Re: [mylyn-integrators] [WikiText] Is it possible to extract attributes

[David Green <dgreen99@xxxxxxxxx>]

Luven,

The good news: The WikiText parser is used behind the scenes in the WikiText editor -- so yes, it can give you the offsets that you need to know where things are in the source markup.  The bad news: this functionality was retrofitted into the WikiText parser and is probably not the most intuitive API.

You're pointed in the right direction.  org.eclipse.mylyn.wikitext.core.parser.Locator is indeed the place to get the offsets that you need.  The WikiText markup editor uses a DocumentBuilder to build a model of the document with exact offsets.  A good place to start looking is org.eclipse.mylyn.internal.wikitext.ui.editor.syntax.FastMarkupPartitioner.PartitionBuilder.

If you feel that there are bugs in the implementation (including documentation bugs or lacking documentation) please post a bug at bugs.eclipse.org under Tools/Mylyn/WikiText.  If you're able to attach JUnit tests that exercise/demonstrate the bug, that's even better.

To satisfy my curiosity, perhaps you could tell me more about your use case: Why do you want to know these offsets?

Regards,

David

On Wed, Jul 15, 2009 at 8:24 AM, siluven <siluven@xxxxxxxx> wrote:

Hello,

now I like to get the begin and end offset of each links in wikitext. Later maybe other components (images).
Is that possible to do this using Mylyn WikiText?

I tried to use getLocator().getDocumentOffset() in method:

void org.eclipse.mylyn.wikitext.core.parser.builder.NoOpDocumentBuilder.link(Attributes attributes, String hrefOrHashName, String text) and
void org.eclipse.mylyn.wikitext.core.parser.builder.NoOpDocumentBuilder.characters(String text)

but I found out that document offset is not updated after calling method link(Attributes attributes, String hrefOrHashName, String text).
The offset is also sometimes decremented. So far I now, it should be only incremented.
I test by using this simple text below as my MediaWiki WikiText:

The '''EditorX''' is an [[text editor|editor]] of small to medium-sized [[text]].
This is a [[test]] too ('''yes''').

The Result:

Offset	Component type	Value	Comment
-1	DOCUMENT_BEGIN	[]
0	BLOCK_BEGIN	[PARAGRAPH]
0	CHARACTERS_GROUP	[The ]
4	SPAN_BEGIN	[BOLD]
7	CHARACTERS_GROUP	[EditorX]
7	SPAN_END	[BOLD]
17	CHARACTERS_GROUP	[ is an ]
7	LINK	[editor]	The offset is anyhow decremented (¿Bug?)
7	CHARACTERS_GROUP	[ of small to medium-sized ]	Now is all offset incorrect
55	LINK	[text]
55	CHARACTERS_GROUP	[.]
83	CHARACTERS_GROUP	[]	New line position is correct now.
83	CHARACTERS_GROUP	[This is a ]
93	LINK	[test]
93	CHARACTERS_GROUP	[ too (]	93 is offset of the link
107	SPAN_BEGIN	[BOLD]	Here is correct again
110	CHARACTERS_GROUP	[yes]
110	SPAN_END	[BOLD]
116	CHARACTERS_GROUP	[).]
118	BLOCK_END	[PARAGRAPH]
118	DOCUMENT_END	[]

For heading I am currently using:

getLocator().getLineDocumentOffset() for beginOffset and
getLocator().getLineDocumentOffset()+getLocator().getLineLength() for endOffset

and it works so far.
It does not work for link because there can be some links in the same line.

Best regards,

Luven

On 6/25/2009 6:59 PM, siluven wrote:

Thank you David,
that is the functionality I need. I've tried also with headings and images.
And it works too.

Best regards,
Luven

On 6/24/2009 6:47 PM, David Green wrote:

You want to do something like this:

public class ExtractHyperlinksBuilder extends NoOpDocumentBuilder {

private Set<String> hyperlinks = new HashSet<String>();

@Override

public void link(Attributes attributes, String hrefOrHashName, String text) {

hyperlinks.add(hrefOrHashName);

}

@Override

public void imageLink(Attributes linkAttributes, Attributes imageAttributes, String href, String imageUrl) {

hyperlinks.add(href);

}

public Set<String> getHyperlinks() {

return hyperlinks;

}

}

MarkupParser parser = new MarkupParser(ServiceLocator.getInstance().getMarkupLanguage("MediaWiki"));

ExtractHyperlinksBuilder builder = new ExtractHyperlinksBuilder();

parser.setBuilder(builder);

Reader markupContent = null;// open reader

try {

parser.parse(markupContent);

} finally {

markupContent.close();

}

// do something with builder.getHyperlinks()

Regards,

David

On Wed, Jun 24, 2009 at 4:13 AM, siluven <siluven@xxxxxxxx> wrote:

Hello everyone,
I am a new Mylyn user. I'm planning to work with Wiki Articles with java and eclipse.

Is that possible to extract attributes like hyperlinks, headings, images, etc. directly from e. g. wikimedia markup-language using WikiText.
like:
- obj.getHyperlinks();

If it is possible or maybe there is solutions for this, how could it be done?

Thank you and best regards
Luven

_______________________________________________
mylyn-integrators mailing list
mylyn-integrators@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/mylyn-integrators

_______________________________________________
mylyn-integrators mailing list
mylyn-integrators@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/mylyn-integrators
  

_______________________________________________
mylyn-integrators mailing list
mylyn-integrators@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/mylyn-integrators
  

_______________________________________________
mylyn-integrators mailing list
mylyn-integrators@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/mylyn-integrators