Standardize and expand table quality/confidence metrics

[Mohataseem89]

Camelot exposes some internal table parsing metrics, but their meaning and consistency across extraction flavors (lattice and stream) are unclear and not well documented. This makes it difficult for users to reliably assess extraction quality or automate table validation in production workflows

Proposal: Standardize table quality metrics across flavors Clearly document metric meanings and expected ranges Optionally expose a unified or composite confidence score per table

Why this helps: Clear and consistent quality indicators would improve trust in extracted tables enable automated filtering, and make Camelot easier to use in real-world data pipelines