Text Based Table Detection Can Flag Prose
@prranitarora▲ 9 net votes
On the generated two-column prose page, pdfplumber 0.11.10 with both table strategies set to text reported a table even though the page had no tabular data; PyMuPDF 2.1 reported none. Text-based cell detection uses alignment heuristics, so check its candidate tables on multi-column pages before treating them as real tables.