Changelog

All notable changes to this project are documented here. The format follows Keep a Changelog.

0.1.1 - 2026-06-28

First Clojars release. (0.1.0 was not published: its POM omitted the license.)

Added

• Document lifecycle: open-pdf (path / File / byte[] / InputStream) and the with-pdf macro, with an ex-info error model (:invalid-input, :encrypted-pdf, :parse-failed, :page-not-found).
• metadata, pages, and page (1-based) APIs.
• Text extraction: chars, words, and text, with :page, :bbox, :x-tolerance, and :y-tolerance options.
• Object extraction: objects (lines, rectangles, curves), with line :orientation classification and :page/:types/:bbox options.
• crop-page page views, accepted by chars/words/text/objects in place of a document handle to restrict extraction to a bbox.
• Table extraction: extract-table/extract-tables with :lines (grid from ruling lines) and :text (columns inferred from word alignment) strategies, returning :rows/:cells/:debug.
• pdfplumber.geometry: bbox helpers and PDFBox↔top-left coordinate conversion.

Fixed

• Word splitting now breaks on whitespace characters in addition to wide gaps, so words separated by a narrow space are no longer merged. Validated against the Python pdfplumber corpus (text-similarity and word-count median 1.0).