B4J Question xmlsax cannot parse correctly when elements and text are mixed

[xulihang]

I am trying to parse xliff which is a kind of xml for localization.

There will be elements like: <g id="1">Styles like <g id="2">bold</g> is supported.</g>

I use xmlsax to parse it and I cannot get the first text part "Styles like " in the _EndElement event.

The result using xml2map:

(MyMap) {g={Attributes={id=1}, g={Attributes={id=2}, Text=bold}}}

I try to use jsoup to parse it and it has the right result:

(TextNode)
Styles like
(Element) <g id="2">
bold
</g>
(TextNode)
bold
(TextNode)  is supported.

Is xmlsax wrapped correctly? I prefer to use xml2map which parses xml to map.

 

[xulihang]

I checked the wrapper's source code. I think there should be a characters event to handle text.

        @Override
        public void characters (char ch[], int start, int length)
        throws SAXException
        {
            String str = new String(ch, start, length);
            sb.append(str);
            ba.raiseEvent2(null, true,charactersEvent ,false, str);
        }
        }

 

[Erel]

The full string should be added. You are right that it will only be raised at the end.

 

[xulihang]

So how to handle this situation?

I have tried to create a Name key to store the name and a children key to store these parts in sequence. But as a result, it cannot be converted back to xml using map2xml.

Map:

(MyMap) {g={Attributes={id=1}, Name=g, children=[Styles like , {Attributes={id=2}, Name=g, children=bold},  is supported.]}}

 

[Erel]

This is indeed a document that is not suitable for XmlSax. Can you upload a full XML example?

 

[xulihang]

I have attached a lite version of the entire xml which is generated by okapi tikal from a docx file.

 

[Erel]

1. Such XML cannot be represented with a map so it will never work with Xml2Map.

2. You can parse it with MiniHtmlParser. I don't know whether it will be simpler than jSoup.

Dim parser As MiniHtmlParser
parser.Initialize
Dim root As HtmlNode = parser.Parse(File.ReadString("C:\Users\H\Downloads\321880_ms.docx.xlf", ""))
parser.PrintNode(root)

 

[xulihang]

MiniHtmlParser is a better solution. I will try to use it in my scenario.

An xml writer is still needed to convert modified existing xml node to string.

 

[Erel]

It should be possible to write it with XmlBuilder.

 

[xulihang]

I found that after a text node is added, XMLBuilder no longer allows adding an element, which will throw an IllegalStateException (see: assertElementContainsNoOrWhitespaceOnlyTextNodes) . I have posted an issue about this.

 

[xulihang]

I tried MiniHtmlParser and it cannot parse <source xml:lang="en"><g id="1">[Global Notes:</g></source> correcly.

I made a new xmlparser based on a new jXmlSax.jar which raises characters event.

I also wrapped javax.xml.parsers.DocumentBuilder to build xml.

Files are attatched.

 

[Erel]

tried MiniHtmlParser and it cannot parse <source xml:lang="en"><g id="1">[Global Notes:</g></source> correcly.

Not sure that you are correct.

Tested with this:

Dim root As HtmlNode = parser.Parse($"<source xml:lang="en"><g id="1">[Global Notes:</g></source>"$)
parser.PrintNode(root)

The output looks correct.

 

[xulihang]

*** root ***
 *** source ***
 |lang: en|
 *** g ***
 |id: 1|
  *** text ***
  |value: [Global Notes:|

The result I got is like this.

 

[Erel]

Looks correct (the attribute xml:lang was renamed to lang but this is probably not the problem that you are talking about).

What is the problem?

 

[xulihang]

The source element should be the parent of the g element.

The result my xmlbuilder gives from the node is like this: <root><source lang="en"/><g id="1">[Global Notes:</g></root>

 

[Erel]

I see. This is indeed not a good use case for MiniHtmlParser as it is not a valid html text.

The source element must be empty: https://developer.mozilla.org/en-US/docs/Web/HTML/Element/source

This code sets the elements that will be closed automatically:

VoidTags = B4XCollections.CreateSet2(Array("!DOCTYPE", "area", "base", "br", "col", _
        "embed", "hr", "img", "input", "link", "meta", "param", "source", "track", "wbr"))