1 <?xml version="1.0" standalone="no"?>
2 <!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML 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">
14 <title>Pazpar2 - User's Guide and Reference</title>
16 <firstname>Sebastian</firstname><surname>Hammer</surname>
19 <firstname>Adam</firstname><surname>Dickmeiss</surname>
22 <firstname>Marc</firstname><surname>Cromme</surname>
25 <firstname>Jakub</firstname><surname>Skoczen</surname>
28 <firstname>Mike</firstname><surname>Taylor</surname>
31 <firstname>Dennis</firstname><surname>Schafroth</surname>
33 <releaseinfo>&version;</releaseinfo>
35 <year>©right-year;</year>
36 <holder>Index Data</holder>
40 Pazpar2 is a high-performance metasearch engine featuring
41 merging, relevance ranking, record sorting,
43 It is middleware: it has no user interface of its own, but can be
44 configured and controlled by an XML-over-HTTP web-service to provide
45 metasearching functionality behind any user interface.
48 This document is a guide and reference to Pazpar2 version &version;.
53 <imagedata fileref="common/id.png" format="PNG"/>
56 <imagedata fileref="common/id.eps" format="EPS"/>
63 <chapter id="introduction">
64 <title>Introduction</title>
66 <section id="what.pazpar2.is">
67 <title>What Pazpar2 is</title>
69 Pazpar2 is a stand-alone metasearch engine with a web-service API, designed
70 to be used either from a browser-based client (JavaScript, Flash,
72 etc.), from server-side code, or any combination of the two.
73 Pazpar2 is a highly optimized client designed to
74 search many resources in parallel. It implements record merging,
75 relevance-ranking and sorting by arbitrary data content, and facet
76 analysis for browsing purposes. It is designed to be data-model
77 independent, and is capable of working with MARC, DublinCore, or any
78 other <ulink url="&url.xml;">XML</ulink>-structured response format
79 -- <ulink url="&url.xslt;">XSLT</ulink> is used to normalize and extract
80 data from retrieval records for display and analysis. It can be used
81 against any server which supports the
82 <ulink url="&url.z39.50;">Z39.50</ulink>,
83 <ulink url="&url.sru;">SRU/SRW</ulink>
84 or <ulink url="&url.solr;">Solr</ulink> protocol. Proprietary
85 backend modules can function as connectors between these standard
86 protocols and any non-standard API, including web-site scraping, to
87 support a large number of other protocols.
90 Additional functionality such as
91 user management and attractive displays are expected to be implemented by
92 applications that use Pazpar2. Pazpar2 itself is user-interface independent.
93 Its functionality is exposed through a simple XML-based web-service API,
94 designed to be easy to use from an Ajax-enabled browser, Flash
95 animation, Java applet, etc., or from a higher-level server-side language
96 like PHP, Perl or Java. Because session information can be shared between
97 browser-based logic and server-side scripting, there is tremendous
98 flexibility in how you implement application-specific logic on top
102 Once you launch a search in Pazpar2, the operation continues behind the
103 scenes. Pazpar2 connects to servers, carries out searches, and
104 retrieves, deduplicates, and stores results internally. Your application
105 code may periodically inquire about the status of an ongoing operation,
106 and ask to see records or result set facets. Results become
107 available immediately, and it is easy to build end-user interfaces than
108 feel extremely responsive, even when searching more than 100 servers
112 Pazpar2 is designed to be highly configurable. Incoming records are
113 normalized to XML/UTF-8, and then further normalized using XSLT to a
114 simple internal representation that is suitable for analysis. By
115 providing XSLT stylesheets for different kinds of result records, you
116 can configure Pazpar2 to work against different kinds of information
117 retrieval servers. Finally, metadata is extracted in a configurable
118 way from this internal record, to support display, merging, ranking,
119 result set facets, and sorting. Pazpar2 is not bound to a specific model
120 of metadata, such as DublinCore or MARC: by providing the right
121 configuration, it can work with any combination of different kinds of data
122 in support of many different applications.
125 Pazpar2 is designed to be efficient and scalable. You can set it up to
126 search several hundred targets in parallel, or you can use it to support
127 hundreds of concurrent users. It is implemented with the same attention
128 to performance and economy that we use in our indexing engines, so that
129 you can focus on building your application without worrying about the
130 details of metasearch logic. You can devote all of your attention to
131 usability and let Pazpar2 do what it does best -- metasearch.
