Change Details

One way that Indic-OCR still beats [[ https://tools.wmflabs.org/ws-google-ocr/ | Wikisource Google OCR ]] is that Indic-OCR correctly detects paragraph breaks (usually) and inserts two linebreaks between each paragraph, while we just put a linebreak at the end of each line without any regard for paragraphs. Would it be possible for us to handle paragraphs better or does Google not give us enough information in the results?

One way that Indic-OCR still beats [[ https://tools.wmflabs.org/ws-google-ocr/ | Wikisource Google OCR ]] is that Indic-OCR correctly detects paragraph breaks (usually) and inserts two linebreaks between each paragraph, while we just put a linebreak at the end of each line without any regard for paragraphs. Let's improve our paragraph handling so that less manual editing is needed.

One way that Indic-OCR still beats [[ https://tools.wmflabs.org/ws-google-ocr/ | Wikisource Google OCR ]] is that Indic-OCR correctly detects paragraph breaks (usually) and inserts two linebreaks between each paragraph, while we just put a linebreak at the end of each line without any regard for paragraphs. Would it be possible for us to handleLet's improve our paragraphs better or does Google not give us enough information in the results? handling so that less manual editing is needed.