You've been warned :-). This patch contains a more
complete mapping of UTF-8 to ASCII. The mappings are
based on those compiled by Richard Mahoney on the
Zebra list: http://lists.indexdata.dk/pipermail/zebralist/2007-August/001707.html
Note to documentation team: we need an area in the
documentation that discusses how Koha handles searches
and indexing for words that contain diacritics, such
as E-ACUTE (vs E without an acute). If you can paste
this list of mappings from this patch directly into
the docs and it preserves the encoding that would be
great.
NOTE: I don't think this patch addresses issues of
combining vs non-combining forms, and may require
a refactor to address that.
Josh
The problem was that the 'mc-' was removed from the checkboxes a while back and
that's what triggers the automatic application of OR boolean searching. I've
added it back to the templates and modified the ccl.properties file to include
mapping for itype,itemtype and ccode
Note: currently only zebraqueue_daemon.pl is known
to use the extended services that require the
Zebra r/w password.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
The following error messages in the Zebra
log should no longer appear:
06:10:25-04/03 zebrasrv(1) [warn] Failed to read character table urx.chr
06:10:25-04/03 zebrasrv(1) [warn] urx.chr [No such file or directory]
To fully install this patch, do a
'make update_zebra_conf'.
Signed-off-by: Chris Cormack <chris@bigballofwax.co.nz>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
To fully install this patch, the following steps
are required:
1. perl Makefile.PL
2. make
3. make update_zebra_conf
4. restart zebrasvr
5. reindex authorities using rebuild_zebra.pl -a -r
Signed-off-by: Chris Cormack <chris@bigballofwax.co.nz>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
To fully install this patch, the following steps are
necessary:
1. perl Makefile.PL
2. make
3. make update_zebra_conf (or make upgrade)
Signed-off-by: Chris Cormack <chris@bigballofwax.co.nz>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
The approach is to use PazPar2 to search just one
target, the biblio Zebra database. The results
of each set are merged by PazPar2 to generate a
hitlist that combines related bibs together; as an
example, if a library has the first Harry Potter
book in three languages and an audiobook format,
the hitlist should ideally return one result
for the work that includes links to the individual
bibs.
The new module C4::Search::PazPar2 implements a
simple client for PazPar2's XML-over-HTTP API. It is
designed to be generic, and thus may end up getting
moved out of Koha to become a stand-alone CPAN module.
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
For DOM indexing, added index_matching_heading option
to create indexes for matching an entire authority
heading -- the index works by indexing a heading
such
150 $aCars$xElectric$zEngland$vScience fiction
as something like
"cars generalsubdiv electric geographicsubdiv england
formsubdiv science fiction"
Also started adjust names of some indexes to conform
to languaged used in the MARC21 and UNIMARC standards, e.g.,
"See" => "See-from"
"See-also" => "See-also-from"
"Conference-name-heading" => "Meeting-name-heading"
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
All new authority features will be based on the DOM indexing.
To update an existing installation, do the following:
[1] run perl Makefile.PL
[2] make
[3] make update_zebra_conf
[4] copy the new koha-conf.xml to $KOHA_CONF
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Currently, MARC authorities are indexed (assuming Zebra
is used) with Zebra's GRS-1 module. However, it does
not appear to be possible to index phrases that cross
subfield boundaries using the GRS-1 module's records.abs
config file's melm, elm, and xelm directives.
Since it is necessary to be able to efficiently search
an entire authority heading (e.g., to see if a given
bib heading is authorized), I'm proposing a switch
to Zebra's DOM XML filter module, which uses XSLT
to generate the words and phrases to be indexed from the
original MARC XML (or ISO2709) record.
