A natural language processing pipeline for analyzing works of fiction, including entity detection, quotation attribution, and character relationship analysis.

• Python 3.9 or higher
• pip (Python package installer)
• Virtual environment (recommended)

1. Clone the repository:

git clone https://github.com/yourusername/booknlp.git
cd booknlp

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows, use: venv\Scripts\activate

3. Install the required packages:

pip install --upgrade pip
pip install -r requirements.txt

4. Install the spaCy model:

python -m spacy download en_core_web_sm

1. Make sure your virtual environment is activated:

source venv/bin/activate  # On Windows, use: venv\Scripts\activate

2. Run BookNLP on a text file:

./run_booknlp.py input_file.txt --output-dir output/directory

• input_file: The text file to process (required)
• --output-dir: Directory where output files will be saved (default: 'output')
• --model: Model size to use - 'big' or 'small' (default: 'small')

The pipeline generates several output files in the specified output directory:

1. {book_id}.tokens: Word-level information including:

   • Paragraph and sentence IDs
   • Word forms and lemmas
   • Part-of-speech tags
   • Dependency relations
   • Event annotations

2. {book_id}.entities: Named entity information including:

   • Entity types
   • Coreference IDs
   • Text spans

3. {book_id}.quotes: Quotation information including:

   • Quoted text
   • Speaker attribution
   • Coreference information

4. {book_id}.supersense: Semantic categories for words:

   • Verb categories
   • Noun categories

5. {book_id}.event: Event annotations including:

   • Event types
   • Participants
   • Temporal information

6. {book_id}.book: JSON file containing:

   • Character information
   • Relationships
   • Actions
   • Attributes

7. {book_id}.book.html: Interactive HTML visualization of the text with:

   • Entity annotations
   • Character relationships
   • Interactive features

# Process a text file named "emma.txt" with maximum accuracy
./run_booknlp.py emma.txt --output-dir output/emma --model big