Hey guys. I recently had to port the mapping work done in
VersificationMgr for use in the VMRCRE and had to fully understand the
scheme and syntax used in the good work done by Костя on this a few
years back.
It was a bit difficult for me to keep each component straight in my mind
initially so during my port, I created a class with clear property names
to hold each value, and an import format which makes it clear what
values do what. Here are the lines of code for reference:
This is the VMRCRE equivelant to a canon.h file (MTNU: the Masoretic OT
+ Nestle/Aland NT) and has mappings between the MTNU and the LXXNU (LXX
= Göttingen/Rahlfs) versifications under the <mappings> section near the
end:
https://crosswire.org/svn/community/trunk/contrib/v11ns/MTNU.xml
Port of VersificationMgr with class to hold the mappings section:
https://git.crosswire.org/main/crosswire-java/-/blob/master/src/org/crosswire/sword/mgr/VersificationMgr.java#L233
Hope this is helpful,
Troy
On 1/24/25 3:11 PM, DM Smith wrote:
On Jan 23, 2025, at 6:40 PM, Arnaud Vié <unas.zole+a...@gmail.com> wrote:
Le jeu. 23 janv. 2025 à 22:18, Fr Cyrille
<fr.cyri...@tiberiade.be> a écrit :
So I think that for versifications, catholic catholic2 and
LXX there is no mapping (which would explain some of the
problems I've had displaying in parallel?).
In the other v11n files I see an entry by instance in vulg:
unsigned char mappings_vulg[] = {
If someone can explain or give me a documentation to
understand how it works, and if this is where the mapping
takes place, I'd like to work on it. Especially as it should
be possible to use Arnaud's work?
You're correct Cyrille, the mapping data is in this char array
defined with the versifications in the canon_*.h files.
This array is injected (for the versifications which have one) and
decoded within versificationmgr.cpp.
The format of this char array is really obscure though, and it's
completely different from the format of mappings used in jsword
(which is a lot clearer and a lot easier to edit and maintain).
From what I understand in the code ("// parse mappings" section of
the loadFromSBook method in versificationmgr.cpp) and
the canon_vulg.h example :
SWORD is built for speed. The format is a C string (characters
followed by a null). SWORD’s format is meant to be obscure to force
the use of SWORD and JSword as the mechanism to read SWORD modules.
JSword is meant to be in lock step with the formats of SWORD. WRT the
mapping files, these were developed independently and at the same
time. I deemed it too hard at the time to bring JSword back into
alignment w SWORD’s format. JSword’s format is human readable,
external, easily maintainable and is read when needed.
1.
The array starts with a sequence of book names, which are the books
present in this bible and absent from KJV.
Each such name is a sequence of letters followed by a null character
serving as delimiter.
In canon_vulg, thoses are the
'E', 'p', 'J', 'e', 'r', 0,
'P', 'r', 'A', 'z', 'a', 'r', 0,
'S', 'u', 's', 0,
'B', 'e', 'l', 0,
2. Then, you have an additional null character indicating the end of
that first section - the rest of the array is encoded completely
differently.
3. The rest of the array is meant to be split in sequences of 7
numbers, each such 7-number sequence corresponding to a mapping rule.
For example, this is a mapping rule :
21, 4, 9, 10, 4, 8, 0,
First digit indicates the book (index in the sequence of books,
starting at 1). In vulg, book 21 is Psalms.
The next 3 digits indicate the destination of the mapping. "4,9,10"
corresponds to chapter 4, verses 9 to 10.
The final 3 digits are the source of the mapping. "4,8,0" corresponds
to chapter 4, verse 8 alone.
So this rule I used as example denotes that verses Ps 4:9-10 of KJV
are mapped to Ps 4:8 in Vulg.
Which corresponds indeed to one rule that is present in the jsword
mapping file.
I guess I could try to build a converter to export the jsword mapping
properties files into this format, so that we could indeed add the
Catholic and Catholic2 mappings that I spent hours building for the
AndBible jsword fork.
That would be greatly appreciated!
But going forward, for maintaining versifications, it would be a lot
better to have a central way of defining all our versifications and
all their mappings in an easily readable and editable format (maybe
the one defined by the Copenhagen Alliance, cf
https://github.com/Copenhagen-Alliance/versification-specification/blob/master/versification-mappings/json-schema/versification_schema.json
), serving as a source of truth for both sword and jsword. Because
even just looking at this Vulg versifications, the mapping have
vastly diverged between sword and jsword...
IIRC, the mapping of deuterocanonical material I don’t think was ever
completed.
Even if we don't go as far as implementing the full modular
versification system that I would like to build, if at the very least
we could have a central, easy-to-maintain place to manage all our
versifications, and then could easily export them to both sword and
jsword, that would be a huge step forward.
I think Chris Little developed an external format that would build the
canon_*.h files. I don’t remember if there was something like this
that was developed for mappings.
Regards,
Arnaud
_______________________________________________
sword-devel mailing list: sword-devel@crosswire.org
http://crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
_______________________________________________
sword-devel mailing list:sword-devel@crosswire.org
http://crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
_______________________________________________
sword-devel mailing list: sword-devel@crosswire.org
http://crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
