vertical_lines & add_lines & add_boxes parameters for find_tables option.

[1504168]

allg_DP_Zwickau (1).pdf

I have couple of questions on those parameters for above attachment.

# Q1 : Why is text being split into multiple columns specially Arzt and Col4
# Q2:  Does join_x_tolerance have any effect on this?

import pymupdf
doc = pymupdf.open('allg_DP_Zwickau (1).pdf')
page = doc[2]
tabs = page.find_tables(strategy='text',clip=(0,page.search_for('Arzt')[0][1],700,page.rect[3]),snap_y_tolerance=7,join_x_tolerance=10)
tabs.tables[0].to_pandas()

Output:

# Q1 : Why Col5 is present when I am explicitly specifying the vertical lines?
       # is first column will be between 42.52 and 63.92? and second column between 63.92 to 106.44?

doc = pymupdf.open('allg_DP_Zwickau (1).pdf')
page = doc[3]
vertical_lines = [42.52,63.92,106.44,148.96,367.22,460]
print(page.get_text('words',clip=(0,page.search_for('Arzt')[0][1],700,page.search_for('Arzt')[0][1]+30)))
tabs = page.find_tables(horizontal_strategy='text',clip=(0,page.search_for('Arzt')[0][1],700,page.rect[3]),vertical_lines = vertical_lines ,snap_y_tolerance=7)
tabs.tables[0].to_pandas()

Output:

# Q1 : Why Col5 is present when I am explicitly specifying the lines?
# Q2 : Can we use the add_boxes param in this PDF example?

doc = pymupdf.open('allg_DP_Zwickau (1).pdf')
page = doc[3]

top = page.search_for('Arzt')[0][1]
y_b= page.rect.y1

print(page.get_text('words',clip=(0,page.search_for('Arzt')[0][1],700,page.search_for('Arzt')[0][1]+30)))

lines = [
    ((42.52,top),(42.52,y_b)),
    ((63.92,top),(63.92,y_b)),
    ((106.44,top),(106.44,y_b)),
    ((148.96,top),(148.96,y_b)),
    ((367.22,top),(367.22,y_b)),
    ((460,top),(460,y_b)),
]

tabs = page.find_tables(horizontal_strategy='text',clip=(0,page.search_for('Arzt')[0][1],700,page.rect[3]),add_lines = lines,snap_y_tolerance=7)
tabs.tables[0].to_pandas()

Output:

I will attach the Jupyter notebook and PDF file as a zip file here.
Invoice Data Extraction.zip