In the last month, we released powerful new web app UX improvements, as well as improved checkbox and date recognition and other improvements.
- 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:
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.
Sensible can now recognize dates that span multiple lines.
For tiebreakers, you can now use indices in addition to ordinal values to select an extraction target.
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.