In the last month, we released powerful new web app UX improvements, as well as improved checkbox and date recognition and other improvements.
Improvement: Web app UX
The web app is now more powerful than ever, with new features coming on the heels of automatically detecting uploaded documents' types and deskewed views of documents.
- Zoom and pan on documents to view small text more easily:
- Click the dot in the gutter next to extracted field IDs to jump to the source SenseML field and highlight the source lines in the PDF:
- Add comments to your SenseML to help with configuration maintenance:
Improvement: Checkbox recognition can bypass pixel rendering
To improve performance, Sensible now preferentially extracts checkbox selection status from PDF metadata, and falls back to pixel recognition of checkboxes only if the metadata is missing.
Improvement: Date types now support multiline recognition
Sensible can now recognize dates that span multiple lines.
Improvement: Indexed tiebreakers
For tiebreakers, you can now use indices in addition to ordinal values to select an extraction target.
Improvement: PDF metadata in extraction
You can now view metadata about the PDF in the extraction response's
file_metadata parameter, for example creation date, author, and modified date. Sensible returns both raw metadata and normalized metadata.