Handle non-finite values in professional chart metadata

[agmangas]

Summary

• The risk assessment professional output could fail when indicator arrays contained non-finite values that reached histogram generation.
• This change filters indicator arrays to finite values before computing histogram bins and summary statistics, and returns an explicit empty-distribution payload when no finite values are available.

Changes

• src/relife_financial/services/risk_assessment.py: Updated professional chart_metadata generation to use np.isfinite(...) filtering and avoid histogram/percentile operations on invalid arrays.
• src/relife_financial/services/risk_assessment.py: Added fallback chart metadata for no-finite-data cases (empty bins, numeric default statistics, and informational note).

[Copilot]

Pull request overview

This PR hardens the professional risk assessment response generation by preventing non-finite simulation values (NaN/Inf) from breaking chart distribution calculations, and by returning an explicit empty-distribution payload when no finite values exist.

Changes:

• Filters indicator arrays to finite values before computing histogram bins and summary statistics in professional chart_metadata.
• Adds an explicit fallback chart_metadata payload (empty bins + defaulted stats + note) when an indicator has no finite simulation values.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.