image-table-ocr/README.org

#+TITLE: Table detection in images and OCR to CSV

* Overview

This python package contains modules to help with finding and extracting tabular
data from a PDF or image into a CSV format.

Given an image that contains a table...

#+ATTR_HTML: :width 25%
[[file:resources/examples/example-page.png]]

Extract the the text into a CSV format...

#+BEGIN_EXAMPLE
PRIZE,ODDS 1 IN:,# OF WINNERS*
$3,9.09,"282,447"
$5,16.66,"154,097"
$7,40.01,"64,169"
$10,26.67,"96,283"
$20,100.00,"25,677"
$30,290.83,"8,829"
$50,239.66,"10,714"
$100,919.66,"2,792"
$500,"6,652.07",386
"$40,000","855,899.99",3
1,i223,
Toa,,
,,
,,"* Based upon 2,567,700"
#+END_EXAMPLE

* Requirements

Along with the python requirements that are listed in setup.py and that are automatically installed when installing this package through pip, there are a few external requirements for some of the modules.

I haven't looked into the minimum required versions of these dependencies, but I'll list the versions that I'm using.

- ~pdfimages~ 20.09.0 of [[https://poppler.freedesktop.org/][Poppler]]
- ~tesseract~ 5.0.0 of [[https://github.com/tesseract-ocr/tesseract][Tesseract]]
- ~mogrify~ 7.0.10 of [[https://imagemagick.org/index.php][ImageMagick]]

* Demo

There is a demo module that will download an image given a URL and try to extract tables from the image and process the cells into a CSV. You can try it out with one of the images included in this repo.

1. ~pip3 install table_ocr~
2. ~python3 -m table_ocr.demo https://raw.githubusercontent.com/eihli/image-table-ocr/master/resources/test_data/simple.png~

That will run against the following image:

#+ATTR_HTML: :width 40%
[[file:resources/test_data/simple.png]]

The following should be printed to your terminal after running the above commands.

#+BEGIN_EXAMPLE
Running `extract_tables.main([/tmp/demo_p9on6m8o/simple.png]).`
Extracted the following tables from the image:
[('/tmp/demo_p9on6m8o/simple.png', ['/tmp/demo_p9on6m8o/simple/table-000.png'])]
Processing tables for /tmp/demo_p9on6m8o/simple.png.
Processing table /tmp/demo_p9on6m8o/simple/table-000.png.
Extracted 18 cells from /tmp/demo_p9on6m8o/simple/table-000.png
Cells:
/tmp/demo_p9on6m8o/simple/cells/000-000.png: Cell
/tmp/demo_p9on6m8o/simple/cells/000-001.png: Format
/tmp/demo_p9on6m8o/simple/cells/000-002.png: Formula
...

Here is the entire CSV output:

Cell,Format,Formula
B4,Percentage,None
C4,General,None
D4,Accounting,None
E4,Currency,"=PMT(B4/12,C4,D4)"
F4,Currency,=E4*C4
#+END_EXAMPLE

* Modules

The package is split into modules with narrow focuses.

- ~pdf_to_images~ uses Poppler and ImageMagick to extract images from a PDF.
- ~extract_tables~ finds and extracts table-looking things from an image.
- ~extract_cells~ extracts and orders cells from a table.
- ~ocr_image~ uses Tesseract to OCR the text from an image of a cell.
- ~ocr_to_csv~ converts into a CSV the directory structure that ~ocr_image~ outputs.

The outputs of a previous module can be used by a subsequent module so that they
can be chained together to create the entire workflow, as demonstrated by the
following shell script.

#+NAME: ocr_tables
#+BEGIN_SRC shell :results none :tangle ocr_tables :tangle-mode (identity #o755)
#!/bin/sh

PDF=$1

python -m table_ocr.pdf_to_images $PDF | grep .png > /tmp/pdf-images.txt
cat /tmp/pdf-images.txt | xargs -I{} python -m table_ocr.extract_tables {}  | grep table > /tmp/extracted-tables.txt
cat /tmp/extracted-tables.txt | xargs -I{} python -m table_ocr.extract_cells {} | grep cells > /tmp/extracted-cells.txt
cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {}

for image in $(cat /tmp/extracted-tables.txt); do
    dir=$(dirname $image)
    python -m table_ocr.ocr_to_csv $(find $dir/cells -name "*.txt")
done
#+END_SRC


The package was written in a [[https://en.wikipedia.org/wiki/Literate_programming][literate programming]] style. The source code at
[[https://eihli.github.io/image-table-ocr/pdf_table_extraction_and_ocr.html]] is
meant to act as the documentation and reference material.