Speed up bib search and OPAC bib display, especially
when the XSLT OPAC results and details display sysprefs are
ON, by passing an existing MARC::Record object to three
functions:
C4::Biblio::get_biblio_authorised_values()
C4::XSLT::XSLTParse4Display()
C4::XSLT::transformMARCXML4XSLT (internal)
These functions previously fetched the bib from the
database, incurring the cost of DB retrieval and MARCXML
parsing even though client code already had a
MARC::Record object available.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
GetMarcFromKohaField returns (0,0) without the $frameworkcode argument.
Note the argument can be "", but it must not be undefined.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Following suggestion by Vincent Danjean for Debian
packaging, 0755 -> 0644 for non-executable
files.
Also removed shebang from a few modules in C4.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
This adds a new function, getitemtypeimagelocation, that returns the image
unmodified for absolute urls and returns the proper intranet or opac path
otherwise. It also updates all of the relevant files to use that function.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
C4::Search::SimpleSearch was alredy patched to let you pass in the number of results you want back.
These instances were not using the new API. This patch makes all calls to SimpleSearch specify a limit.
I improved the documentation of SimpleSearch a bit to include the third returned value.
I believe there's a bug in C4::Output::pagination_bar, in that it doesn't deal well with URLs
with only one pair of parameter=value passed to it. I'm getting around this by passing in a second
pair that does nothing.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
The in transit status now displays as such in the
OPAC search results. In the non-XSLT version, the status
is checked only for bibs having 20 or fewer items to avoid
extra hits on the database during result set presentation.
In the XSLT version, all items are checked.
Note that because an item's transfer status is not
stored in the MARC record, the transfer status
has no effect when limiting a search by item
availability. For a future version, the transit status
should be added to the Zebra index.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
If a search contains a digit, it is almost certainly an identifier, and should
not be stemmed. This is particularly relevant for ISBNs and ISSNs, which
can contain the letter "X" - for example, C4::Search::_build_stemmend_operand
would reduce "014100018X" to "x ", which for a MARC21 database would bring
up irrelevant results (e.g., "23 x 29 cm." from the 300$c).
With this patch, supplying a search term that contains one or more digits
followed by "X" will no longer retrieve irrelevant results. This applies
to catalogs using Zebra and query stemming.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Explicitly destroy ZOOM Query and ResultSet objects created
in C4::Search::SimpleSearch() - for long-running users
of SimpleSearch, such as link_bibs_to_authorities.pl or
the matcher used by the staging MARC import, the failure
to do this causes a memory leak in both the client Perl
script and (in the case of ZOOM ResultSet objects) a
corresponding resource leak in zebrasrv for the life
of the Z39.50 connection.
With this change, link_bibs_to_authorities.pl will be
be able to process large bib datasets without leaking
memory due to the Z39.50 lookups it does.
Similar changes are indicated for all uses of
ZOOM that could last longer than a single CGI
query.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
This patch allows to translate facets label with standart
templates / PO files. Facets are still hard coded in Koha.pm.
Template wait those hard coded facets: Topics, Places, Titles,
Autors and Libraries.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Corrected bug that prevented the 'not' operator
from working consistently - i.e., a search of
'mice not men' would not always work.
Also added test cases for NoZebra, so far focusing on
NZanalyse and adding and deleting bibs. A couple of
the test cases are currently known to fail and
therefore are marked TODO. The tests in question
are to verify that rows in nozebra are removed if
no bib is linked to the relevant word. However, it
looks like such rows are retained, just with
nozebra.biblionubmers set to ''. Is there any
reason to keep these rows?
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
When running Koha in NoZebra mode under Perl 5.10, a search
containing a operator (e.g., "mice and men") could cause
a bib search to enter into an infinite loop in NZanalyse.
This possibility arises from the fact that NZanalyse used
to use capture variables from regular expressions without
verifying whether the regular expressions actually matched.
It was compounded by the fact that NZanalyse is recursive;
however, because $1, $2, etc. are dynamically scoped,
*they are not automatically cleared when NZanalyse calls
itself*.
Consequently, if the search string contains a boolean
operator, it would be split into
left = mice
operator = and
right = men
Then NZanalyse would be called recursively on the search
string 'mice'. However, because $1, $2, and $3 are not
automatically cleared when the function is called again,
and because they are not cleared if a match fails, the code
would fail to recognize that 'mice' is leaf, and would
call NZanalyse('mice') repeatedly, to the promotion of
warm server rooms.
The wrinkle in this is that because of a bug in Perl 5.8, a
failing matches can sometimes alter the capture variables, thus
avoiding the infinite recursion. However, this bug was fixed in
Perl 5.10, leading to the NZanalyse bug becoming evident.
