1 <?xml version="1.0" standalone="no"?>
2 <!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook V4.4//EN"
3 "http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd"
5 <!ENTITY % local SYSTEM "local.ent">
7 <!ENTITY % entities SYSTEM "entities.ent">
9 <!ENTITY % idcommon SYSTEM "common/common.ent">
12 <refentry id="pazpar2_conf">
14 <productname>Pazpar2</productname>
15 <productnumber>&version;</productnumber>
16 <info><orgname>Index Data</orgname></info>
20 <refentrytitle>Pazpar2 conf</refentrytitle>
21 <manvolnum>5</manvolnum>
22 <refmiscinfo class="manual">File formats and conventions</refmiscinfo>
26 <refname>pazpar2_conf</refname>
27 <refpurpose>Pazpar2 Configuration</refpurpose>
32 <command>pazpar2.conf</command>
37 <title>DESCRIPTION</title>
39 The Pazpar2 configuration file, together with any referenced XSLT files,
40 govern Pazpar2's behavior as a client, and control the normalization and
41 extraction of data elements from incoming result records, for the
42 purposes of merging, sorting, facet analysis, and display.
46 The file is specified using the option -f on the Pazpar2 command line.
47 There is not presently a way to reload the configuration file without
48 restarting Pazpar2, although this will most likely be added some time
56 The configuration file is XML-structured. It must be well-formed XML. All
57 elements specific to Pazpar2 should belong to the namespace
58 <literal>http://www.indexdata.com/pazpar2/1.0</literal>
59 (this is assumed in the
60 following examples). The root element is named "<literal>pazpar2</literal>".
61 Under the root element are a number of elements which group categories of
62 information. The categories are described below.
65 <refsect2 id="config-threads">
66 <title>threads</title>
68 This section is optional and is supported for Pazpar2 version 1.3.1 and
69 later . It is identified by element "<literal>threads</literal>" which
70 may include one attribute "<literal>number</literal>" which specifies
71 the number of worker-threads that the Pazpar2 instance is to use.
72 A value of 0 (zero) disables worker-threads (all work is carried out
76 <refsect2 id="config-file">
79 This configuration takes one attribute <literal>path</literal> which
80 specifies a path to search for local files, such as XSLTs and settings.
81 The path is a colon separated list of directories. Its default value
82 is "<literal>.</literal>" which is equivalent to the location of the
83 main configuration file (where indeed the file element is given).
86 <refsect2 id="config-server">
89 This section governs overall behavior of a server endpoint. It is identified
90 by the element "server" which takes an optional attribute, "id", which
91 identifies this particular Pazpar2 server. Any string value for "id"
96 elements are described below. From Pazpar2 version 1.2 this is
99 <variablelist> <!-- level 1 -->
104 Configures the webservice -- this controls how you can connect
105 to Pazpar2 from your browser or server-side code. The
106 attributes 'host' and 'port' control the binding of the
107 server. The 'host' attribute can be used to bind the server to
108 a secondary IP address of your system, enabling you to run
109 Pazpar2 on port 80 alongside a conventional web server. You
110 can override this setting on the command line using the option -h.
119 If this item is given, Pazpar2 will forward all incoming HTTP
120 requests that do not contain the filename 'search.pz2' to the
121 host and port specified using the 'host' and 'port'
122 attributes. The 'myurl' attribute is required, and should provide
123 the base URL of the server. Generally, the HTTP URL for the host
124 specified in the 'listen' parameter. This functionality is
125 crucial if you wish to use
126 Pazpar2 in conjunction with browser-based code (JS, Flash,
127 applets, etc.) which operates in a security sandbox. Such code
128 can only connect to the same server from which the enclosing
129 HTML page originated. Pazpar2s proxy functionality enables you
130 to host all of the main pages (plus images, CSS, etc) of your
131 application on a conventional webserver, while efficiently
132 processing webservice requests for metasearch status, results,
139 <term>icu_chain</term>
142 Specifies character set normalization for relevancy / sorting /
143 mergekey and facets - for the server. These definitions serves as
144 default for services that don't have these given. For the meaning
145 of these settings refer to the
146 <xref linkend="icuchain"/> element inside service.
152 <term>relevance / sort / mergekey / facet</term>
155 Obsolete. Use element icu_chain instead.
161 <term>settings</term>
164 Specifies target settings for the server.. These settings serves
165 as default for all services which don't have these given.
166 The settings element requires one attribute 'src' which specifies
167 a settings file or a directory . If a directory is given all
168 files with suffix <filename>.xml</filename> is read from this
170 <xref linkend="target_settings"/> for more information.
176 <term id="service_conf">service</term>
179 This nested element controls the behavior of Pazpar2 with
180 respect to your data model. In Pazpar2, incoming records are
181 normalized, using XSLT, into an internal representation.
182 The 'service' section controls the further processing and
183 extraction of data from the internal representation, primarily
184 through the 'metadata' sub-element.
187 Pazpar2 version 1.2 and later allows multiple service elements.
188 Multiple services must be given a unique ID by specifying
189 attribute <literal>id</literal>.
190 A single service may be unnamed (service ID omitted). The
191 service ID is referred to in the
192 <link linkend="command-init"><literal>init</literal></link> webservice
193 command's <literal>service</literal> parameter.
