| Time |
S |
Nick |
Message |
| 14:38 |
|
nahuel |
gmcharlt, hi |
| 14:38 |
|
nahuel |
gmcharlt, around ? |
| 14:42 |
|
gmcharlt |
hi nahuel |
| 14:50 |
|
nahuel |
gmcharlt, what's up ? |
| 14:50 |
|
gmcharlt |
not much |
| 14:50 |
|
gmcharlt |
working on a Unicode bug |
| 14:50 |
|
gmcharlt |
(and I truly hate Perl's Unicode support) |
| 14:50 |
|
nahuel |
;) |
| 14:51 |
|
nahuel |
gmcharlt, are you a zebra killer ? |
| 14:51 |
|
gmcharlt |
testing a change to use an XSL stylesheet to format the metdata for the oai service |
| 14:51 |
|
gmcharlt |
turns out that if the stylesheet omits the XML declaration |
| 14:51 |
|
gmcharlt |
XML::LibXSLT won't recognize if the output is UTF-8 |
| 14:51 |
|
nahuel |
and with binmode utf8 ? |
| 14:52 |
|
gmcharlt |
meaning that you have to explicitly turn on the output string's utf8 flag |
| 14:52 |
|
gmcharlt |
yeah, binmode utf8 on STDOUT doesn't help |
| 14:54 |
|
gmcharlt |
so what's your zebra question? |
| 14:59 |
|
nahuel |
gmcharlt, it's an authorities search problem |
| 14:59 |
|
nahuel |
"somethine like (this)" doesn't find anything |
| 14:59 |
|
nahuel |
but "something like this" find something |
| 15:00 |
|
nahuel |
the strange stuff is that in the word-phrase-utf.chr there are the () interpreted as "space" |
| 15:00 |
|
nahuel |
I don't understand wtf |
| 15:00 |
|
gmcharlt |
ok, setting example in my test database |
| 15:01 |
|
nahuel |
does the 'autority' database uses the default.idx that point to word-phrase-utf.chr ? |
| 15:03 |
|
gmcharlt |
yes |
| 15:03 |
|
nahuel |
so I really don't understand what's going on |
| 15:08 |
|
gmcharlt |
nahuel: ok, I just created an authority records in my test DB with a 150$aFoobar barbaz (quux) |
| 15:08 |
|
gmcharlt |
a "topical term" "contains" search of "foobar barbarz (quux)" does work for me |
| 15:08 |
|
gmcharlt |
how are you doing the search, exactly? |
| 15:09 |
|
gmcharlt |
one big difference is that my test DB is using DOM indexing for authorities, not GRS-1 |
| 15:09 |
|
nahuel |
/cgi-bin/koha/opac-search.pl?q=su:Eglise 30-600 (Eglise primitive) |
| 15:09 |
|
nahuel |
ah |
| 15:09 |
|
nahuel |
i'm using grs-1 |
| 15:09 |
|
nahuel |
arf |
| 15:10 |
|
nahuel |
/cgi-bin/koha/opac-search.pl?q=su:"Eglise 30-600 (Eglise primitive)" |
| 15:10 |
|
gmcharlt |
ah, but it looks like you're doing a subject *bib* search, not an autohrity search |
| 15:10 |
|
gmcharlt |
lemme check |
| 15:10 |
|
nahuel |
subjects are autorities, no ? |
| 15:11 |
|
gmcharlt |
subjects can be controlled by authority records, yes |
| 15:11 |
|
gmcharlt |
but there's a difference between searching on a bib index |
| 15:12 |
|
gmcharlt |
and search on an authority index |
| 15:12 |
|
nahuel |
ho, and su: is a bib search ? |
| 15:12 |
|
gmcharlt |
yes |
| 15:12 |
|
nahuel |
aaaah, ok |
| 15:13 |
|
gmcharlt |
just a sec, I'm trying this out in my test DB |
| 15:13 |
|
nahuel |
ok |
| 15:20 |
|
gmcharlt |
nahuel: 650 $aBarbaz tree (quux) |
| 15:20 |
|
gmcharlt |
both /cgi-bin/koha/opac-search.pl?q=su:"barbaz tree (quux)" |
| 15:20 |
|
gmcharlt |
and /cgi-bin/koha/opac-search.pl?q=su:"barbaz tree quux" work for me |
| 15:21 |
|
gmcharlt |
