Repository navigation
Releases
Releases · IsHubolla/layout-aware-pdf2json
Initial release
layout-aware-pdf2json v1.0.0
Initial release. A command-line tool that extracts structured content — text
blocks and tables — from PDF files and exports it as JSON, built on
MuPDF.
Features
- Table reconstruction — detects tables via MuPDF's structured text
GRID API, merges split/overlapping cells, and groups them into logical
blocks, including multi-row/column spans. - Text block extraction — pulls free-text content alongside tables.
- Reading order — orders all page items (text blocks + tables) using a
recursive XY-cut algorithm, correctly handling single-column,
multi-column, and asymmetric layouts. - JSON output — one JSON file per run, with one entry per page
containingblocksandtables, each with bounding boxes and reading
order.
Downloads
Prebuilt binaries below, one archive per platform:
layout-aware-pdf2json-v1.0.0-ubuntu.tar.gzlayout-aware-pdf2json-v1.0.0-rhel.tar.gzlayout-aware-pdf2json-v1.0.0-darwin.tar.gzlayout-aware-pdf2json-v1.0.0-windows.zip
Usage
./mupdfwrapper <input.pdf> [output.json]Requirements & building from source
See the README for build instructions and
notes on MuPDF's licensing (AGPL-3.0 / commercial dual license).
License
AGPL-3.0 — see COPYING.txt.