196 <variablelist> <!-- Level 2 -->
198 <term>metadata</term>
201 One of these elements is required for every data element in
202 the internal representation of the record (see
203 <xref linkend="data_model"/>. It governs
204 subsequent processing as pertains to sorting, relevance
205 ranking, merging, and display of data elements. It supports
206 the following attributes:
209 <variablelist> <!-- level 3 -->
214 This is the name of the data element. It is matched
215 against the 'type' attribute of the
217 in the normalized record. A warning is produced if
218 metadata elements with an unknown name are
220 normalized record. This name is also used to
222 data elements in the records returned by the
223 webservice API, and to name sort lists and browse
233 The type of data element. This value governs any
234 normalization or special processing that might take
235 place on an element. Possible values are 'generic'
236 (basic string), 'year' (a range is computed if
237 multiple years are found in the record). Note: This
238 list is likely to increase in the future.
247 If this is set to 'yes', then the data element is
248 includes in brief records in the webservice API. Note
249 that this only makes sense for metadata elements that
250 are merged (see below). The default value is 'no'.
259 Specifies that this data element is to be used for
260 sorting. The possible values are 'numeric' (numeric
261 value), 'skiparticle' (string; skip common, leading
262 articles), and 'no' (no sorting). The default value is
266 When 'skiparticle' is used, some common articles from the
267 English and German languages are ignored. At present the
268 list is: 'the', 'den', 'der', 'die', 'des', 'an', 'a'.
274 <term id="metadata-rank">rank</term>
277 Specifies that this element is to be used to
279 records against the user's query (when ranking is
281 The valus is of the form
285 where M is an integer, used as a
286 weight against the basic TF*IDF score. A value of
287 1 is the base, higher values give additional weight to
288 elements of this type. The default is '0', which
289 excludes this element from the rank calculation.
292 F is a CCL field and N is the multipler for terms
293 that matches those part of the CCL field in search.
294 The F+N combo allows the system to use a different
295 multipler for a certain field. For example, a rank value of
296 "<literal>1 au 3</literal>" gives a multipler of 3 for
297 all terms part of the au(thor) terms and 1 for everything else.
300 For Pazpar2 1.6.13 and later, the rank may also defined
301 "per-document", by the normalization stylesheet.
304 The per field rank was introduced in Pazpar2 1.6.15. Earlier
305 releases only allowed a rank value M (simple integer).
307 See <xref linkend="relevance_ranking"/> for more
313 <term>termlist</term>
316 Specifies that this element is to be used as a
317 termlist, or browse facet. Values are tabulated from
318 incoming records, and a highscore of values (with
319 their associated frequency) is made available to the
320 client through the webservice API.
322 are 'yes' and 'no' (default).
331 This governs whether, and how elements are extracted
332 from individual records and merged into cluster
333 records. The possible values are: 'unique' (include
334 all unique elements), 'longest' (include only the
335 longest element (strlen), 'range' (calculate a range
336 of values across all matching records), 'all' (include
337 all elements), or 'no' (don't merge; this is the
341 Pazpar 1.6.24 also offers a new value for merge, 'first', which
342 is like 'all' but only takes all from first database that returns
343 the particular metadata field.
349 <term>mergekey</term>
352 If set to '<literal>required</literal>', the value of this
353 metadata element is appended to the resulting mergekey if
354 the metadata is present in a record instance.
355 If the metadata element is not present, the a unique mergekey
356 will be generated instead.
359 If set to '<literal>optional</literal>', the value of this
360 metadata element is appended to the resulting mergekey if the
361 the metadata is present in a record instance. If the metadata
362 is not present, it will be empty.
365 If set to '<literal>no</literal>' or the mergekey attribute is
366 omitted, the metadata will not be used in the creation of a
373 <term id="facetrule">facetrule</term>
376 Specifies the ICU rule set to be used for normalizing
377 facets. If facetrule is omitted from metadata, the
378 rule set 'facet' is used.
384 <term id="limitcluster">limitcluster</term>
387 Allow a limit on merged metadata. The value of this attribute
388 is the name of actual metadata content to be used for matching
389 (most often same name as metadata name).
393 Requires Pazpar2 1.6.23 or later.
400 <term id="metadata_limitmap">limitmap</term>
403 Specifies a default limitmap for this field. This is to avoid mass
404 configuring of targets. However it is important to review/do
405 this on a per target since it is usually target-specific.
406 See limitmap for format.
412 <term id="metadata_facetmap">facetmap</term>
415 Specifies a default facetmap for this field. This is to avoid mass
416 configuring of targets. However it is important to review/do
417 this on a per target since it is usually target-specific.
418 See facetmap for format.
424 <term id="icurule">icurule</term>
427 Specifies the ICU rule set to be used for normalizing
428 metadata text. The "display" part of the rule is kept
429 in the returned metadata record (record+show commands), the
430 end result - normalized text - is used for performing
431 within-cluster merge (unique, longest, etc). If the icurule is
432 omitted, type generic (text) is converted as follows:
433 any of the characters "<literal> ,/.:([</literal>" are
434 chopped of prefix and suffix of text content
435 <emphasis>unless</emphasis> it includes the
436 characters "<literal>://</literal>" (URL).
440 Requires Pazpar2 1.9.0 or later.
450 This attribute allows you to make use of static database
451 settings in the processing of records. Three possible values
452 are allowed. 'no' is the default and doesn't do anything.
453 'postproc' copies the value of a setting with the same name
454 into the output of the normalization stylesheet(s). 'parameter'
455 makes the value of a setting with the same name available
456 as a parameter to the normalization stylesheet, so you
457 can further process the value inside of the stylesheet, or use
458 the value to decide how to deal with other data values.