what does the heading look like in your bib record? |
| 15:21 |
|
nahuel |
not for me on a new 3.0.1 install :s |
| 15:21 |
|
nahuel |
2s |
| 15:24 |
|
nahuel |
gmcharlt, http://biblibre.pastebin.com/m40681189 |
| 15:28 |
|
gmcharlt |
could you paste in the Zebra logs for the two searches? |
| 15:29 |
|
gmcharlt |
based on your record, was the search supposed to be "Asc?tisme 30-600 (Eglise primitve) ? |
| 15:33 |
|
nahuel |
no |
| 15:33 |
|
nahuel |
scuse |
| 15:33 |
|
nahuel |
i pasted a bad record |
| 15:35 |
|
nahuel |
gmcharlt, http://biblibre.pastebin.com/m7875867e |
| 15:36 |
|
nahuel |
606$a and 606$z are concatened |
| 15:52 |
|
frederic |
gmcharlt: What kind of work are you doing on Koha OAI server? |
| 15:52 |
|
gmcharlt |
frederic: some bugfixes and performance improvements |
| 15:53 |
|
gmcharlt |
but biggest is adding ability to use XSL transform of the MARCXML to generate metadata instead of hardcoding |
| 15:53 |
|
frederic |
I'm doing an OAI server and I'm facing several issues. |
| 15:53 |
|
frederic |
yes |
| 15:53 |
|
frederic |
http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=2990 |
| 15:53 |
|
gmcharlt |
and will add support for returning MARCXML, not just DC |
| 15:53 |
|
frederic |
It seems OAI server has been done for Koha v2. |
| 15:54 |
|
frederic |
I also would like to be able to implement metadata support for schema other than Dublin Core. |
| 15:54 |
|
frederic |
Koha OAI server should be modified to be extensible. An XSLT style sheet could be a solution to return any kind of format... |
| 15:54 |
|
frederic |
I'm about to do it but would wait something will be available on your side... |
| 16:01 |
|
gmcharlt |
frederic: I have a proof of concept set of patches, but they're not particularly clean yet |
| 16:02 |
|
gmcharlt |
if you have more time than I at the moment, I can send you the patches and you can look at |
| 16:02 |
|
gmcharlt |
and use as a basis or ignore as you see fit |
| 16:08 |
|
frederic |
gmcharlt: I need to find a solution quickly. So will I have time. For now, OAI server is unusable: no subjects, no items, and un-extensible. |
| 16:08 |
|
frederic |
Please send me your patches. I will try them and see how they deal with my need. Thanks. |
| 16:08 |
|
gmcharlt |
just occurred to me - do you have a style sheet to do UNIMARC to DC? |
| 16:09 |
|
frederic |
no yet. Have to write it. |
| 16:10 |
|
frederic |
I have both MARC21 and UNIMARC DBs. |
| 16:10 |
|
frederic |
But my need is yet for UNIMARC |
| 16:49 |
|
gmcharlt |
frederic: sent you OAI patch |
| 16:49 |
|
gmcharlt |
as I said, very messy, so use for inspiration |
| 16:50 |
|
gmcharlt |
IMO, the current design of C4::OAI::* is suited to a situation where you're cobbling together the DC from a relational database or the like |
| 16:51 |
|
gmcharlt |
but XSLT makes much more sense if you're starting from a metadata format like MARC21 or UNIMARC that is serialized as XML |
| 00:03 |
|
headbang |
hi, any active users on libki around? |
| 00:28 |
|
gmcharlt |
headbang: dunno. not me, at any rate. you may have better luck on #code4lib on freenode |
| 00:28 |
|
headbang |
ah tnx for the possible right hint! :) |
