Incorrect table extraction despite visible grid lines (nested boxes)

[sjainam]

Describe the bug
pdfplumber fails to detect table grid lines when using table_settings with vertical_strategy="lines" and horizontal_strategy="lines", even though the PDF visually contains clear table borders and grid lines. As a result, tables are not detected or extracted as expected.

Have you tried repairing the PDF?

• Yes. The issue persists even when opening the PDF using:
• pdfplumber.open("sample.pdf", repair=True)

Code to reproduce the problem

im = pdfplumber.open("/home/jainamshah/Downloads/111.pdf").pages[0].to_image()
im.debug_tablefinder(table_settings= {"vertical_strategy": "lines","horizontal_strategy": "lines"})

PDF file

• Attached PDF: sample.pdf
• The PDF contains a clearly visible table with grid lines (both vertical and horizontal).
• No selectable text is present (vector-based drawing).

Expected behavior
Since the table grid is clearly visible, pdfplumber should:

• Detect vertical and horizontal ruling lines
• Identify consistent intersections
• Correctly infer table structure
• Extract rows and columns accurately using line-based strategies
• No cell should overlap, merge into, or appear inside another cell

Actual behavior

• Table grid lines are detected only partially or as fragmented segments
• Intersections are inconsistent
• extract_tables() fails or returns malformed tables
• Line-based strategies do not produce reliable results
• also into One box another box is wrog

Screenshots

Environment

• pdfplumber version: 0.11.9
• Python version: 3.12.3
• OS: Linux

Additional context
I have already raised Same type of Issue :
#175

[monchin]

Seems there are many white (invisible) lines in this pdf. What if you extract all edges with TableFinder(page, {"vertical_strategy": "lines_strict", "horizontal_strategy": "lines_strict"}).edges and filter out all white edges (math.isclose(stroking_color, 1.0)) and use the rest edges as explicit strategy? If that works, it could be concluded that this issue is just raised by white color, and maybe a color filter setting (a boolean exclude_white_edges or some arg to filter) could be added to solve it.

[monchin]

First this pdf, "lines_strict" strategy would be a better choice, because there are some invisible rects in the table which would disrupt the judge of table extraction.
monchin/tablers#20 has added "exclude_white_edges" and you can try tablers==0.5.0 to see whether it is good for you. @sjainam

For pdfplumber, if @jsvine thinks "exclude_white_edges" is worth being added, I'd like to make a PR. However, "exclude_white_edges" should only be valid on "lines_strict" lines.