461 The purpose of using settings in this way can either be to
462 control the behavior of normalization stylesheet in a database-
463 dependent way, or to easily make database-dependent values
464 available to display-logic in your user interface, without having
465 to implement complicated interactions between the user interface
466 and your configuration system.
471 </variablelist> <!-- attributes to metadata -->
477 <term id="servicexslt" xreflabel="xslt">xslt</term>
480 Defines a XSLT stylesheet. The <literal>xslt</literal>
481 element takes exactly one attribute <literal>id</literal>
482 which names the stylesheet. This can be referred to in target
483 settings <xref linkend="pzxslt"/>.
486 The content of the xslt element is the embedded stylesheet XML
491 <term id="icuchain" xreflabel="icu_chain">icu_chain</term>
494 Specifies a named ICU rule set. The icu_chain element must include
495 attribute 'id' which specifies the identifier (name) for the ICU
497 Pazpar2 uses the particular rule sets for particular purposes.
498 Rule set 'relevance' is used to normalize
499 terms for relevance ranking. Rule set 'sort' is used to
500 normalize terms for sorting. Rule set 'mergekey' is used to
501 normalize terms for making a mergekey and, finally. Rule set 'facet'
502 is normally used to normalize facet terms, unless
503 <xref linkend="facetrule">facetrule</xref> is given for a
507 The icu_chain element must also include a 'locale'
508 attribute which must be set to one of the locale strings
509 defined in ICU. The child elements listed below can be
510 in any order, except the 'index' element which logically
511 belongs to the end of the list. The stated tokenization,
512 transformation and charmapping instructions are performed
513 in order from top to bottom.
515 <variablelist> <!-- Level 2 -->
520 The attribute 'rule' defines the direction of the
521 per-character casemapping, allowed values are "l"
522 (lower), "u" (upper), "t" (title).
527 <term>transform</term>
530 Normalization and transformation of tokens follows
531 the rules defined in the 'rule' attribute. For
532 possible values we refer to the extensive ICU
533 documentation found at the
534 <ulink url="&url.icu.transform;">ICU
535 transformation</ulink> home page. Set filtering
536 principles are explained at the
537 <ulink url="&url.icu.unicode.set;">ICU set and
538 filtering</ulink> page.
543 <term>tokenize</term>
546 Tokenization is the only rule in the ICU chain
547 which splits one token into multiple tokens. The
548 'rule' attribute may have the following values:
549 "s" (sentence), "l" (line-break), "w" (word), and
550 "c" (character), the later probably not being
551 very useful in a pruning Pazpar2 installation.
557 From Pazpar2 version 1.1 the ICU wrapper from YAZ is used.
558 Refer to the <ulink url="&url.yaz.yaz-icu;">yaz-icu</ulink>
559 utility for more information.
565 <term>relevance</term>
568 Specifies the ICU rule set used for relevance ranking.
569 The child element of 'relevance' must be 'icu_chain' and the
570 'id' attribute of the icu_chain is ignored. This
571 definition is obsolete and should be replaced by the equivalent
574 <icu_chain id="relevance" locale="en">..<icu_chain>
584 Specifies the ICU rule set used for sorting.
585 The child element of 'sort' must be 'icu_chain' and the
586 'id' attribute of the icu_chain is ignored. This
587 definition is obsolete and should be replaced by the equivalent
590 <icu_chain id="sort" locale="en">..<icu_chain>
597 <term>mergekey</term>
600 Specifies ICU tokenization and transformation rules
601 for tokens that are used in Pazpar2's mergekey.
602 The child element of 'mergekey' must be 'icu_chain' and the
603 'id' attribute of the icu_chain is ignored. This
604 definition is obsolete and should be replaced by the equivalent
607 <icu_chain id="mergekey" locale="en">..<icu_chain>
617 Specifies ICU tokenization and transformation rules
618 for tokens that are used in Pazpar2's facets.
619 The child element of 'facet' must be 'icu_chain' and the
620 'id' attribute of the icu_chain is ignored. This
621 definition is obsolete and should be replaced by the equivalent
624 <icu_chain id="facet" locale="en">..<icu_chain>
631 <term>ccldirective</term>
634 Customizes the CCL parsing (interpretation of query parameter
636 The name and value of the CCL directive is gigen by attributes
637 'name' and 'value' respectively. Refer to possible list of names
640 url="http://www.indexdata.com/yaz/doc/tools.html#ccl.directives.table">
647 <varlistentry id="service-rank">
651 Customizes the ranking (relevance) algorithm. Also known as
652 rank tweaks. The rank element
653 accepts the following attributes - all being optional:
660 Attribute 'cluster' is a boolean
661 that controls whether Pazpar2 should boost ranking for merged
662 records. Is 'yes' by default. A value of 'no' will make
663 Pazpar2 average ranking of each record in a cluster.
671 Attribute 'debug' is a boolean
672 that controls whether Pazpar2 should include details
673 about ranking for each document in the show command's
674 response. Enable by using value "yes", disable by using
675 value "no" (default).
683 Attribute 'follow' is a a floating point number greater than
684 or equal to 0. A positive number will boost weight for terms
685 that occur close to each other (proximity, distance).
686 A value of 1, will double the weight if two terms are in
687 proximity distance of 1 (next to each other). The default
688 value of 'follow' is 0 (order will not affect weight).
