Render DataFrames as interactive notebook tables

[eosfor]

Summary

• add a packaged Verso.DataFrame extension for Microsoft.Data.Analysis.DataFrame values
• render DataFrames as theme-aware, scrollable HTML tables with column names and types, sticky headers, null styling, HTML encoding, and a bounded 100-row preview
• recognize DataFrames and unwrap PowerShell PSObject values through reflection
• integrate the formatter into the CLI, Blazor editor, and VS Code host
• add unit tests plus built-in extension discovery coverage
• document DataFrame table support and its place in the execution pipeline

Why

PowerShell data-analysis modules can expose Microsoft.Data.Analysis.DataFrame, but the default output only reveals the object's Columns and Rows properties. A notebook needs a readable table while preserving extension isolation from the language kernel that owns the runtime assembly.

The extension deliberately has no compile-time dependency on Microsoft.Data.Analysis. It recognizes and reads the public DataFrame shape through reflection, avoiding type-identity conflicts when PowerShell loads a different version or load context.

User impact

Both implicit PowerShell output and explicit display now produce the same table:

Import-Module DataFrame
$penguins = Import-DataFrame ./data/penguins.csv

$penguins
Display $penguins

Manual validation used the Palmer Penguins dataset: 344 rows and 8 columns, with the preview correctly reporting Showing 100 of 344 rows.

Validation

• dotnet test Verso.sln --no-restore
• Verso.DataFrame.Tests: 11 passed
• CLI discovery reports verso.dataframe.formatter
• package inspection confirms that Verso.DataFrame has no Microsoft.Data.Analysis dependency
• manually verified implicit output and Display $penguins using the PowerShell DataFrame module

The full solution test suite passes. Existing dependency-analysis and documentation warnings are unchanged.