text-peeler 1.0.0

PDF text, table, image, and form extraction utilities

Text Peeler

One command to extract text, tables, images, and forms from any PDF.

Text Peeler analyzes your document, picks the right extraction strategy, and delivers clean output in the format you need. Digital PDFs, scanned pages, mixed documents, ebooks: one tool handles them all.

Install

pip install text-peeler

OCR support requires Tesseract:

brew install tesseract    # macOS
apt install tesseract-ocr # Debian/Ubuntu

See the full installation guide for all options.

Quick Start

# Auto-detect and extract
text-peeler-detect report.pdf

# Extract text from a digital PDF
text-peeler-native report.pdf

# Pull tables as JSON
text-peeler-tables report.pdf --format json

# Run every relevant extractor at once
text-peeler-ensemble report.pdf

Or use the shell router:

./extract.sh auto report.pdf
./extract.sh tables report.pdf output.csv --format csv

See the quickstart guide for more examples.

What It Does

Mode	Purpose
native	Digital PDFs with selectable text
scanned	Image-only PDFs via OCR
mixed	Per-page routing (native or OCR)
tables	Structured table extraction
images	Embedded images with surrounding context
forms	Fillable form field extraction
epub	EPUB ebook chapter extraction
ebook	Legacy ebook formats (.mobi, .lit, .prc)
detect	Analyze a PDF and recommend extractors
ensemble	Run all relevant extractors, merge into one JSON

Output Formats

Every extractor supports multiple output formats. Choose what fits your pipeline.

Extractor	Default	Supported
native	txt	txt, json, md
scanned	txt	txt, json, md
mixed	txt	txt, json, md
tables	json	json, csv, txt, md
images	json	json, md, txt
forms	json	json, txt, md, csv

See output format details for schema documentation.

Use Cases

Text Peeler is built for pipelines. Here are the most common workflows:

• LLM Ingestion : Feed PDFs into language models as clean, structured text
• Batch Processing : Process hundreds of mixed PDFs in a single pass
• Scanned Documents : OCR pipeline with smart per-page text density checks
• Table Extraction : Pull tabular data out of PDFs as CSV or JSON
• Form Extraction : Extract fillable form fields with type-aware parsing
• Image Extraction : Recover embedded images with page context
• Ebook Conversion : Chapter-level text from EPUB and legacy ebook formats
• Document Auditing : Characterize a PDF without extracting anything

Architecture

Each extractor is a standalone Python script. No shared base class, no deep inheritance. Shared formatting lives in output_utils.py.

text_peeler/
├── detect.py           # PDF analysis + routing
├── ensemble.py         # Multi-extractor runner
├── output_utils.py     # txt/json/md/csv formatting
└── extractors/
    ├── native.py       # pymupdf text
    ├── scanned.py      # pymupdf + tesseract OCR
    ├── mixed.py        # per-page routing
    ├── tables.py       # pdfplumber tables
    ├── images.py       # pymupdf images
    ├── forms.py        # pymupdf form widgets
    ├── epub.py         # ebooklib EPUB
    └── ebook.py        # Calibre ebook conversion

See the architecture guide for implementation details.

Documentation

Guide	Description
Installation	All install methods, system dependencies, troubleshooting
Quickstart	First extraction in under a minute
CLI Reference	Every flag and option for every mode
Output Formats	JSON schemas, CSV layouts, Markdown structure
Architecture	How the pieces fit together

See Also

Gutenfetchen (PyPI) - Bulk download and process public domain texts from Project Gutenberg. Pairs well with Text Peeler for building large text corpora from mixed sources.

License

AGPL-3.0