696 Attribute 'lead' is a floating point number.
697 It controls if term weight should be reduced by position
698 from start in a metadata field. A positive value of 'lead'
699 will reduce weight as it apperas further away from the lead
700 of the field. Default value is 0 (no reduction of weight by
709 Attribute 'length' determines how/if term weight should be
710 divided by lenght of metadata field. A value of "linear"
711 divide by length. A value of "log" will divide by log2(length).
712 A value of "none" will leave term weight as is (no division).
713 Default value is "linear".
719 Refer to <xref linkend="relevance_ranking"/> to see how
720 these tweaks are used in computation of score.
723 Customization of ranking algorithm was introduced with
724 Pazpar2 1.6.18. The semantics of some of the fields changed
725 in versions up to 1.6.22.
730 <varlistentry id="sort-default">
731 <term>sort-default</term>
734 Specifies the default sort criteria (default 'relevance'),
735 which previous was hard-coded as default criteria in search.
736 This is a fix/work-around to avoid re-searching when using
737 target-based sorting. In order for this to work efficient,
738 the search must also have the sort critera parameter; otherwise
739 pazpar2 will do re-searching on search criteria changes, if
740 changed between search and show command.
743 This configuration was added in pazpar2 1.6.20.
753 Specifies a variable that will be inherited by all targets defined in settings
755 <set name="test" value="en"..<set>
762 <term>settings</term>
765 Specifies target settings for this service. Refer to
766 <xref linkend="target_settings"/>.
771 <varlistentry id="service-timeout">
775 Specifies timeout parameters for this service.
776 The <literal>timeout</literal>
777 element supports the following attributes:
778 <literal>session</literal>, <literal>z3950_operation</literal>,
779 <literal>z3950_session</literal> which specifies
780 'session timeout', 'Z39.50 operation timeout',
781 'Z39.50 session timeout' respectively. The Z39.50 operation
782 timeout is the time Pazpar2 will wait for an active Z39.50/SRU
783 operation before it gives up (times out). The Z39.50 session
784 time out is the time Pazpar2 will keep the session alive for
785 an idle session (no operation).
788 The following is recommended but not required:
789 z3950_operation (30) < session (60) < z3950_session (180) .
790 The default values are given in parantheses.
793 The Z39.50 operation timeout may be set per database. Refer to
794 <xref linkend="pztimeout"/>.
798 </variablelist> <!-- Data elements in service directive -->
801 </variablelist> <!-- Data elements in server directive -->
806 <title>EXAMPLE</title>
808 Below is a working example configuration:
812 <?xml version="1.0" encoding="UTF-8"?>
813 <pazpar2 xmlns="http://www.indexdata.com/pazpar2/1.0">
815 <threads number="10"/>
817 <listen port="9004"/>
820 <metadata name="title" brief="yes" sortkey="skiparticle"
821 merge="longest" rank="6"/>
822 <metadata name="isbn" merge="unique"/>
823 <metadata name="date" brief="yes" sortkey="numeric"
824 type="year" merge="range" termlist="yes"/>
825 <metadata name="author" brief="yes" termlist="yes"
826 merge="longest" rank="2"/>
827 <metadata name="subject" merge="unique" termlist="yes" rank="3" limitmap="local:"/>
828 <metadata name="url" merge="unique"/>
829 <icu_chain id="relevance" locale="el">
830 <transform rule="[:Control:] Any-Remove"/>
832 <transform rule="[[:WhiteSpace:][:Punctuation:]] Remove"/>
835 <settings src="mysettings"/>
836 <timeout session="60"/>
844 <refsect1 id="config-include">
845 <title>INCLUDE FACILITY</title>
847 The XML configuration may be partitioned into multiple files by using
848 the <literal>include</literal> element which takes a single attribute,
849 <literal>src</literal>. The <literal>src</literal> attribute is
850 regular Shell like glob-pattern. For example,
852 <include src="/etc/pazpar2/conf.d/*.xml"/>
856 The include facility requires Pazpar2 version 1.2.
860 <refsect1 id="target_settings">
861 <title>TARGET SETTINGS</title>
863 Pazpar2 features a cunning scheme by which you can associate various
864 kinds of attributes, or settings with search targets. This can be done
865 through XML files which are read at startup; each file can associate
866 one or more settings with one or more targets. The file format is generic
867 in nature, designed to support a wide range of application requirements.
868 The settings can be purely technical things, like, how to perform a title
869 search against a given target, or it can associate arbitrary name=value
870 pairs with groups of targets -- for instance, if you would like to
871 place all commercial full-text bases in one group for selection
872 purposes, or you would like to control what targets are accessible
873 to users by default. Per-database settings values can even be used
874 to drive sorting, facet/termlist generation, or end-user interface display
879 During startup, Pazpar2 will recursively read a specified directory
880 (can be identified in the pazpar2.cfg file or on the command line), and
881 process any settings files found therein.
885 Clients of the Pazpar2 webservice interface can selectively override
886 settings for individual targets within the scope of one session. This
887 can be used in conjunction with an external authentication system to
888 determine which resources are to be accessible to which users. Pazpar2
889 itself has no notion of end-users, and so can be used in conjunction
890 with any type of authentication system. Similarly, the authentication
891 tokens submitted to access-controlled search targets can similarly be
892 overridden, to allow use of Pazpar2 in a consortial or multi-library
893 environment, where different end-users may need to be represented to
894 some search targets in different ways. This, again, can be managed
895 using an external database or other lookup mechanism. Setting overrides
896 can be performed either using the
897 <link linkend="command-init">init</link> or the
898 <link linkend="command-settings">settings</link> webservice
903 In fact, every setting that applies to a database (except pz:id, which
904 can only be used for filtering targets to use for a search) can be overridden
905 on a per-session basis.