The file authority-zebra-indexdefs.xml is an XSLT stylesheet
to implement the new indexing regime. It is based on the
MARC21 authority record.abs with the following changes:
* addition of 148/448/548
* changed name of "see" indexes to "see-from"
* changed name of "see-also" indexes to "see-also-from"
* added index on the subject thesaurus based on
the 008/11 and 040$f
* added indexes on the full heading
authority-zebra-indexdefs.xml was generated from
authority-koha-indexdefs.xml via the XSL transform
koha-indexdefs-to-zebra.xsl. authority-koha-indexdefs.xml
is the actual master version of the indexing definitions,
and was created to provide a much more compact syntax
over the raw XSLT that is to be passed to Zebra.
An experimental schema for Koha indexing definitions is
under way; my aim is to propose a simple format that can
be readily worked with, and perhaps even generated as
a serialization of indexing definitions that are set up
via administration settings in the Koha database itself.
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
* Defined local field 942$a to store the authority type
for MARC21 instead of 152$b
* Added 942$b to MARC21 authority framework.
* Added auth_header.authid and auth_header.authtypecode
to appropriate subfields in MARC21 authority framework.
* Started work on two new modules:
C4::AuthoritiesMarc::MARC21
C4::AuthoritiesMarc::UNIMARC
These modules will be used to extract MARC-format-specific
behavior out of C4::AuthoritiesMarc
* Updated Zebra config for MARC21 to use only the 942$a
for the authority type.
* For MARC21, added logic to move 152$b to 942$a for
existing authority records. Specifically, AddAuthority
now does this move when a record is saved, while
GetAuthority and GetAuthorityXML do this when
extracting a record for other use. This logic
is temporary, and can hopefully be removed later, once
use of 152$b in MARC21 authorities is confirmed to be
absent for Koha users. I will also create a batch
job to do this update in one fell swoop.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Needed to restore OpenSearch capabilities, and did the following while
I was at it:
* add support for unAPI: http://unapi.info/
* add basic support for COinS and OpenURL:
http://ocoins.info;
http://www.niso.org/committees/committee_ax.html
* ^^ Gives us Zotero Support!
* adding some XSLT stylesheets for handling additional transformations
NOTE: English and MARC21 specific unfortunately
* adding back opensearch/rss feed <link>s for autodiscovery
TODO: after the installation, to get the Zebra system running on an external
port it's necessary to hand-edit the configs. I'm looking into Virtual Hosts
which could solve that problem (run on both the socket and a port).
Need to add better error handling to the unapi and opensearch scripts
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
This commit is a partial fix to authority control for MARC21,
and better authority handling in general (for UNIMARC too).
Before this patch, authority searching, editing, saving, was
not functioning, or was extremely buggy.
WARNING: You will need to re-index your authority database after
applying this commit.
The following changes have been made:
* Normalizing record.abs index names (in both MARC21 and UNIMARC)
* Synching authorities/bib1.att, ccl.properties, AuthoritiesMarc.pm
with new indexes (UNIMARC too)
* Clean up biblios/bib1.att (remove duplicate att defs)
* Clean up authorities-* templates to conform to new styles
* Fixed search failure when using Default framework (now searches
All)
Also included are several fixes to the built-in SRU server for
Authority and Biblio, it's recommended that you update your
koha-conf.xml file:
* adding explain-authorities.xml and explain-biblios.xml
* adding necessary info to koha-conf.xml to enable SRU/W
* adding several example XSLT stylesheets, that can be used
for SRU on-the-fly transformations (to MODS, DC, RDF, etc.)
Still remaining for 3.0 are the following tasks:
* update MARC21 frameworks (authority and cross-reference bib)
* update display code/templates in authority results list
* update search code/templates to utilize index points
* implement 'grouping' of authtypes for searching (Name, Title, Subject)
* repair utility to import auths and perform matching
* repair bibliographic import to match auths and warn if no match
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
This parameter, initialized from LOG_DIR during installation,
allows scripts to specify a common directory for logs.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
* prior to this commit, virtual shelvesn't did not function in
the OPAC! Now they do, except for deletion from virtual shelves
in list form
* I've re-named 'Virtual Shelves' to 'Lists' as per our agreed
upon convention
* while vshelves aren't perfect yet, they're in enough of a working
state for the RC1 now
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