134 Pazpar2 is our attempt to re-think the traditional paradigms for
135 implementing and deploying metasearch logic, with an uncompromising
136 approach to performance, and attempting to make maximum use of the
137 capabilities of modern browsers. The demo user interface that
138 accompanies the distribution is but one example. If you think of new
139 ways of using Pazpar2, we hope you'll share them with us, and if we
140 can provide assistance with regards to training, design, programming,
141 integration with different backends, hosting, or support, please don't
142 hesitate to contact us. If you'd like to see functionality in Pazpar2
143 that is not there today, please don't hesitate to contact us. It may
144 already be in our development pipeline, or there might be a
145 possibility for you to help out by sponsoring development time or
146 code. Either way, get in touch and we will give you straight answers.
152 Pazpar2 is covered by the GNU General Public License (GPL) version 2.
153 See <xref linkend="license"/> for further information.
157 <section id="connectors">
158 <title>Connectors to non-standard databases</title>
160 If you need to access commercial or open access resources that don't support
161 Z39.50 or SRU, one approach would be to use a tool like <ulink
162 url="&url.simpleserver;">SimpleServer</ulink> to build a
163 gateway. An easier option is to use Index Data's <ulink
164 url="&url.mkc;">MasterKey Connect</ulink>
165 service, which will expose virtually <emphasis>any</emphasis> resource
166 through Z39.50/SRU, dead easy to integrate with Pazpar2.
167 The service is hosted, so all you have to do is to let us
168 know which resources you are interested in, and we operate the gateways,
169 or Connectors for you for a low annual charge.
170 Types of resources supported include
171 commercial databases, free online resources, and even local resources;
172 almost anything that can be accessed through a web-facing user
173 interface can be accessed in this way.
174 Contact <email>info@indexdata.com</email> for more information.
175 See <xref linkend="masterkey_connect"/> for an example.
180 <title>A note on the name Pazpar2</title>
182 The name Pazpar2 derives from three sources. One one hand, it is
183 Index Data's second major piece of software that does parallel
184 searching of Z39.50 targets. On the other, it is a near-homophone
185 of Passpartout, the ever-helpful servant in Jules Verne's novel
186 Around the World in Eighty Days (who helpfully uses the language
187 of his master). Finally, "passe par tout" means something like
188 "passes through anything" in French -- on other words, a universal
189 solution, or if you like a MasterKey.
194 <chapter id="installation">
195 <title>Installation</title>
197 The Pazpar2 package includes documentation as well
198 as the Pazpar2 server. The package also includes a simple user
199 interface called "test1", which consists of a single HTML page and a single
200 JavaScript file to illustrate the use of Pazpar2.
203 Pazpar2 depends on the following tools/libraries:
205 <varlistentry><term><ulink url="&url.yaz;">YAZ</ulink></term>
208 The popular Z39.50 toolkit for the C language.
209 YAZ <emphasis>must</emphasis> be compiled with
210 <ulink url="&url.libxml2;">Libxml2</ulink>/<ulink url="&url.libxslt;">Libxslt</ulink> support.
213 It is highly recommended that YAZ is also compiled with
214 <ulink url="&url.icu;">ICU</ulink> support.
221 In order to compile Pazpar2, a C compiler which supports C99 or later
225 <section id="installation.unix">
226 <title>Installation from source on Unix (including Linux, MacOS, etc.)</title>
228 The latest source code for Pazpar2 is available from
229 <ulink url="&url.pazpar2.download;"/>.
230 Most Unix-based operating systems have the required
231 tools available as binary packages.
232 For example, if Libxml2/libXSLT libraries
233 are already installed as development packages, use these.
237 Ensure that the development libraries and header files are
238 available on your system before compiling Pazpar2. For installation
239 of YAZ, refer to the Installation chapter of the YAZ manual at
240 <ulink url="&url.yaz.install;"/>.
243 Once the dependencies are in place, Pazpar2 can be unpacked and
244 installed as follows:
247 tar xzf pazpar2-VERSION.tar.gz
254 The <literal>make install</literal> will install manpages as well as the
255 Pazpar2 server, <literal>pazpar2</literal>,
256 in PREFIX<literal>/sbin</literal>.
257 By default, PREFIX is <literal>/usr/local/</literal> . This can be
258 changed with configure option <option>--prefix</option>.
262 <section id="installation.win32">
263 <title>Installation from source on Windows</title>
265 Pazpar2 can be built for Windows using
266 <ulink url="&url.vstudio;">Microsoft Visual Studio</ulink>.
267 The support files for building YAZ on Windows are located in the
268 <filename>win</filename> directory. The compilation is performed
269 using the <filename>win/makefile</filename> which is to be
270 processed by the NMAKE utility part of Visual Studio.
