[Feature Request] Evaluate Tapirscan as an alternative barcode reader

Description I built [Tapirscan]( an open-source barcode reader with Python bindings. ZXing reads clear, aligned codes well; I wanted something that also holds up when codes are tilted, skewed or blurred, which also happens with scanned and photographed documents. My [public benchmark]( compares Tapirscan with ZXing (via ZXing WASM, the same decoder family Paperless uses through ZXing-C++). Tapirscan had higher recall on rotated, perspective-distorted and blurred images. These are retail product photos, not documents, so they don’t show how it would do on separator pages or ASN labels. Would there be interest in comparing the two on document barcodes? I have a local prototype that can switch Paperless’s barcode reading to Tapirscan, and I’d be happy to measure missed and incorrect reads on representative scans before anyone considers a change. For your internal test set, both ZXing and Tapirscan perfectly read all codes. Other This post and the local prototype were prepared wi…