906 This allows the client to override specific CCL fields for
907 searching, etc., to meet the needs of a session or user.
911 Finally, as an extreme case of this, the webservice client can
912 introduce entirely new targets, on the fly, as part of the
913 <link linkend="command-init">init</link> or
914 <link linkend="command-settings">settings</link> command.
915 This is useful if you desire to manage information
916 about your search targets in a separate application such as a database.
917 You do not need any static settings file whatsoever to run Pazpar2 -- as
918 long as the webservice client is prepared to supply the necessary
919 information at the beginning of every session.
924 The following discussion of practical issues related to session
925 and settings management are cast in terms of a user interface based on
926 Ajax/Javascript technology. It would apply equally well to many other
927 kinds of browser-based logic.
932 Typically, a Javascript client is not allowed to directly alter the
933 parameters of a session. There are two reasons for this. One has to do
934 with access to information; typically, information about a user will
935 be stored in a system on the server side, or it will be accessible in
936 some way from the server. However, since the Javascript client cannot
937 be entirely trusted (some hostile agent might in fact 'pretend' to be
938 a regular ws client), it is more robust to control session settings
939 from scripting that you run as part of your webserver. Typically, this
940 can be handled during the session initialization, as follows:
944 Step 1: The Javascript client loads, and asks the webserver for a
945 new Pazpar2 session ID. This can be done using a Javascript call, for
946 instance. Note that it is possible to submit Ajax HTTPXmlRequest calls
947 either to Pazpar2 or to the webserver that Pazpar2 is proxying
948 for. See (XXX Insert link to Pazpar2 protocol).
952 Step 2: Code on the webserver authenticates the user, by database lookup,
953 LDAP access, NCIP, etc. Determines which resources the user has access to,
954 and any user-specific parameters that are to be applied during this session.
958 Step 3: The webserver initializes a new Pazpar2 settings, and sets
959 user-specific parameters as necessary, using the init webservice
960 command. A new session ID is returned.
964 Step 4: The webserver returns this session ID to the Javascript
965 client, which then uses the session ID to submit searches, show
970 Step 5: When the Javascript client ceases to use the session,
971 Pazpar2 destroys any session-specific information.
975 <title>SETTINGS FILE FORMAT</title>
977 Each file contains a root element named <settings>. It may
978 contain one or more <set> elements. The settings and set
979 elements may contain the following attributes. Attributes in the set
980 node overrides those in the setting root element. Each set node must
981 specify (directly, or inherited from the parent node) at least a
982 target, name, and value.
990 This specifies the search target to which this setting should be
991 applied. Targets are identified by their Z39.50 URL, generally
992 including the host, port, and database name, (e.g.
993 <literal>bagel.indexdata.com:210/marc</literal>).
994 Two wildcard forms are accepted:
995 * (asterisk) matches all known targets;
996 <literal>bagel.indexdata.com:210/*</literal> matches all
997 known databases on the given host.
1000 A precedence system determines what happens if there are
1001 overlapping values for the same setting name for the same
1002 target. A setting for a specific target name overrides a
1003 setting which specifies target using a wildcard. This makes it
1004 easy to set defaults for all targets, and then override them
1005 for specific targets or hosts. If there are
1006 multiple overlapping settings with the same name and target
1007 value, the 'precedence' attribute determines what happens.
1010 For Pazpar2 1.6.4 or later, the target ID may be user-defined, in
1011 which case, the actual host, port, etc is given by setting
1012 <xref linkend="pzurl"/>.
1020 The name of the setting. This can be anything you like.
1021 However, Pazpar2 reserves a number of setting names for
1022 specific purposes, all starting with 'pz:', and it is a good
1023 idea to avoid that prefix if you make up your own setting
1024 names. See below for a list of reserved variables.
1032 The value of the setting. Generally, this can be anything you
1033 want -- however, some of the reserved settings may expect
1034 specific kinds of values.
1039 <term>precedence</term>
1042 This should be an integer. If not provided, the default value
1043 is 0. If two (or more) settings have the same content for
1044 target and name, the precedence value determines the outcome.
1045 If both settings have the same precedence value, they are both
1046 applied to the target(s). If one has a higher value, then the
1047 value of that setting is applied, and the other one is ignored.
1054 By setting defaults for target, name, or value in the root
1055 settings node, you can use the settings files in many different
1056 ways. For instance, you can use a single file to set defaults for
1057 many different settings, like search fields, retrieval syntaxes,
1058 etc. You can have one file per server, which groups settings for
1059 that server or target. You could also have one file which associates
1060 a number of targets with a given setting, for instance, to associate
1061 many databases with a given category or class that makes sense
1062 within your application.
1066 The following examples illustrate uses of the settings system to
1067 associate settings with targets to meet different requirements.
1071 The example below associates a set of default values that can be
1072 used across many targets. Note the wildcard for targets.
1073 This associates the given settings with all targets for which no
1074 other information is provided.
