The task
How can a reader find equipment and technical specifications inside scanned manuals without exposing the original document collection?
My solution
I built VOPT, an offline pipeline that extracts candidate specifications, preserves their evidence, supports review, and releases a controlled metadata catalog for discovery.
Private document processing and public discovery have separate responsibilities. Page extraction preserves units, qualifiers, and source locations. Reviewed observations enter a versioned catalog, where an inspectable query plan finds relevant equipment and specifications.
How it works
Python · local OCR · SQLite FTS · versioned metadata- 01
Extract
Read page text and geometry, using local OCR for image pages. Preserve the supporting passage.
- 02
Review
Inspect candidate observations and their units, scope, and provenance before release.
- 03
Discover
Search a controlled metadata snapshot by equipment, attribute, and value.
Try it
Interactive browser demonstrationInteractive browser demo
Find the specification in the source
Choose equipment and a specification. Inspect the value alongside its operating conditions, quoted evidence, and manual page.
Loading the bundled evidence…
Methods / evidence / reproduction
Read the research.
A source-inspected development comparison records 13 of 17 selected extraction matches for the layout-aware rules and 4 of 17 for the line-rule baseline. The paper identifies the evaluation set and its interpretation.
The paper connects the research question to the implementation, worked examples, and evidence. The project directory includes the runnable software, documentation, and tests.