如何使用 Jsoup 将 HTML 文本（包括无序列表和有序列表）转换为纯文本？

Question

I need to convert HTML to plain text for sending it per mail. Currently I'm using

Jsoup.parse(html).wholeText();

This preserves line breaks, but not lists. Something like

 - List item
 - List item 2
   - Nested list item

gets converted to List itemList item2Nested list item

How can I keep most of the text formatting, but remove all HTML tags with images, links etc.?

Answer 1

What you're asking for is to render HTML (not parse it; though parsing it is, naturally, part of any HTML rendering engine). Not render it the way eg Chromium would render it (as an image to a screen), but to render it into a string.

This is highly complicated, and involves CSS support as well. In basis, what you are asking for is multiple personyears of effort, and as far as I know no library exists that did it. You can have a look at text-based HTML renderers such as Lynx or w3m - you can probably install them, execute these with ProcessBuilder (this does, of course, make your app entirely arch+OS dependent, and you'll have to ship a w3m or lynx binary for each and every platform you want to support, or ask the one who installs your app to take care of also installing a lynx and/or w3m and telling your app where it is). Note that lynx/w3m tend to assume full terminal support, meaning: Bold, colours, etc.

Imagine an HTML page that doesn't use <ul> and <li> to create a bulleted list, but instead uses some CSS to make something that looks a lot like a bulleted list. Or what if inline CSS is used to align something to the right. Presumably then you would expect the string to also do this right alignment, except that is completely impossible unless either [A] you know the size of the 'window' the string will be rendered into or [B] the output is not basic text strings but some sort of markup language that supports right aligning (be it HTML or RTF or similar), or [C] terminal command sequences are available to move the cursor around.

This should highlight how your question is in essence 'weird' - it's either incredibly complicated, or a seemingly arbitrary tiny subselection of what HTML can do.

If the latter piques your interest, it isn't too difficult to just write a simplistic tree walker that specifically inserts newlines and "\n * " any time a <li> element inside a <ul> is visited, and a String.format("\n%2d. ") anytime a <li> is visited inside an <ol> .

In other words, given that what you ask for is either impossible or is an arbitrary choice of HTML and CSS stylings that you do and don't want to support, write it yourself. If truly you are only interested specifically in <ol> / <ul> based lists and nothing else, this will be about a page full of code and no more.