PDF to HTML

Changed in 1.21.0: Structured HTML, Options.

PdfToHtml turns any PDF, not only those PdfPecker made, into one HTML file that looks like the PDF. Each page is an image of the page, and the page's text is laid invisibly over it line by line, so it can be selected, copied, found with the browser's search (whole phrases too) and read by screen readers. The PDF's web, mail and phone links and its links between pages keep working. Use it to show statements, invoices or reports in a web page or an e-mail portal without a PDF viewer. The structured mode writes the text instead as headings, paragraphs, lists and tables that flow, for search indexes, language models and small screens.

// Any PDF as one self-contained HTML page
string html = PdfToHtml.Convert(File.ReadAllBytes("statement.pdf"));

// File to file, or straight to a stream a few pages at a time, so a long PDF is never held in memory whole
PdfToHtml.ConvertFile("report.pdf", "report.html");
using (var file = File.Create("report.html")) PdfToHtml.Convert(pdfBytes, file);

// Some pages of an encrypted PDF, as sharper JPEG images
string first = PdfToHtml.Convert(pdfBytes, new PdfHtmlOptions
{
    Password = "pw",
    Pages = PdfPageRange.Parse("1-3"),
    Dpi = 200,
    ImageFormat = PdfImageFormat.Jpeg,
    Quality = 80,
});

Convert also takes a Stream, and every method takes a CancellationToken.

The HTML

The output is one self-contained HTML file with no scripts: the page images, the style sheet and the small web font of the text layer are all inside it, so it can be saved, mailed or served as it is. Pages scale to the width of the window, down to a phone screen, and print on paper of their own size. Every page is a div with the class pdf-page and the id page- followed by its page number, so report.html#page-3 opens the third page. Each line of the text layer is one box with its words inside, so the browser's find matches phrases across words, not only single words.

The pages are drawn by PdfPecker's own PDF renderer and the text comes from text extraction, so everything those can read works here: any fonts, JPEG 2000, JBIG2 and CCITT images, encrypted PDFs, form fields and annotations. Pages are converted in parallel.

Structured HTML New in 1.21.0

Mode = PdfHtmlMode.Structured writes the pages' text as flowing headings, paragraphs, lists and real <table> elements in reading order, with nothing placed at a fixed position. A tagged PDF (PDF/UA, or any PDF with a structure tree) is read from its tags, which say exactly what each heading, list item and table cell is. Other PDFs are read from their layout: tables come from the rules and borders the PDF draws and from text that lines up in columns, with header rows, spanned cells, wrapped cells and indented row labels.

// Plain structure: one readable style, no pictures
string text = PdfToHtml.Convert(pdfBytes, new PdfHtmlOptions { Mode = PdfHtmlMode.Structured });

// The same structure with the PDF's fonts, colours, backgrounds, borders and pictures
string styled = PdfToHtml.Convert(pdfBytes,
    new PdfHtmlOptions { Mode = PdfHtmlMode.Structured, IncludeStyles = true, IncludeImages = true });

IncludeStyles keeps how the page looks as CSS classes of each page: the fonts, sizes and colours of the text, the backgrounds, rounded corners and border lines of table cells and coloured boxes (dashed ones too), where text sits in its cell, and the widths of tables and columns. The text still flows. IncludeImages adds the pictures the pages draw, such as a logo, as JPEG or PNG images where they sit in the flow; a picture that fills most of a page (a scan under its text), image masks, turned pictures and pictures too small to see are left out. The image options of the table below do not apply to structured HTML.

Options Changed in 1.21.0

What it is not

Visual HTML keeps the PDF's layout rather than reflowing it into paragraphs and tables, and the text is shown in the image rather than in the PDF's fonts, so it zooms as an image does. Structured HTML reflows, but an untagged PDF's structure is read from its layout, so unusual layouts, floats and charts can come out as plain paragraphs. Text that the PDF draws only as outlines or images, such as a scan, can be seen but not selected, since there is no OCR. See Limitations.