0.2.2 -- Starting with an analysis goal, or get suggestions from the agent.

[Chenglong-MS]

You can start your exploration with a goal, or, tab and see if the agent can recommend some good exploration ideas for you. Oh, I find Claude models works best on suggesting good analysis objectives at the moment.

tab-and-see.mov

[Copilot]

Pull Request Overview

This PR enhances the data exploration workflow by introducing a goal-driven chart recommendation panel, incorporating automated “explorative questions” into table metadata, and cleaning up legacy sync logic and model-testing patterns.

• Add a new ChartRecBox component for NL-driven chart suggestions and integrate it into the visualization view
• Extend DictTable and the slice to carry and populate explorativeQuestions from the agent
• Remove outdated syncTableWithServer calls, consolidate model-slot readiness logic, and sanitize table names across data loaders

Reviewed Changes

Copilot reviewed 19 out of 19 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/views/VisualizationView.tsx	Integrate ChartRecBox into chart selection when a table is focused
src/views/TableSelectionView.tsx	Remove deprecated syncTableWithServer calls
src/views/ModelSelectionDialog.tsx	Refactor model-slot status display and button disabling logic
src/views/EncodingShelfCard.tsx	Adjust styling and placeholder logic for “Auto” chart types
src/views/DataThread.tsx	Refine thread rendering to show thread origin and trigger cards
src/views/DataFormulator.tsx	Add global model testing; gate main view on model-slot readiness
src/views/DBTableManager.tsx	Initialize explorativeQuestions on load; tweak error/query buttons
src/views/ChartRecBox.tsx	Introduce new component for exploration goals and suggestions
src/components/ComponentType.tsx	Add explorativeQuestions field and update createDictTable signature
src/app/dfSlice.tsx	Populate explorativeQuestions; tighten createNewChart payload
py-src/data_formulator/data_loader/*.py	Sanitize table names in all ingest_data_from_query methods
py-src/data_formulator/agents/agent_data_load.py	Extend system prompt to request explorative questions
py-src/data_formulator/agents/agent_query_completion.py	Remove redundant “[REASONING]” markers in query completion output

Comments suppressed due to low confidence (2)

src/views/ChartRecBox.tsx:1

• [nitpick] The new ChartRecBox component contains substantial logic and UI interactions but lacks corresponding unit or integration tests. Consider adding tests to cover its key behaviors (e.g., table selection, prompt handling, API calls).

// Copyright (c) Microsoft Corporation.

py-src/data_formulator/data_loader/s3_data_loader.py:206

• sanitize_table_name is used here but not imported in this module. Add a proper import for sanitize_table_name or fully qualify its reference to avoid a NameError at runtime.

        self.ingest_df_to_duckdb(df, sanitize_table_name(name_as))