1076 <settings target="*">
1078 <!-- This file introduces default settings for pazpar2 -->
1080 <!-- mapping for unqualified search -->
1081 <set name="pz:cclmap:term" value="u=1016 t=l,r s=al"/>
1083 <!-- field-specific mappings -->
1084 <set name="pz:cclmap:ti" value="u=4 s=al"/>
1085 <set name="pz:cclmap:su" value="u=21 s=al"/>
1086 <set name="pz:cclmap:isbn" value="u=7"/>
1087 <set name="pz:cclmap:issn" value="u=8"/>
1088 <set name="pz:cclmap:date" value="u=30 r=r"/>
1090 <set name="pz:limitmap:title" value="rpn:@attr 1=4 @attr 6=3"/>
1091 <set name="pz:limitmap:date" value="ccl:date"/>
1093 <!-- Retrieval settings -->
1095 <set name="pz:requestsyntax" value="marc21"/>
1096 <set name="pz:elements" value="F"/>
1098 <!-- Query encoding -->
1099 <set name="pz:queryencoding" value="iso-8859-1"/>
1101 <!-- Result normalization settings -->
1103 <set name="pz:nativesyntax" value="iso2709"/>
1104 <set name="pz:xslt" value="../etc/marc21.xsl"/>
1112 The next example shows certain settings overridden for one target,
1113 one which returns XML records containing DublinCore elements, and
1114 which furthermore requires a username/password.
1116 <settings target="funkytarget.com:210/db1">
1117 <set name="pz:requestsyntax" value="xml"/>
1118 <set name="pz:nativesyntax" value="xml"/>
1119 <set name="pz:xslt" value="../etc/dublincore.xsl"/>
1121 <set name="pz:authentication" value="myuser/password"/>
1127 The following example associates a specific name/value combination
1128 with a number of targets. The targets below are access-restricted,
1129 and can only be used by users with special credentials.
1131 <settings name="pz:allow" value="0">
1132 <set target="funkytarget.com:210/*"/>
1133 <set target="commercial.com:2100/expensiveDb"/>
1141 <title>RESERVED SETTING NAMES</title>
1143 The following setting names are reserved by Pazpar2 to control the
1144 behavior of the client function.
1150 <term>pz:allow</term>
1153 Allows or denies access to the resources it is applied to. Possible
1154 values are '0' and '1'.
1155 The default is '1' (allow access to this resource).
1161 <term>pz:apdulog</term>
1164 If the 'pz:apdulog' setting is defined and has other value than 0,
1165 then Z39.50 APDUs are written to the log.
1171 <term>pz:authentication</term>
1174 Sets an authentication string for a given database. For Z39.50,
1175 this is carried as part of the Initialize Request. In order to carry
1176 the information in the "open" elements, separate
1177 username and password with a slash (In Z39.50 it is a VisibleString).
1178 In order to carry the information in the idPass elements, separate
1179 username term, password term and, optionally, a group term with a
1181 If three terms are given, the order is
1182 <emphasis>user, group, password</emphasis>.
1183 If only two terms are given, the order is
1184 <emphasis>user, password</emphasis>.
1187 For HTTP based procotols, such as SRU and Solr, the authentication
1188 string includes a username term and, optionally, a password term.
1189 Each term is separated by a single blank. The
1190 authentication information is passed either by HTTP basic
1191 authentication or via URL parameters. The mode is operation is
1192 determined by <literal>pz:authentication_mode</literal> setting.
1198 <term>pz:authentication_mode</term>
1201 Determines how authentication is carried in HTTP based protocols.
1202 Value may be "<literal>basic</literal>" or "<literal>url</literal>".
1208 <term>pz:block_timeout</term>
1211 (Not yet implemented).
1212 Specifies the time for which a block should be released anyway.
1218 <term>pz:cclmap:xxx</term>
1221 This establishes a CCL field definition or other setting, for
1222 the purpose of mapping end-user queries. XXX is the field or
1223 setting name, and the value of the setting provides parameters
1224 (e.g. parameters to send to the server, etc.). Please consult
1225 the YAZ manual for a full overview of the many capabilities of
1226 the powerful and flexible CCL parser.
1229 Note that it is easy to establish a set of default parameters,
1230 and then override them individually for a given target.
1236 <term>pz:elements</term>
1239 The element set name to be used when retrieving records from a
1246 <term>pz:extendrecs</term>
1249 If a show command goes to the boundary of a result set for a
1250 database - depends on sorting - and pz:extendrecs is set to a positive
1251 value. then Pazpar2 wait for show to fetch pz:extendrecs more
1252 records. This setting is best used if a database does native
1253 sorting, because the result set otherwise may be completely
1254 re-sorted during extended fetch.
1255 The default value of pz:extendrecs is 0 (no extended fetch).
1259 The pz:extendrecs setting appeared in Pazpar2 version 1.6.26.
1260 But the bahavior changed with the release of Pazpar2 1.6.29.
1267 <term>pz:facetmap:<replaceable>name</replaceable></term>
1270 Specifies that for field <replaceable>name</replaceable>, the target
1271 supports (native) facets. The value is the name of the
1272 field on the target.
1276 At this point only Solr targets have been tested with this
1287 This setting can't be 'set' -- it contains the ID (normally
1288 ZURL) for a given target, and is useful for filtering --
1289 specifically when you want to select one or more specific
1290 targets in the search command.
1295 <varlistentry id="limitmap">
1296 <term>pz:limitmap:<replaceable>name</replaceable></term>
1299 Specifies attributes for limiting a search to a field - using
1300 the limit parameter for search. It can be used to filter locally
1301 or remotely (search in a target). In some cases the mapping of
1302 a field to a value is identical to an existing cclmap field; in
1303 other cases the field must be specified in a different way - for
1304 example to match a complete field (rather than parts of a subfield).
