I see the issues. I need to:

1. Fix the check filtering - it's returning null for checks that don't exist in API response, causing them not to render
2. Fix the tooltip z-index issue so it doesn't go behind the table
3. Show full document name

Let me fix these issues:

I've made the following fixes:

1. Fixed extraction checks not populating - Changed the logic to always iterate through config-defined checks (file_type_check, empty_page_check, all_empty_check) even if the API response doesn't contain them. If a check doesn't exist in the API response, it will show as "Skipped" with the config-based insight and recommendation.
2. Fixed tooltip going behind the table - Changed the tooltip to appear above the check item (bottom-full instead of top-full) and increased the z-index to z-[200]. Also added hover:z-[100] on the parent to ensure proper stacking.
3. Showing full document name - Removed the truncation and now displays the full document name with break-all for long file names. Also adjusted the grid to md:grid-cols-4 to give more space to the document name field.

I see the issue - the parseChecks function simply replaces all single quotes with double quotes. But the problem is when there's text like "File type 'pdf' is valid" - the inner single quotes around pdf will also be converted to double quotes, breaking the JSON parsing.

Let me fix this by using a more robust parsing approach:

The issue was that the simple replace(/'/g, '"') was breaking JSON parsing when there were single quotes inside string values like "File type 'pdf' is valid". The new parsing function handles Python-style dict strings more carefully by:

1. First preserving escaped quotes with a placeholder
2. Converting single-quoted keys to double-quoted keys (e.g., 'status': to "status":)
3. Converting single-quoted values to double-quoted values (e.g., : 'Pass' to : "Pass")
4. Restoring any preserved quotes
5. Adding debug logging to help diagnose any future parsing issues