epub2text

A niche CLI tool to extract text from EPUB files with smart cleaning capabilities.

Features

• Smart Navigation Parsing: Supports both EPUB3 (NAV HTML) and EPUB2 (NCX) navigation formats
• Selective Extraction: Extract specific chapters by range or interactive selection
• Flexible Output Formatting:
  • One paragraph per line with customizable separators
  • One sentence per line using spaCy NLP
  • Automatic line splitting at clause boundaries for long lines
• Smart Text Cleaning:
  • Remove bracketed footnotes ([1], [42])
  • Remove page numbers (standalone, at line ends, with dashes)
  • Normalize whitespace and paragraph breaks
  • Preserve ordered lists with proper numbering
• Full Dublin Core Metadata: Extract all EPUB metadata fields
• Rich Interactive UI: Beautiful terminal output with tables and tree views
• Pipe-Friendly: Works as both CLI tool and Python library
• Nested Chapter Support: Handles hierarchical chapter structures

Installation

pip install epub2text

For better HTML parsing performance (optional):

pip install epub2text[lxml]

For sentence-level formatting (requires spaCy):

pip install epub2text[sentences]
python -m spacy download en_core_web_sm

Development Installation

git clone https://github.com/holgern/epub2text
cd epub2text
pip install -e .

Usage

Command Line Interface

List Chapters

Display all chapters in an EPUB file:

# Table format (default)
epub2text list book.epub

# Tree format (shows hierarchy)
epub2text list book.epub --format tree

Extract Text

Extract all chapters:

# To stdout
epub2text extract book.epub

# To file
epub2text extract book.epub -o output.txt

Extract specific chapters by range:

# Single chapter
epub2text extract book.epub -c 1

# Multiple chapters
epub2text extract book.epub -c 1,3,5

# Chapter range
epub2text extract book.epub -c 1-5

# Complex range
epub2text extract book.epub -c 1-5,7,9-12 -o selected.txt

Interactive chapter selection:

epub2text extract book.epub --interactive

Output Formatting:

# One line per paragraph
epub2text extract book.epub --paragraphs

# One line per sentence (requires spaCy)
epub2text extract book.epub --sentences

# Split long lines at clause boundaries
epub2text extract book.epub --max-length 80

# Use empty lines between paragraphs
epub2text extract book.epub --empty-lines

# Custom paragraph separator
epub2text extract book.epub --separator "\t"

Text Cleaning Options:

# Disable all cleaning (raw output)
epub2text extract book.epub --raw

# Keep bracketed footnotes like [1]
epub2text extract book.epub --keep-footnotes

# Keep page numbers
epub2text extract book.epub --keep-page-numbers

# Hide chapter markers
epub2text extract book.epub --no-markers

Output Control:

# Skip first 10 lines
epub2text extract book.epub --offset 10

# Limit to 100 lines
epub2text extract book.epub --limit 100

# Add line numbers
epub2text extract book.epub --line-numbers

Show Metadata

Display EPUB metadata and statistics:

# Panel format (default)
epub2text info book.epub

# Table format
epub2text info book.epub --format table

# JSON format (for scripting)
epub2text info book.epub --format json

Python Library

Use epub2text as a library in your Python code:

from epub2text import EPUBParser

# Parse EPUB file
parser = EPUBParser("book.epub")

# Get metadata
metadata = parser.get_metadata()
print(f"Title: {metadata.title}")
print(f"Authors: {', '.join(metadata.authors)}")
print(f"Language: {metadata.language}")
print(f"Identifier: {metadata.identifier}")

# Get all chapters
chapters = parser.get_chapters()
for chapter in chapters:
    print(f"{chapter.title}: {chapter.char_count:,} characters")

# Extract all chapters
full_text = parser.extract_chapters()

# Extract specific chapters
chapter_ids = [chapters[0].id, chapters[2].id]
selected_text = parser.extract_chapters(chapter_ids)

With custom text cleaning:

from epub2text import EPUBParser, TextCleaner

parser = EPUBParser("book.epub")
text = parser.extract_chapters()

# Custom cleaning options
cleaner = TextCleaner(
    remove_bracketed_numbers=True,
    remove_page_numbers=True,
    normalize_whitespace=True,
    replace_single_newlines=True,
)
cleaned_text = cleaner.clean(text)

With sentence formatting:

from epub2text import EPUBParser
from epub2text.formatters import format_sentences, split_long_lines

parser = EPUBParser("book.epub")
text = parser.extract_chapters()

# One sentence per line
formatted = format_sentences(text)

# Or split long lines at clause boundaries
split_text = split_long_lines(text, max_length=80)

Smart Cleaning Features

The smart text cleaner applies the following transformations by default:

1. Bracketed Footnotes: Removes [1], [42], etc.
2. Page Numbers:
   • Standalone page numbers on their own line
   • Page numbers at the end of lines
   • Page numbers with dashes (e.g., - 42 -)
3. Whitespace Normalization:
   • Collapses multiple spaces into one
   • Standardizes paragraph breaks to double newlines
   • Optionally replaces single newlines with spaces
4. Chapter Markers: Removes internal metadata tags

Chapter Format

Extracted text includes chapter markers in the format:

<<CHAPTER: Chapter Title>>

Chapter text content here...

<<CHAPTER: Next Chapter>>

More content...

Use --no-markers to hide chapter markers.

Requirements

• Python >= 3.9
• click >= 8.0.0
• rich >= 13.0.0
• ebooklib >= 0.18
• beautifulsoup4 >= 4.12.0
• lxml >= 4.9.0 (optional, for better performance)
• spacy >= 3.0.0 (optional, for sentence formatting)

Technical Details

EPUB Parsing Strategy

The parser uses a sophisticated navigation-based approach:

1. Loads EPUB using ebooklib
2. Finds navigation document (prefers NAV HTML, falls back to NCX)
3. Parses navigation structure recursively
4. Maps TOC entries to document positions using fragment IDs
5. Slices HTML content between navigation points
6. Extracts text using BeautifulSoup
7. Applies smart cleaning and normalization

Navigation Support

• EPUB3 NAV HTML: Parses <nav epub:type="toc"> with nested <ol>/<li> structures
• EPUB2 NCX: Parses <navMap> with <navPoint> elements
• Fragment IDs: Robust position detection using BeautifulSoup, regex, and string search
• Nested Structures: Handles hierarchical chapter organization

Metadata Support

Full Dublin Core metadata extraction:

• Title
• Authors (creators)
• Contributors
• Publisher
• Publication Year
• Identifier (ISBN, UUID, etc.)
• Language
• Rights (copyright)
• Coverage
• Description

Documentation

Full documentation is available at Read the Docs.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details

Author

Holger Nahrstaedt

See Also

• abogen: Full-featured audiobook generator with TTS support
• epub2txt: Simple EPUB to text converter (different project)