Bash script to convert tables inserted in PDFs into CSVs in batch mode using the python module table-ocr

batch_ocr_to_table

#!/bin/sh

# Convert each PDF to separate images per page 
for pdf in ./*.pdf;do
    python3 -m table_ocr.pdf_to_images $pdf;
done;

# Extract tables from each page in .png output
for images in $(find . -name "*.png"); do
      python3 -m table_ocr.extract_tables $images;
done;

# Extract cells from each table in .png outputs
for tables in $(find . -name "table-*"); do
     python3 -m table_ocr.extract_cells $tables;
done; 

# Apply OCR to each image of each cell, output as .txt
for cells in $(find . -name "0*-*.png"); do
      python3 -m table_ocr.ocr_image $cells;
done;

# Build CSVs with the different .txt files of each PDF analysed.
folders=();
lenght=0;
for folder in "$(find -type d -iname "*-*")";do
    folders+=($folder);
done;

for i in "${folders[@]}"; do
    python3 -m table_ocr.ocr_to_csv $(find $i/cells/ -name "*.gt.txt") > "$lenght".csv;
    ((++lenght));
done

Hi @eihli I wrote this bash script using table-ocr. Basically it is used to export to individual csv files the tables that it extracts from the pdfs in batch mode. I would like to improve the nomenclature of the output csv files to know which pdf they correspond to. I hope to be able to integrate it soon. Thanks for such a great tool!