Full data type treeviewitems

[James Robinson (jlrobins)]

Summary of Changes

• New model object FlinkTypeNode class to bridge the Flink full datatype parser ouput interface FlinkType objects introduced here to become fully fledged children of the Flink Database view.
  • ROW, MAP column types are recusively expandable (having children) now.
  • ARRAY and MULTISET types are special cased to not appear like they have children (they both have a single child node, the type being array'd or multiseted over) unless is directly or indirectly a array/multiset of ROW or MAP, in which case we can expand to see the ROW or MAP children. Rather, we annotate the member type as being '[]' or ' MULTISET', more closely matching how regular programming languages model these things. So:
    • An ARRAY<INT> will not be expandable, but instead will look like 'INT []'.
    • An ARRAY<ROW<id INT, name VARCHAR>> will be expandable, showing two children id INT and name VARCHAR, and the description for the ARRAY node itself will read 'ROW []'.
    • An ARRAY<ARRAY<ROW<id INT, name VARCHAR>>> will be expandable -- the determination that there's a ROW or MAP is done recursively.
• Adjust the Flink full datatype parser and output interfaces to also keep track of the original full data type string (segment) so that we can display it in the tooltip of a FlinkTypeNode.
• FlinkTypeNode objects create their own TreeViewItems just like FlinkRelation does.
• So many new tests.

Click-testing instructions

1. Log in to CCloud, navigate to realworld-data cluster in the Flink Database view.
2. Expand tables and views.
3. Check out how table spotify-listening-data columns show up. The ROW columns are now expandable, all the way down.

Caveats

• There's really only ~575 added or changed lines of actual code here --- FlinkTypeNode, hooking up a getChildren() to FlinkRelationColumn, autmenting the type parser to keep track of the substring range for this single datatype. All of the rest of the changes are tests, tests , tests.
• We now have support for showing showing different icons for different Flink types, but the only suitable icon to my eyes at this point is one for arrays, looks nice like '[]'. Respelling what some of the new constants in our icons.ts file refer to (and / or creating new icons) is for the future.

Pull request checklist

Please check if your PR fulfills the following (if applicable):

Tests

• Added new
• Updated existing
• Deleted existing

Release notes

• Does anything in this PR need to be mentioned in the user-facing CHANGELOG?

[Copilot]

Pull request overview

Adds first-class TreeView support for fully-expanded Flink SQL data types in the Flink Database view by introducing a FlinkTypeNode model, extending the full datatype parser output with an exact fullDataTypeString segment for tooltips, and wiring expand/collapse + icons through the existing view provider patterns.

Changes:

• Introduce FlinkTypeNode as a tree-aware wrapper over parsed FlinkType, including tooltips, icons, IDs, and child expansion.
• Extend the Flink datatype parser/model to retain fullDataTypeString per parsed node, and add a shared display formatter (formatFlinkTypeForDisplay).
• Update Flink Database view + relation column models to surface nested type children and add extensive unit test coverage.

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
src/viewProviders/flinkDatabase.ts	Integrates FlinkTypeNode into Flink Database tree expansion/parenting and reveal behavior.
src/viewProviders/flinkDatabase.test.ts	Adds unit tests for expanding/revealing FlinkTypeNode and column type children.
src/utils/flinkTypes.ts	Adds unified Flink type display formatting and updates SQL type cleaning.
src/utils/flinkTypes.test.ts	Adds tests for display formatting across scalar/compound/container types.
src/parsers/flinkTypeParser.ts	Captures fullDataTypeString for parsed nodes (for tooltip fidelity).
src/parsers/flinkTypeParser.test.ts	Updates/extends parser tests to validate fullDataTypeString behavior.
src/models/flinkTypes.ts	Extends Flink type model with fullDataTypeString.
src/models/flinkTypeNode.ts	New model bridging parsed types into TreeView-friendly nodes (children/tooltips/icons/search).
src/models/flinkTypeNode.test.ts	Comprehensive unit tests for FlinkTypeNode behavior and rendering metadata.
src/models/flinkRelation.ts	Updates column rendering/expandability via parser + FlinkTypeNode, adds icons + caching/error logging.
src/models/flinkRelation.test.ts	Updates tests for new column formatting, tooltip behavior, and type-children expansion.
src/icons.ts	Adds icon constants for ROW/ARRAY/MULTISET.
FULL_DATATYPE_PARSER_PROJECT.md	Project notes documenting goals/approach for the parser + UI integration.

[Copilot]

Pull request overview

Copilot reviewed 13 out of 13 changed files in this pull request and generated 8 comments.

[Copilot]

Pull request overview

Copilot reviewed 13 out of 13 changed files in this pull request and generated 5 comments.

[sonarqube-confluent]

Quality Gate passed

Issues
0 New issues
9 Fixed issues
0 Accepted issues

Measures
0 Security Hotspots
95.6% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube

[Dave Shoup (shouples)]

Love this! 🙌 Only concerns are about the icon enum values and usage, but we'll need to do some icon-related follow-up before the 2.3 release anyway.

[Dave Shoup (shouples)]

Minor: this currently makes it look like all other types are the same as UDFs:

[James Robinson (jlrobins)]

hah -- I was thinking 'f' for 'field', forgot that over there it was 'f' for 'function.' Sigh.

That said, a followup item for this parser is to have each UDF param or return type also be a child node of the UDF, then possibly have them be expandable if is compound like here. So:

• MyFunctionB:
  • Parameters:
    • a INT
    • B INT
  • Return type:
    • INT

But if it returned, say, a ROW type:

* MyFunctionB:
    * Parameters:
        * a INT
        * B INT
    * Return type:
        * ROW:
            * name: VARCHAR 
            * age: INT
            * ...