Issue with Bounding Boxes of Tables that span across multiple pages

### Expected Behaviour
When I retrieve a table from a PDF doc that spans across multiple pages, I am expecting the following behavior:
"Bounds": Are the bounds of the table on the page where the table starts.
"ClipBounds": Are the bounds of the table on the next page.

### Actual Behaviour
Unfortunately, the "Bounds" of the table part on the first page where the table starts are returned incorrectly by the API.
More specifically, the list of bounds returned seams to be correct except for the last element of the bound:
"Bounds": [correct, correct, correct, INCORRECT]
The incorrect value returned as the last element of the "Bounds" is equal to the last value of "ClipBounds" (which are the bounds of the continued table of the NEXT page). So "Bounds"[3] == "ClipBounds" [3] which is obviously not correct.

### Reproduce Scenario (including but not limited to)
Take a PDF spanning across two pages and print the "Bounds" returned onto the first page of the PDF.

#### Sample Code that illustrates the problem
Python Extract API.

Can someone please help to fix this?

Reproduce the reported case with the Python Extract API using a PDF table that spans two pages. Inspect the returned Bounds and ClipBounds values and compare them with the table geometry on each page. Done means the first page's Bounds[3] reflects that page, while the continued table's ClipBounds remains tied to the next page.

Written by the indexing model from the issue text.