adding minify option in Parser, so the parsed Document occupies less memory.

[oymy]

When parsing large html, the memory usage of the parsed Document is huge.

So maybe we can add minify option in Parser,
so the parsed Document could have less nodes, and occupies less memory.

[jhy]

I'm always interested in new ideas on how to use less memory! Can you tell me more about your idea and your use case?

• What memory budget per Document do you have / need?
• What size are the input documents?
• What are you actually doing with the documents (selecting / editing / outputting...?)
• What nodes do you not want to retain when parsing? How would you consider controlling that?

One feature I've been thinking about for a while is to implement a streaming type parser that would just emit tokens / nodes, and perhaps the current stack, but not retain a full DOM. But I see this as being a difficult interface for users of the library to work with. So am interested in hearing your use case (and everyone's! Others please feel free to comment also) and what developer experience would be best.

[oymy]

Thanks for the reply.
Sorry I didn't explain my idea.

Rarely I have Document that may occupies 300M.
The size of of the input document may extends 200M,
I use jsoup to format the html to xml ,then export to pdf.
What I don't want to retain in parsing is empty nodes like "\n" in the html. because they don't affect the final look of the pdf.

for example: comparing the parsed documents of the following two lines. the first line has 5 more empty nodes than the second:
"<tr> <td>a</td> <td>a</td> <td>a</td> <td>a</td> </tr>"
"<tr><td>a</td><td>a</td><td>a</td><td>a</td></tr>"

so minify the html in parsing can save memory spaces.

[jhy]

Thanks -- it makes sense. Have you measured the memory impact of removing those nodes?

Would you be able to share an example of the document? Please contact me directly ([email protected]) if you can. Or, is there an example file I could use as a proxy?

I have been thinking of adding a stream() function to the Parser that would emit a stream of Elements / Nodes as the document is consumed. And one could then node.remove() the unneeded content. Or, just process the content in a streaming mode and delete the processed content.

Would be keen to get real-world examples and beta testers for this functionality.

[jhy]

I've built out a new feature -- StreamParser -- that should address this. Take a look at the examples in #2096. Would be great if you can give it a try.