This repository provides Python scripts for converting PDF documents into structured JSON files. It uses libraries like pdfplumber and pdfminer for text extraction and formatting.
- Extract text and titles from PDFs based on font size and formatting.
- Support for both
pdfplumberandpdfminerlibraries. - Handles custom page ranges for focused data extraction.
- Output JSON structure with clear title and content separation.
- Optional support for non-English text (e.g., Hindi).
Make sure you have the following libraries installed:
pdfplumberpdfminer.sixjson
Install them using:
pip install pdfplumber pdfminer.sixIdentify the largest font size (excluding digits) in a specified range of pages:
import pdfplumber
path = "example.pdf"
def is_digit(char):
return char.isdigit()
def extract_font_info(page):
max_font_size = 0
for element in page.chars:
current_char = element['text']
if is_digit(current_char):
continue
current_font_size = element['size']
if current_font_size > max_font_size:
max_font_size = current_font_size
return max_font_size
with pdfplumber.open(path) as pdf:
total_pages = len(pdf.pages)
start_page = int(total_pages * 0.5)
end_page = int(total_pages * 0.6)
largest_font_size = 0
for page_number in range(start_page, end_page):
page = pdf.pages[page_number]
current_page_max_font_size = extract_font_info(page)
if current_page_max_font_size > largest_font_size:
largest_font_size = current_page_max_font_size
print(f"Largest Font Size (excluding digits): {largest_font_size}")Once the largest font size is identified, extract text data into JSON format:
import pdfplumber
import json
def extract_text_from_pdf(path, largest_font_size):
temp_pdf = []
temp_dict = {"title": "", "data": ""}
with pdfplumber.open(path) as pdf:
for page in pdf.pages:
temp_title, page_data = extract_page_data(page.chars, largest_font_size)
if temp_title:
if len(temp_dict["data"]) > 0:
temp_pdf.append(temp_dict)
temp_dict = {"title": temp_title, "data": page_data}
else:
temp_dict["data"] += page_data
if len(temp_dict["data"]) > 0:
temp_pdf.append(temp_dict)
return temp_pdf
def extract_page_data(chars, largest_font_size):
temp_title = ""
page_data = ""
for span in chars:
if span['size'] >= largest_font_size:
temp_title += span['text']
else:
page_data += span['text']
return temp_title, page_data
def main():
path = "example.pdf"
largest_font_size = 24 # Replace with calculated value
temp_pdf = extract_text_from_pdf(path, largest_font_size)
with open(f"{path}_output.json", "w", encoding='utf-8') as out_file:
json.dump(temp_pdf, out_file, indent=6)
if __name__ == "__main__":
main()from pdfminer.high_level import extract_pages
from pdfminer.layout import LTTextContainer, LTChar
import json
def extract_text_with_pdfminer(path, largest_font_size):
temp_pdf = []
page_data = ""
temp_dict = {}
for page_layout in extract_pages(path):
for element in page_layout:
if isinstance(element, LTTextContainer):
for text_line in element:
for chr in text_line:
if isinstance(chr, LTChar):
if chr.size >= largest_font_size:
if len(page_data) > 0:
temp_dict['data'] = page_data
temp_pdf.append(temp_dict)
temp_dict = {}
temp_dict["title"] = text_line.get_text()
page_data = ""
else:
page_data += text_line.get_text()
with open(f"{path}_output.json", "w", encoding='utf-8') as out_file:
json.dump(temp_pdf, out_file, indent=6, ensure_ascii=False)
path = "example.pdf"
largest_font_size = 24 # Replace with calculated value
extract_text_with_pdfminer(path, largest_font_size)The scripts are optimized for large PDFs by focusing on specific page ranges and excluding numeric characters during font analysis.
The output JSON format:
[
{
"title": "Title Extracted",
"data": "Associated content"
},
{
"title": "Another Title",
"data": "More content"
}
]Flow chart Image
Contributions are welcome! Feel free to submit issues or pull requests.
This project is licensed under the MIT License. See the LICENSE file for details.