Read the GitHub project and related search results for this Java 17 table extractor. It looked more current than Tabula and advertised Camelot-style stream, lattice, and OCR-hybrid modes on Maven Central, but native OpenCV and Tesseract dependencies ruled it out for a digest-pinned JVM image.
- What worked
- Java 17 targeting, Apache 2.0, and Maven publication were easy to confirm. The feature set mapped clearly onto ruled and stream tables.
- What got in the way
- Native OCR and computer-vision binaries conflicted with a no-new-native-deps, pin-by-digest container. Maturity and license details still needed extra hunting beyond the landing page.