1307 The value of limitmap may have one of three forms: referral to
1308 an existing CCL field, a raw PQF string or a local limit. Leading string
1309 determines type; either <literal>ccl:</literal> for CCL field,
1310 <literal>rpn:</literal> for PQF/RPN, or <literal>local:</literal>
1311 for filtering in Pazpar2. The local filtering may be followed
1312 by a field a metadata field (default is to use the name of the
1316 For Pazpar2 version 1.6.23 and later the limitmap may include multiple
1317 specifications, separated by <literal>,</literal> (comma).
1319 <literal>ccl:title,local:ltitle,rpn:@attr 1=4</literal>.
1323 The limitmap facility is supported for Pazpar2 version 1.6.0.
1324 Local filtering is supported in Pazpar2 1.6.6.
1331 <term>pz:maxrecs</term>
1334 Controls the maximum number of records to be retrieved from a
1335 server. The default is 100.
1341 <term>pz:memcached</term>
1344 If set and non-empty,
1345 <ulink url="&url.libmemcached;">libMemcached</ulink> will
1346 configured and enabled for the target.
1347 The value of this setting is same as the ZOOM option
1348 <literal>memcached</literal>, which in turn is the configuration
1349 string passed to the <function>memcached</function> function
1350 of <ulink url="&url.libmemcached;">libMemcached</ulink>.
1353 This setting is honored in Pazpar2 1.6.39 or later. Pazpar2 must
1354 be using YAZ version 5.0.13 or later.
1360 <term>pz:redis</term>
1363 If set and non-empty,
1364 <ulink url="&url.redis;">redis</ulink> will
1365 configured and enabled for the target.
1366 The value of this setting is exactly as the redis option for
1370 This setting is honored in Pazpar2 1.6.43 or later. Pazpar2 must
1371 be using YAZ version 5.2.0 or later.
1377 <term>pz:nativesyntax</term>
1380 Specifies how Pazpar2 shoule map retrieved records to XML. Currently
1381 supported values are <literal>xml</literal>,
1382 <literal>iso2709</literal> and <literal>txml</literal>.
1385 The value <literal>iso2709</literal> makes Pazpar2 convert retrieved
1386 MARC records to MARCXML. In order to convert to XML, the exact
1387 chacater set of the MARC must be known (if not, the resulting
1388 XML is probably not well-formed). The character set may be
1389 specified by adding:
1390 <literal>;</literal><replaceable>charset</replaceable> to
1391 <literal>iso2709</literal>. If omitted, a charset of
1392 MARC-8 is assumed. This is correct for most MARC21/USMARC records.
1395 The value <literal>txml</literal> is like <literal>iso2709</literal>
1396 except that records are converted to TurboMARC instead of MARCXML.
1399 The value <literal>xml</literal> is used if Pazpar2 retrieves
1400 records that are already XML (no conversion takes place).
1406 <term>pz:negotiation_charset</term>
1409 Sets character set for Z39.50 negotiation. Most targets do not support
1410 this, and some will even close connection if set (crash on server
1411 side or similar). If set, you probably want to set it to
1412 <literal>UTF-8</literal>.
1418 <term>pz:piggyback</term>
1421 Piggybacking enables the server to retrieve records from the
1422 server as part of the search response in Z39.50. Almost all
1423 servers support this (or fail it gracefully), but a few
1424 servers will produce undesirable results.
1425 Set to '1' to enable piggybacking, '0' to disable it. Default
1426 is 1 (piggybacking enabled).
1431 <term>pz:pqf_prefix</term>
1434 Allows you to specify an arbitrary PQF query language substring.
1435 The provided string is prefixed to the user's query after it has been
1436 normalized to PQF internally in pazpar2.
1437 This allows you to attach complex 'filters' to queries for a given
1438 target, sometimes necessary to select sub-catalogs
1439 in union catalog systems, etc.
1445 <term>pz:pqf_strftime</term>
1448 Allows you to extend a query with dates and operators.
1449 The provided string allows certain substitutions and serves as a
1451 The special two character sequence '%%' gets converted to the
1452 original query. Other characters leading with the percent sign are
1453 conversions supported by strftime.
1454 All other characters are copied verbatim. For example, the string
1455 <literal>@and @attr 1=30 @attr 2=3 %Y %%</literal>
1456 would search for current year combined with the original PQF (%%).
1459 This setting can also be used as more general alternative to
1460 pz:pqf_prefix -- a way of embedding the submitted query
1461 anywhere in the string rather than appending it to prefix. For
1462 example, if it is desired to omit all records satisfying the
1463 query <literal>@attr 1=pica.bib 0007</literal> then this
1464 subquery can be combined with the submitted query as the second
1465 argument of <literal>@andnot</literal> by using the
1466 pz:pqf_strftime value <literal>@not %% @attr 1=pica.bib
1473 <term>pz:preferred</term>
1476 Specifies that a target is preferred, e.g. possible local, faster
1477 target. Using block=preferred on <link linkend="command-show">
1478 show command</link> will wait for all these
1479 targets to return records before releasing the block.
1480 If no target is preferred, the block=preferred will identical to
1481 block=1, which release when one target has returned records.
1487 <term>pz:present_chunk</term>
1490 Controls the chunk size in present requests. Pazpar2 will
1491 make (maxrecs / chunk) request(s). The default is 20.
1497 <term>pz:queryencoding</term>
1500 The encoding of the search terms that a target accepts. Most
1501 targets do not honor UTF-8 in which case this needs to be specified.
