Skip to content

CLAUDE.md

Latest commit

 

History

History
50 lines (32 loc) · 2.08 KB

CLAUDE.md

File metadata and controls

50 lines (32 loc) · 2.08 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

Language-experiments is a Python project that converts literary texts into visual representations using linguistic analysis. It maps word and bigram frequency data to color values, creating images that represent textual patterns.

Running the Code

# Generate visualization from a text file
python book_png.py <path_to_text_file>
# Output: <filename>-lenbig.png

No build system or test framework exists. Dependencies: NLTK, PIL/Pillow, Flask, SQLAlchemy.

Architecture

Core Visualization Engine (book_png.py)

Four visualization methods that convert tokenized text to PNG images:

  1. blue_red_gradient_words - Maps word frequency to blue-red gradient using logarithmic scaling
  2. random_colored_words - Assigns consistent random colors per unique word
  3. conditional_probabilities_bigrams - Colors based on bigram transition probabilities (disabled)
  4. possible_bigrams - Colors based on bigram diversity per word (currently active)

Processing pipeline: Text → nltk.word_tokenize() → frequency analysis → color mapping → square PNG output

Image dimensions are automatically calculated as √n × √n where n = number of words/bigrams.

Web Application (web/textnav.py)

Skeleton Flask app with SQLAlchemy models (User, Book, Picture, PictureParams) using SQLite. No routes implemented - database schema only.

Key Patterns

  • Functional decomposition: each visualization method is self-contained
  • NLTK for all NLP operations (tokenization, bigrams, frequency distributions)
  • PIL for pixel-by-pixel image generation
  • Square grid layout maps linear sequences to 2D images
  • RGB channels encode different frequency metrics

Code Notes

  • Written for Python 2.x (uses deprecated patterns like has_key(), old Flask-SQLAlchemy imports)
  • No error handling or input validation
  • Active function in book_png.py is possible_bigrams() at the bottom of the file