pdf-table2json

Extract tables data from pdf files To JSON

• Locate the table with oepncv and read the contents with a text reader (Your table should be blocked by a border)
• (If you don't have a border, add a border through adjustment)
• The pdf must be readable by a text reader. Drag on pdf to see if the text is captured.
• Check version before install
  • 1.0.1 : Only the basic table is supported
  • 2.0.1 : Handling tables with separate headers or cells (example)

Current Status(Change the values ​​if adjustments are needed)

• Finds a table with a horizontal length greater than 1000 and a height greater than 100.
• Cells are excluded if their width or height is equal to the width or height of the table, or if the width or height of the cell is less than 10.
• Adding border lines to areas with a color of (230, 230, 230) and a width of 1000 or more to recognize them as table regions.
• Removing watermark images with a color of (213, 213, 213 == #D5D5D5).
• The specific string list is removed from the PDF text for the purpose of removing text watermarks (Currently empty).

Installation

• Rquired Python >= 3.8
• install with pip

pip install pdf-table2json 

Example

import

import pdf_table2json.converter as converter

path = "PATH/PDF_NAME.pdf"
result = converter.main(path, json_file_out=True, image_file_out=True)
print(result)

CLI

python converter.py -i "pdf_path/pdf_name.pdf" [-j] [-o]

• "-i", "--input", required=True, help="[Required] Input PDF file path"
• "-j", "--json_file", action="store_true", help="[Optional] Create JSON Data file"
• "-o", "--image_file", action="store_true", help="[Optional] Save Image Data file"

version-1.0.1

• Only the basic table is supported. (Supports only tables with horizontal headers).

• The number of headers and the number of cells must be the same

• Example Table

  Header 1	Header 2	Header 3
  cel1	cel2	cel3
  cel1	cel2	cel3
  cel1	cel2	cel3

• Example

  • converter.py

  import pdf_table2json.converter as converter
  
  path = "PATH/PDF_NAME.pdf"
  result = converter.main(path, json_file_out=True, image_file_out=True)
  print(result)
  

version-2.0.1 or Higher

1. Tables in general format that can be processed in version 1.0.1 can be processed.

   • Example Table

     Header 1	Header 2	Header 3
     cel1	cel2	cel3
     cel1	cel2	cel3
     cel1	cel2	cel3

2. Table with separated header and subheader

   • Example Table

     Header 1	Header 2
     Sub Header 1	Sub Header 2
     cel1	cel2	cel3
     cel1	cel2	cel3
     cel1	cel2	cel3

   • Output

     • Delete separated parent header, use child header

       Header 1 : cel1
       Sub Header 1 : cel2
       Sub Header 2 : cel3
       

3. Tables with columns separated, except for the first cell

   • Example Table

     Header 1	Header 2	Header 3
     cel1	cel2	cel3
     cel1	cel2-1	cel3-1
     	cel2-2	cel3-2

   • Output

     • Add to data in the top row (with "@")

       Header 1 : cel1
       Header 2 : cel2
       Header 3 : cel3
       Header 1 : cel1
       Header 2 : cel2-1@cel2-2
       Header 3 : cel3-1@cel3-2
       

• Use Example

  • converter_2.py

  import pdf_table2json.converter_2 as converter_2
  
  path = "PATH/PDF_NAME.pdf"
  result = converter_2.main(path, json_file_out=True, image_file_out=True)
  print(result)
  

License

• GPL-3.0 license

Contact

• Reporting a bug
• @yousojeong

Read Text From PDF library

• PyMuPDF GitHub