1502 Each term in a query will be converted if this setting is given.
1508 <term>pz:recordfilter</term>
1511 Specifies a filter which allows Pazpar2 to only include
1512 records that meet a certain criteria in a result.
1513 Unmatched records will be ignored.
1514 The filter takes the form name, name~value, or name=value, which
1515 will include only records with metadata element (name) that has the
1516 substring (~value) given, or matches exactly (=value).
1517 If value is omitted all records with the named metadata element
1518 present will be included.
1523 <varlistentry id="requestsyntax">
1524 <term>pz:requestsyntax</term>
1527 This specifies the record syntax to use when requesting
1528 records from a given server. The value can be a symbolic name like
1529 marc21 or xml, or it can be a Z39.50-style dot-separated OID.
1535 <term>pz:sort</term>
1538 Specifies sort criteria to be applied to the result set.
1539 Only works for targets which support the sort service.
1544 <varlistentry id="pzsortmap">
1545 <term>pz:sortmap:<replaceable>field</replaceable></term>
1548 Specifies native sorting for a target where
1549 <replaceable>field</replaceable> is a sort criterion (see command
1550 show). The value has two components separated by a colon: strategy and
1551 native-field. Strategy is one of <literal>z3950</literal>,
1552 <literal>type7</literal>, <literal>cql</literal>,
1553 <literal>sru11</literal>, or <literal>embed</literal>.
1554 The second component, native-field, is the field that is recognized
1559 Only supported for Pazpar2 1.6.4 and later.
1569 This setting enables
1570 <ulink url="&url.sru;">SRU</ulink>/<ulink url="&url.solr;">Solr</ulink>
1572 It has four possible settings.
1573 'get', enables SRU access through GET requests. 'post' enables SRU/POST
1574 support, less commonly supported, but useful if very large requests are
1575 to be submitted. 'soap' enables the SRW (SRU over SOAP) variation of
1579 A value of 'solr' enables Solr client support. This is supported
1580 for Pazpar version 1.5.0 and later.
1586 <term>pz:sru_version</term>
1589 This allows SRU version to be specified. If unset Pazpar2
1590 will the default of YAZ (currently 1.2). Should be set
1591 to 1.1 or 1.2. For Solr, the current supported/tested version
1598 <term>pz:termlist_term_count</term>
1601 Specifies number of facet terms to be requested from the target.
1602 The default is unspecified e.g. server-decided. Also see pz:facetmap.
1608 <term>pz:termlist_term_factor</term>
1611 Specifies whether to use a factor for pazpar2 generated facets (1)
1613 When mixing locally generated (by the downloaded (pz:maxrecs) samples)
1614 facet with native (target-generated) facets, the later will
1615 dominated the dominate the facet list since they are generated
1616 based on the complete result set.
1617 By scaling up the facet count using the ratio between total hit
1618 count and the sample size,
1619 the total facet count can be approximated and thus better compared
1620 with native facets. This is not enabled by default.
1626 <varlistentry id="pztimeout">
1627 <term>pz:timeout</term>
1630 Specifies timeout for operation (eg search, and fetch) for
1631 a database. This overrides the z3650_operation timeout
1632 that is given for a service. See <xref linkend="service-timeout"/>.
1636 The timeout facility is supported for Pazpar2 version 1.8.4 and later.
1642 <varlistentry id="pzurl">
1646 Specifies URL for the target and overrides the target ID.
1650 <literal>pz:url</literal> is only recognized for
1651 Pazpar2 1.6.4 and later.
1658 <term id="pzxslt" xreflabel="pz:xslt">pz:xslt</term>
1661 Is a comma separated list of of stylesheet names that specifies
1662 how to convert incoming records to the internal representation.
1665 For each name, the embedded stylesheets (XSL) that comes with the
1666 service definition are consulted first and takes precedence over
1667 external files; see <xref linkend="servicexslt"/>
1668 of service definition).
1669 If the name does not match an embedded stylesheet it is
1670 considered a filename.
1673 The suffix of each file specifies the kind of tranformation.
1674 Suffix "<literal>.xsl</literal>" makes an XSL transform. Suffix
1675 "<literal>.mmap</literal>" will use the MMAP transform (described below).
1678 The special value "<literal>auto</literal>" will use a file
1679 which is the <link linkend="requestsyntax">pz:requestsyntax's</link>
1681 <literal>'.xsl'</literal>.
1684 When mapping MARC records, XSLT can be bypassed for increased
1685 performance with the alternate "MARC map" format. Provide the
1686 path of a file with extension ".mmap" containing on each line:
1688 <field> <subfield> <metadata element></programlisting>
1695 To map the field value specify a subfield of '$'. To store a
1696 concatenation of all subfields, specify a subfield of '*'.
1702 <term>pz:zproxy</term>
1705 The 'pz:zproxy' setting has the value syntax
1706 'host.internet.adress:port', it is used to tunnel Z39.50
1707 requests through the named Z39.50 proxy.
1717 <title>SEE ALSO</title>
1720 <refentrytitle>pazpar2</refentrytitle>
1721 <manvolnum>8</manvolnum>
1724 <refentrytitle>yaz-icu</refentrytitle>
1725 <manvolnum>1</manvolnum>
1728 <refentrytitle>pazpar2_protocol</refentrytitle>
1729 <manvolnum>7</manvolnum>
1734 <!-- Keep this comment at the end of the file
1737 nxml-child-indent: 1