.}

[cubic-dev-ai]

5 issues found across 25 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="frontend/src/components/Elements/Dataframe.tsx">

<violation number="1" location="frontend/src/components/Elements/Dataframe.tsx:78">
P2: Column filter always stores text input, but numeric columns use different filter semantics in TanStack, causing incorrect filtering for number fields.</violation>

<violation number="2" location="frontend/src/components/Elements/Dataframe.tsx:324">
P2: Expanded row details ignore column visibility and reveal hidden column data.</violation>
</file>

<file name="backend/chainlit/translations/zh-TW.json">

<violation number="1" location="backend/chainlit/translations/zh-TW.json:264">
P3: `zh-TW` translation uses a Simplified character in `filterPlaceholder`, causing mixed-script UI text.</violation>
</file>

<file name="backend/chainlit/element.py">

<violation number="1" location="backend/chainlit/element.py:450">
P2: Dataframe now serializes new flags, but Element.from_dict() still lacks a dataframe case and will rebuild these elements as File, losing dataframe type/options when rehydrating from dicts.</violation>
</file>

<file name="backend/chainlit/translations/kn.json">

<violation number="1" location="backend/chainlit/translations/kn.json:261">
P3: The new Kannada translation for `deselectAll` contains a Telugu character (U+0C15) in the middle of the word, resulting in mixed-script, potentially unreadable UI text for kn users.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[cubic-dev-ai]

1 issue found across 4 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="backend/chainlit/element.py">

<violation number="1" location="backend/chainlit/element.py:185">
P1: `Element.from_dict()` constructs `Dataframe` without `data`, causing runtime `TypeError` during dataframe deserialization.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}