A 43-page borderless call log (with animals)

This PDF is a service call report covering 911 incidents at the Rainforest Cafe in Niagara Falls, NY — we’re hunting for the animal calls. The data is a spreadsheet printed to PDF: wildly varied column widths, no cell borders anywhere, whole columns that sit empty for pages at a time, and a stray “2770 Records Found” line waiting to land in the middle of your data.

from natural_pdf import PDF

pdf = PDF("pdfs/24480polcompleted.pdf")
pdf.show(cols=3, limit=9)

Select the spreadsheet pages

The first four pages are cover letters. The table starts on page 5:

pages = pdf.pages[4:]
pages.show(cols=6)

Exclude the record count

The last page ends with “2770 Records Found” — a line that would otherwise be swallowed into the table as a bogus row:

pages[-1].show(crop=200)

Instead of matching the exact string, match the shape of it with a regex — then the same code works next year when the count is different:

(
  pages[-1]
  .find_all(r'text:regex(\d+ Records Found)')
  .show(crop=100)
)

Anything you can find, you can exclude. Two options, depending on how much you trust the document:

Exclude on the last page

The count is always on the last page, so a selector exclusion on that one page is enough:

pages[-1].add_exclusion(r'text:regex(\d+ Records Found)')

<Page number=47 index=46>

Exclude on all possible pages

If you’re not sure the count only appears at the end — or you’ll be feeding in other documents like this one — register it on the PDF, and every page filters it:

pdf.add_exclusion(
  lambda page: page.find_all(r'text:regex(\d+ Records Found)')
)

<PDF source='pdfs/24480polcompleted.pdf' pages=47>

Record counts: excluded.

pages[-1].show(exclusions='red')

Look at the table structure

pages[0].show()

No borders, so we’ll draw our own with guides:

• Vertical boundaries at the start of each column header, reused on every page.
• Horizontal boundaries wherever text starts with NF- — every row begins with a case number like NF-00051026-24, so multi-line rows can’t split into two records.

Why not just extract_table() per page and concatenate? Because if an entire column is empty on some page, plain extraction has no way to know a column is missing there — a grid built from the headers does.

Reusable guides

Vertical guides come from the header text. The last column (“Main Officer”) has no boundary to its right, so outer="last" counts the space after the final guide as a column:

from natural_pdf.guides import Guides

guide = Guides(pages[0])
columns = ['Number', 'Date Occurred', 'Time Occurred', 'Location', 'Call Type', 'Description', 'Disposition', 'Main Officer']
guide.vertical.from_content(columns, outer="last")
guide.horizontal.from_content(
  lambda p: p.find_all('text:starts-with(NF-)')
)
guide.show()

The lambda matters: it means “find the NF- rows on whatever page this guide is applied to,” not just the first page.

Now apply the guide to every page in one call. The headers only exist on page one, so header="first":

table_result = guide.extract_table(pages, header="first")
df = table_result.to_df()
df.head()

	Number	Date Occurred	Time Occurred	Location	Call Type	Description	Disposition	Main Officer
0	NF-00051317-24	11/17/2024	708	300 3RD ST, NIAGARA FALLS, NEW YORK	STOLEN VEHICLE	VEH WAS PARKED IN CITY LOT. NO LONGER THERE. >	01 -CRIME REPORT	BAUMANN, JACK
1	NF-00051302-24	11/17/2024	208	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	None	37-CHECKED - ALL OK	ALCIME, ADONIS S
2	NF-00051298-24	11/17/2024	153	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	CLAPS, BRANDAN
3	NF-00051026-24	11/15/2024	141	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	GIZZARELLI, RYAN
4	NF-00051010-24	11/14/2024	2357	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	Sheraton Hotel	37-CHECKED - ALL OK	SHIHADAH, BAHA A

Guides with loops

The same thing, spelled out. Build the base grid from the first page:

from natural_pdf.guides import Guides

base = Guides(pages[0])
columns = ['Number', 'Date Occurred', 'Time Occurred', 'Location', 'Call Type', 'Description', 'Disposition', 'Main Officer']
base.vertical.from_content(columns, outer="last")
base.horizontal.from_content(pages[0].find_all('text:starts-with(NF-)'))
base.show()

Extract the first page — the only one with headers:

first_table = base.extract_table().to_df()
first_table.head()

	Number	Date Occurred	Time Occurred	Location	Call Type	Description	Disposition	Main Officer
0	NF-00051317-24	11/17/2024	708	300 3RD ST, NIAGARA FALLS, NEW YORK	STOLEN VEHICLE	VEH WAS PARKED IN CITY LOT. NO LONGER THERE. >	01 -CRIME REPORT	BAUMANN, JACK
1	NF-00051302-24	11/17/2024	208	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	None	37-CHECKED - ALL OK	ALCIME, ADONIS S
2	NF-00051298-24	11/17/2024	153	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	CLAPS, BRANDAN
3	NF-00051026-24	11/15/2024	141	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	GIZZARELLI, RYAN
4	NF-00051010-24	11/14/2024	2357	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	Sheraton Hotel	37-CHECKED - ALL OK	SHIHADAH, BAHA A

Then walk the remaining pages, reusing the column positions but recomputing the rows from each page’s own NF- anchors. to_df(header=columns) sets the headers manually so the frames can stack:

dataframes = [first_table]

for page in pages[1:]:
    guides = Guides(page)
    guides.vertical = base.vertical
    guides.horizontal.from_content(page.find_all('text:starts-with(NF-)'))
    single_df = guides.extract_table().to_df(header=columns)
    dataframes.append(single_df)
print("We made", len(dataframes), "dataframes")

We made 43 dataframes

import pandas as pd

df = pd.concat(dataframes, ignore_index=True)
df.head()

	Number	Date Occurred	Time Occurred	Location	Call Type	Description	Disposition	Main Officer
0	NF-00051317-24	11/17/2024	708	300 3RD ST, NIAGARA FALLS, NEW YORK	STOLEN VEHICLE	VEH WAS PARKED IN CITY LOT. NO LONGER THERE. >	01 -CRIME REPORT	BAUMANN, JACK
1	NF-00051302-24	11/17/2024	208	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	None	37-CHECKED - ALL OK	ALCIME, ADONIS S
2	NF-00051298-24	11/17/2024	153	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	CLAPS, BRANDAN
3	NF-00051026-24	11/15/2024	141	300 3RD ST, NIAGARA FALLS,	PREMISE CHECK	None	37-CHECKED - ALL OK	GIZZARELLI, RYAN
4	NF-00051010-24	11/14/2024	2357	300 3RD ST, NIAGARA FALLS, NEW YORK	PREMISE CHECK	Sheraton Hotel	37-CHECKED - ALL OK	SHIHADAH, BAHA A

From here the animal hunt is a pandas filter on Call Type — the PDF part is done.