273 Ensure that the development libraries and header files are
274 available on your system before compiling Pazpar2. For installation
276 the Installation chapter of the YAZ manual at
277 <ulink url="&url.yaz.install;"/>.
278 It is easiest if YAZ and Pazpar2 are unpacked in the same
279 directory (side-by-side).
282 The compilation is tuned by editing the makefile of Pazpar2.
283 The process is similar to YAZ. Adjust the various directories
284 <literal>YAZ_DIR</literal>, <literal>ZLIB_DIR</literal>, etc.,
288 Compile Pazpar2 by invoking <application>nmake</application> in
289 the <filename>win</filename> directory.
290 The resulting binaries of the build process are located in the
291 <filename>bin</filename> of the Pazpar2 source
292 tree - including the <filename>pazpar2.exe</filename> and necessary DLLs.
295 The Windows version of Pazpar2 is a console application. It may
296 be installed as a Windows Service by adding option
297 <literal>-install</literal> for the pazpar2 program. This will
298 register Pazpar2 as a service and use the other options provided
299 in the same invocation. For example:
302 ..\bin\pazpar2 -install -f pazpar2.cfg -l pazpar2.log
304 The Pazpar2 service may now be controlled via the Service Control
305 Panel. It may be unregistered by passing the <literal>-remove</literal>
309 ..\bin\pazpar2 -remove
314 <section id="installation.test1">
315 <title>Installation of test interfaces</title>
317 In this section we show how to make available the set of simple
318 interfaces that are part of the Pazpar2 source package, and which
319 demonstrate some ways to use Pazpar2. (Note that Debian users can
320 save time by just installing the package <literal>pazpar2-test1</literal>.)
323 A web server, such as Apache, must be installed and running on the system.
327 Start the Pazpar2 daemon using the 'in-source' binary of the Pazpar2
328 daemon. On Unix the process is:
331 cp pazpar2.cfg.dist pazpar2.cfg
332 ../src/pazpar2 -f pazpar2.cfg
337 copy pazpar2.cfg.dist pazpar2.cfg
338 ..\bin\pazpar2 -f pazpar2.cfg
340 This will start a Pazpar2 listener on port 9004. It will proxy
341 HTTP requests to port 80 on localhost, which we assume will be the regular
342 HTTP server on the system. Inspect and modify pazpar2.cfg as needed
343 if this is to be changed. The pazpar2.cfg file includes settings from the
344 file <filename>settings/edu.xml</filename>
349 The test UIs are located in <literal>www</literal>. Ensure that this
350 directory is available to the web server by copying
351 <literal>www</literal> to the document root,
352 using Apache's <literal>Alias</literal> directive, or
353 creating a symbolic link: for example, on a Debian or Ubuntu
354 system with Apache2 installed from the standard package, you might
355 make the link as follows:
358 sudo ln -s `pwd`/www /var/www/pazpar2-demo
363 This makes the test applications visible at
364 <ulink url="http://localhost/pazpar2-demo/"/>
365 but they can not be run successfully from that URL, as they submit
366 search requests back to the server form which they were served,
367 and Apache2 doesn't know how to handle them. Instead, the test
368 applications must be accessed from Pazpar2 itself, acting as a
369 proxy to Apache2, at the URL
370 <ulink url="http://localhost:9004/pazpar2-demo/"/>
374 From here, the demo applications can be
375 accessed: <literal>test1</literal>, <literal>test2</literal> and
376 <literal>jsdemo</literal>
377 are pure HTML+JavaScript setups, needing no server-side
379 <literal>demo</literal>
380 requires PHP on the server.
383 If you don't see the test interfaces, check whether they are available
384 on port 80 (i.e. directly from the Apache2 server). If not, the
385 Apache configuration is incorrect.
388 In order to use Apache as frontend for the interface on port 80
389 for public access etc., refer to
390 <xref linkend="installation.apache2proxy"/>.
394 <section id="installation.debian">
395 <title>Installation on Debian GNU/Linux and Ubuntu</title>
397 Index Data provides Debian and Ubuntu packages for Pazpar2.
398 As of February 2010, these
399 are prepared for Debian versions Etch, Lenny and Squeeze; and for
400 Ubuntu versions 8.04 (hardy), 8.10 (intrepid), 9.04 (jaunty) and
401 9.10 (karmic). These packages are available at
402 <ulink url="&url.pazpar2.download.debian;"/> and
403 <ulink url="&url.pazpar2.download.ubuntu;"/>.