The Perl bug is described at http://rt.perl.org/rt3/Public/Bug/Display.html?id=19049
and the fix http://public.activestate.com/cgi-bin/perlbrowse/p/29279.
The fix to the Koha code is to check whether each regexp
that uses capture variables matches or fails, then act
accordingly.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Including the subtitle in search results and OPAC and staff
details templates. The subtitle is expected to be in the
MARC subfield that is mapped to the (now nonexistant) column
bibliosubtitle.subtitle.
This is a fairly hackish patch that is meant to be minimally
disruptive while allowing subtitles to be displayed for those
not using the experimental XSLT bib display feature. It
does *not* cover all cases where you have a MARC
record and want to display its subtitle.
UNIMARC notes: at the moment, none of the fr-FR UNIMARC
frameworks appear to be map any subfields to bibliosubtitle.subtitle.
This could be changed, but caveat programmer: the UNIMARC 200$e
is repeatable (meaning that get_koha_field_from_marc will put in ' | '
between multiple subtitles) and one may want to also include the
200$d (parallel title). A GetMarcSubtitle routine may be in order
if absolutely necessary.
This points to a long-term need to make display of MARC (and later, hopefully,
non-MARC) bib records more flexible - the XSLT feature is one approach;
there may be others.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
I've refactored the subs I added in the previous commit so that they make a little
more sense and are in better places in the code base. I was really hoping to make use
of existing subs, but they all seemed so specific to particular uses.
The icons now show up on the OPAC item details page.
TODO: The icons still don't show up in the OPAC search results page.
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
This is an improvement of a patch that has previously been rejected: I wanted to display holdingbranch, kados prefer homebranch
(and in fact it depends on the library choice...)
now, the library can choose, thanks to HomeOrHoldingBranch syspref.
Note that I have changed the description of the syspref to reflect this new use
Signed-off-by: Chris Cormack <crc@liblime.com>
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>
Disable or Enable the display of query terms in the biblio.
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
This patch is meant to guarantee that a bulkedit
does not try to edit an item tag embedded in a MARC
biblio without updating the items feature. It is
not a comprehensive fix of the bulkedit feature, which
currently does not appear to be functional and
needs some thought:
* The general search results is probably not the
best place to put this feature -- it should
probably be in tools.
* A bulk edit of something like items is desireable,
but needs to be designed so that it respects
business logic for circulation and acquisitions.
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
Prior to this fix, the status fields had three 'off' values, NULL, "",
and 0. I've reduced it to two in the db, removing the option for NULL, and
setting the default value to 0, however, we need to verify that we don't ever
write out as "" as this needlessly complicates the indexing process,
critical for searching or limiting by status (e.g., availability). Also,
queries that attempt to write a NULL value to one of these fields will fail
(based on my tests).
This patch includes the following changes:
* Updated the database definition for notforloan, damaged, itemlost, and
wthdrawn in kohastructure.sql to forbid NULL and default to 0; MySQL
can't forbid other values (such as empty ""), so this has to be handled
at the application layer and REQUIRES further patching.
* Fixed the 'limit by availability' query node in Search.pm to use a
much less confusing definition of 'available'
* Added code to set values to 0 where they are NULL or empty ( "" ) for
notforloan, damaged, itemlost or wthdrawn in both the MARC and the items
table:
* Biblio.pm -> AddBiblioAndItems
* catalogue/updateitem.pl
* SEE NOTE BELOW, REQUIRES UPDATE TO THE REST OF KOHA'S ITEM MGT!
* Removed code in bulkmarcimport.pl that sets notforloan status depending
on item-level or bib-level itemtype -- that flag is designed to be set
only to override the notforloan setting for the item's (or bib's,
depending on the syspref) assigned itemtype (it doesn't need to override
to 'for loan', only to 'not for loan').
added $dbh->do("truncate zebraqueue"); when operation is 'delete'
* I updated some notes in catalogue/updateitem.pl as to why ModItem can't be
used -- we don't have _a_ place where we can change the item and marc :/
I've tested the following:
bulkmarcimport.pl..........................MARC/items OK
Staged Records Import......................NOT OK
updateitem.pl (via moredetail.pl)..........MARC/items OK
circulation.pl.............................NOT OK
returns.pl.................................NOT OK
addbiblio.pl...............................NOT OK
additem.pl.................................NOT OK
Basically, there isn't a single place to apply this patch that will
update both item data and MARC data in one place ... a future patch
needs to address this issue.
Signed-off-by: Galen Charlton <galen.charlton@liblime.com>
Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>
