Break New-FinOpsTestData into smaller composable commands

Summary

Refactor New-FinOpsTestData into smaller, composable commands — one per dataset type — to improve maintainability and allow independent use.

Motivation

New-FinOpsTestData is a large (~2,400+ line) monolithic function that generates all 5 FOCUS datasets. Breaking it into smaller commands would:

• Make each dataset generator independently testable and usable
• Reduce complexity per function
• Allow users to generate only the datasets they need
• Follow PowerShell best practices for single-responsibility commands

Proposed commands

• New-FinOpsTestCostData — generates Costs dataset rows
• New-FinOpsTestPriceData — generates Prices dataset rows
• New-FinOpsTestCommitmentData — generates CommitmentDiscountUsage dataset rows
• New-FinOpsTestRecommendationData — generates Recommendations dataset rows
• New-FinOpsTestTransactionData — generates Transactions dataset rows
• New-FinOpsTestData — orchestrator that calls the above and handles file output, storage upload, and manifest generation

Considerations

• Shared helper logic (provider templates, column definitions, FOCUS version filtering) should be extracted into internal/private functions
• Each sub-command should accept common parameters (-Providers, -FocusVersion, -Months, -Seed, etc.)
• Maintain backward compatibility — New-FinOpsTestData should continue to work as-is

Context

Deferred from PR #2006 (New-FinOpsTestData and Remove-FinOpsTestData).

Start by locating the New-FinOpsTestData command and reviewing the deferred work from PR #2006. Trace how it generates the five FOCUS datasets and handles file output, storage upload, and manifest generation. Done means the five dataset commands are independently usable while New-FinOpsTestData preserves backward-compatible behavior.