v0.4.2

Fix handling of documents without page spans.

v0.4.1

Fix potential memory leak by closing PDFs explicitly.

v0.4.0

Added pdf.layout.fetch and pdf.spans.manual recipes.

v0.3.0

This release adds support to pages UI introduced in Prodigy v1.17.0.

v0.2.2

This release adds porting over the meta field to the resultant dataset so that for example the document information can be preserved.

v0.2.1

Bumped Pillow dependecy to <11.0.0

v0.2.0

This release adds support for pdf.ocr.correct. This recipe adds OCR to annotated segments from the pdf.image.manual recipe and gives a textbox for corrections. This recipe uses pytessaract under the hood.

v0.1.0

This initial release introduces a pdf.image.manual recipe that adds a workflow to annotate PDF pages as if they are images. We also add CI and unit tests.