Re: [edict-jmdict] Re: code to parse/format edict format

["Jim Breen" <jimbreen@*********>]

On 24/05/07, wmaton <wmaton@yahoo.com> wrote:
> --- In edict-jmdict@yahoogroups.com, "Stuart McGraw" <smcg4191@...> wrote:
>  > Does anyone have or know where I can get
>  > Perl (or other language) code that will parse
>  > or format edict-style text?
>
>  There's a C language example in Jeffrey Friedl's lookup program, which
>  serves as the backend of the J-E server I run.  You can find it at any
>  Nihongo mirror near you, or in my unofficial patch collection page at:
>
>  http://www.wfms.org/lookup/
>
>  I do remember that there was a list version someplace, but that was
>  eons ago.
>
>  Also, Jim Breen's xjdict (c'mon Jim, you know it still works....!)
>  is another example in C.

The xjdic tarball is on the Monash ftp site. Like WWWJDIC it doesn't do much
meaningful parsing of the EDICT entry - just fiddles a few things to make the
output look better, e.g. / converts to ;, etc. The indexer parses the entry
but only looks for text strings or whitespace.

Probably the most useful parser I have is the one that converts EDICT-format
entries into my internal database. It takes something like:

KKK1;KKK2 [rrrr1;rrrr2] /(n,vs) (obsc) (see KKKK3) blah/

And turns it into

#E nnnnnnn
KKK1
KKK2
#R
rrr1
rrr2
#AU 2007-05-06 Entry created
#M
#G n vs
#X KKK3
#MI obsc
blah

I have put the (totally undocumented) source of this at:
http://www.csse.monash.edu.au/~jwb/buildzip.zip

That .zip also has a file called "tags" which is used to
sort out the n,vs,obsc, etc. situations. It's not tested
yet for the new "#SL" markup which generates the <lsource...>

Undocumented "features" of the utility are:
- it only does tags, xrefs, etc. properly (more or less) for
the first sense. I have to fix things by hand later for multi-sense
conversions;
- for multiple senses, it needs the later senses to be flagged by a "#".
When converting a batch I just use an editor to change:

.... /(1) senses1 /(2) sense 2/

to

.... /sense 1/#sense 2/

I'm sure there are other things I've overlooked.

Hope this helps

Jim

--
Jim Breen
Honorary Senior Research Fellow
Clayton School of Information Technology,
Monash University, VIC 3800, Australia
http://www.csse.monash.edu.au/~jwb/