PDF Word Counter

A tool to search PDF files for specific words and generate frequency statistics. Supports both PyPDF2 and pdfminer.six backends, with output to Astropy or Pandas formats.

Features

• Case-sensitive/insensitive search
• Unicode text normalization (NFC, NFD, NFKC, NFKD)
• Parallel processing using threads
• Output to Astropy tables or Pandas DataFrames
• Select specific pages or page ranges
• CLI and Python API
• Filename parsing to extract metadata
• Optional columns and sorting

Installation

Basic installation:

pip install pdf-word-counter

Full installation (with all optional dependencies):

pip install pdf-word-counter[miner]    # + pdfminer.six
pip install pdf-word-counter[astropy]  # + Astropy Tables
pip install pdf-word-counter[full]     # All extras

Coomand Line Usage

Basic Example

# Count words in all PDFs (case-insensitive) and print table
pdf-word-counter "AI,ML" papers/*.pdf --show

Advanced Examples

# Search pages 1-5 with Unicode normalization
pdf-word-counter "café" doc.pdf --pages 1-5 --unicode --form NFKC

# Case-sensitive search with 4 parallel workers
pdf-word-counter "Python" docs/*.pdf --case --workers 4

# Case-sensitive search with pdfminer
pdf-word-counter "Python" docs/*.pdf --case --miner --outfile counts.csv

# Extract project names from filenames like "ProjectX_2023.pdf"
pdf-word-counter "algorithm" *.pdf --dsep "{'_': {'project': 0}}" 

CLI Options

Filename Parsing Options

Output Column Toggles

Python API

from pdf_word_counter import search_pdfs

results = search_pdfs(
    ["document1.pdf", "document2.pdf"],
    ["word1", "word2"],
    output_file="results.csv",
    show=True,
    sort_cols=["count"],
    ignore_case=True,
    use_pdfminer=False,
    workers=4,
)

License

MIT

How to Cite

This code was used in the research published as:

García-Benito, Rubén (2025). Beyond Universality: Cultural Diversity in Music and Its Implications for Sound Design and Sonification. Audio Mostly & ICAD Joint Conference (AM.ICAD 2025). Association for Computing Machinery, June 30 – July 4, 2025, Coimbra, Portugal

The code version corresponding to this publication is archived and can be cited using the DOI:

If you use this software in your research or adaptations, please cite the above paper and this repository accordingly.