[jsword-devel] [sword-devel] Web Front-end for JSword (iPhone targeted)

[Greg Hellings]

On Sat, Apr 19, 2008 at 1:11 PM, DM Smith <dmsmith555 at yahoo.com> wrote:
>
>
> On Apr 19, 2008, at 1:40 PM, Greg Hellings wrote:
>
> On Sat, Apr 19, 2008 at 12:20 PM, DM Smith <dmsmith555 at yahoo.com> wrote:
>
>  On Apr 19, 2008, at 12:43 PM, Greg Hellings wrote:
>
>
> On Sat, Apr 19, 2008 at 8:42 AM, DM Smith <dmsmith555 at yahoo.com>
>
> wrote:
>
> On Apr 19, 2008, at 5:22 AM, Greg Hellings wrote:
>
>
>
>
>
>
>
> For more details see my post on jsword-devel:
>
>
>
> http://www.crosswire.org/pipermail/jsword-devel/2008-April/
>
> 002453.html
>
>
>
> I read through it.  It seems most of the problem you're encountering
>
> is with client-side XSL transformations.  I don't know what exactly
>
> you're doing with the XML on the client side.  My version is just
>
> sending the output of getRawText(user_key) across the wire.  I'd like
>
> to send the transformed HTML, but I get a NullPointerExcepion when I
>
> invoke ConverterFactory.getConverter() from my servlet.  When I
>
> iterate over the Converters that it recognizes, the only one is the
>
> ConfigurableSwingConverter, but even trying to retrieve that Converter
>
> throws the exception on line 55 of ConverterFactory.java.
>
>  Try including the bibledesktop*.jar. It is in there.
>
> All of those are included, and the class exists within the project (I
> can instantiate the SAX Converter directly - though I don't know the
> URI that it's looking for).  It just seems that ConverterFactory isn't
> picking up on its existence for some reason.
>
> That's odd.
> Then use the APIExample's example of searchAndShow. It gives how you can
> supply your own stylesheet, but it uses BibleDesktop's simple.xsl. Note,
> this needs to be tweaked to produce appropriate output for the web.
> Essentially that's all that the ConverterFactory is doing in a round about
> way.

This worked perfectly (once I figured out where GWT expected to find
the relative resource path!).

>
>
>
>
>
>
>  If I
>
> understood more of how to pull the pure XML from the module and where
>
> to find the XSLT, I could do the transformation client-side with GWT's
>
> built-in XSL processor, as it would only make sense for me to be
>
> displaying HTML and not RawText in a browser window.
>
>  See o.c.jsword.bridge.DwrBridge and o.c.jsword.examples.APIExample
>  for, well, examples.
>
> APIExamples has been my crutch throughout the whole project.  It's the
> code from getStyledText() in the examples that I'm using to try and
> convert to HTML on the server side.  However, the above
> ConverterFactory error appears to be the only thing hampering my
> efforts.  I'll take a look at DwrBridge.
>
> DwrBridge expects client side xslt.

Since I'm targeting a mobile platform, I think that transforming on
the server is probably the best.

>
>
>
>
>
> There are still some rough edges with the interface - primarily it
>
>
> doesn't properly detect the books that are in a module at any given
>
>
> moment.  It does detect chapters-in-book and verses-in-chapter,
>
>
> though.  For now, you just have to assume that if you get to the
>
>
> verse
>
>
> display and there is no text (and no error box), then that module
>
>
> must
>
>
> not support that verse/book.  There are also additional GUI tweaks
>
>
> that will be needed to smooth the functioning of it, but this is
>
>
> just
>
>
> a quickly hobbled together patch-up.
>
>
>
>
>
> One of the things that I think that is needed in JSword is that the
>
> global key list actually reflect the content of the module. For
>
> Bibles
>
> and commentaries, it assumes Gen-Rev. I'll probably do the same thing
>
> I did for dictionaries there too.
>
>
>
> I had tried that method (get the global key list, then try the
>
> .contains() method) and to my surprise, it told me that all of the
>
> Bible was available in the LXX.  I likewise tried catching exceptions
>
> from the .getKey() method, but that seems to only throw them if the
>
> Key doesn't exist in the Bible at all, rather than if it doesn't exist
>
> in this module.  However, having each Bible support the
>
> .booksInModule() as well as its own implementation of
>
> .chaptersInBook() and ..versesInChapter() would allow for ease of use
>
> if alternate versification ever pops up.  Currently the BibleInfo
>
> class works fine for the latter two, but I can't figure out any way to
>
> narrow down the books available.  It sounds like this is a "ToDo" on
>
> your list, though.
>
>  Books available can be done *today* by checking chapter 1, verse 1 of
>  each book.
>
> I must not be checking them correctly.  I've done it by (1) calling
> getGlobalKeyList() and then checking "<book> 1:1" for all 66 books and
> seeing if they're in the returned global key list and (2) calling
> getKey() on "<book> 1:1" and catching exceptions - if an exception
> occurs then I assume the book is not in the module and if none occurs
> I add the book name to the list of valid books.  However, both of the
> above methods return all 66 books as valid regardless of which module
> I'm using.  I couldn't find any examples of how to check it.  What is
> the correct method?
>
> The global key list is a guess. It is not based upon the content of the
> module. It is merely "Gen-Rev".
>
> To check if a verse is actually present you need something like the
> following:
> Key key = book.getKey("Gen.1.1");
> String raw = book.getRawText(key);
> boolean hasVerse =  (raw.length() > 0);

Worked like a charm!

--Greg

>
>
> _______________________________________________
>  jsword-devel mailing list
>  jsword-devel at crosswire.org
>  http://www.crosswire.org/mailman/listinfo/jsword-devel
>
>