[jsword-devel] strongs search

[DM Smith]

Mullins, Steven wrote:
> I had to delete my KJV module and reinstall it to get it 
> to work right.  Now I can now search the hebrew strongs
> just fine.  Both strong:h430 and strong:h0430 work 
> thanks to an existing regular expression matcher in 
> OSISUtil.java:
>
> strongsNumber = "strong:([GH])0*([0-9]+)";
>
> However, the greek strongs searches will not work for me,
> though I see no reason why not.
>   
There's a tool that can be used to view lucene indexes. I don't remember 
what it is called or where to find it. But that would help figure out 
what is going on.

> The LXX module is displaying strongs numbers as [xxxx] like 
> the MorphGreek module.  The xls changes to fix MorphGNT have 
> affected it. I assume that the lemma is the same as the 
> MorphGNT.  My 1.0.8 version does not have a "view source" 
> option to verify.  At least I can't find the view source
> button.
>   

ViewSource was made a hidden option. Just edit 
~/.jsword/desktop.properties Advanced.ViewSource=true (or something like 
that).

> Really the LXX module is more correct in calling the strongs
> number lemma.Strong, since it actually is.  The MorphGNT would
> be more correct if it used say lemma.Lexical, since it has
> little to do with a strongs number.
>   
This would be a good discussion for sword-devel. The prefixes have 
seemed to be a bit ad-hoc.

There are a couple of tools and examples to dig into a module:
The program o.c.j.bridge.BookExporter will dump a module into SWORD's 
imp format.
In JSword o.c.j.examples both APIExamples and BibleToOSIS can be used to 
dump the content of the Bible.


> DM, what do you suggest to differentiate between LXX and 
> MorphGNT lemmas?
>
> I was working on a code to allow searchs by the "lex:" field
> for the MorphGNT module, and using DM's helpful suggestions
> following the pattern of used for strongs numbers.  But since
> I can't get strongs to work in greek, I'm sorta stuck.  I'm
> attaching the code I have so far for the lex lemma.  I think 
> I'm going down the right track, but want to make sure.
>   
The code looks fine.
Given how arbitrary the lemma attribute is, we might want to take a page 
from the C++ developers for now.

Their approach is that the lemma has several different values in it. 
Either of the form A:B or just A.
When A = strong: we index it as a Strong's number otherwise we index it 
as a "lemma" having 1 or 2 words. So one could then search lemma:["X Y 
value"] where it was lemma="X.Y:value".

I think I'd suggest a speedup change to the different getXYZ(Element) 
methods to get everything out of an element by it's type, E.g. searching 
for <w> elements we get all Strong's numbers, all morphs and all lemmas.

My goal was feature first, then optimize. I just never got to the optimize.

When you are happy with what you want to add, send a final patch and 
I'll apply it.

Hope this helps,
    DM
> Take care,
>
> Steve