407 <section id="installation.apache2proxy">
408 <title>Apache 2 Proxy</title>
412 url="http://httpd.apache.org/docs/2.2/mod/mod_proxy.html">
415 which allows Pazpar2 to become a backend to an Apache 2
416 based web service. The Apache 2 proxy must operate in the
417 <emphasis>Reverse</emphasis> Proxy mode.
421 On a Debian based Apache 2 system, the relevant modules can
424 sudo a2enmod proxy_http proxy_balancer
429 Traditionally Pazpar2 interprets URL paths with suffix
430 <literal>/search.pz2</literal>.
433 url="http://httpd.apache.org/docs/2.2/mod/mod_proxy.html#proxypass">
436 directive of Apache must be used to map a URL path
437 the the Pazpar2 server (listening port).
442 The ProxyPass directive takes a prefix rather than
443 a suffix as URL path. It is important that the Java Script code
444 uses the prefix given for it.
448 <example id="installation.apache2proxy.example">
449 <title>Apache 2 proxy configuration</title>
451 If Pazpar2 is running on port 8004 and the portal is using
452 <filename>search.pz2</filename> inside portal in directory
453 <filename>/myportal/</filename> we could use the following
454 Apache 2 configuration:
457 <IfModule mod_proxy.c>
461 AddDefaultCharset off
466 ProxyPass /myportal/search.pz2 http://localhost:8004/search.pz2
477 <title>Using Pazpar2</title>
479 This chapter provides a general introduction to the use and
480 deployment of Pazpar2.
483 <section id="architecture">
484 <title>Pazpar2 and your systems architecture</title>
486 Pazpar2 is designed to provide asynchronous, behind-the-scenes
487 metasearching functionality to your application, exposing this
488 functionality using a simple webservice API that can be accessed
489 from any number of development environments. In particular, it is
490 possible to combine Pazpar2 either with your server-side dynamic
491 website scripting, with scripting or code running in the browser, or
492 with any combination of the two. Pazpar2 is an excellent tool for
493 building advanced, Ajax-based user interfaces for metasearch
494 functionality, but it isn't a requirement -- you can choose to use
495 Pazpar2 entirely as a backend to your regular server-side scripting.
496 When you do use Pazpar2 in conjunction
497 with browser scripting (JavaScript/Ajax, Flash, applets,
498 etc.), there are special considerations.
502 Pazpar2 implements a simple but efficient HTTP server, and it is
503 designed to interact directly with scripting running in the browser
504 for the best possible performance, and to limit overhead when
505 several browser clients generate numerous webservice requests.
506 However, it is still desirable to use a conventional webserver,
507 such as Apache, to serve up graphics, HTML documents, and
508 server-side scripting. Because the security sandbox environment of
509 most browser-side programming environments only allows communication
510 with the server from which the enclosing HTML page or object
511 originated, Pazpar2 is designed so that it can act as a transparent
512 proxy in front of an existing webserver (see <xref
513 linkend="pazpar2_conf"/> for details).
514 In this mode, all regular
515 HTTP requests are transparently passed through to your webserver,
516 while Pazpar2 only intercepts search-related webservice requests.
520 If you want to expose your combined service on port 80, you can
521 either run your regular webserver on a different port, a different
522 server, or a different IP address associated with the same server.
526 Pazpar2 can also work behind
527 a reverse Proxy. Refer to <xref linkend="installation.apache2proxy"/>)
528 for more information.
529 This allows your existing HTTP server to operate on port 80 as usual.
530 Pazpar2 can be started on another (internal) port.
534 Sometimes, it may be necessary to implement functionality on your
535 regular webserver that makes use of search results, for example to
536 implement data import functionality, emailing results, history
537 lists, personal citation lists, interlibrary loan functionality,
538 etc. Fortunately, it is simple to exchange information between
539 Pazpar2, your browser scripting, and backend server-side scripting.
540 You can send a session ID and possibly a record ID from your browser
541 code to your server code, and from there use Pazpar2s webservice API
542 to access result sets or individual records. You could even 'hide'
543 all of Pazpar2s functionality between your own API implemented on
544 the server-side, and access that from the browser or elsewhere. The
545 possibilities are just about endless.
549 <section id="data_model">
550 <title>Your data model</title>
552 Pazpar2 does not have a preconceived model of what makes up a data
553 model. There are no assumptions that records have specific fields or
554 that they are organized in any particular way. The only assumption
555 is that data comes packaged in a form that the software can work
556 with (presently, that means XML or MARC), and that you can provide
557 the necessary information to massage it into Pazpar2's internal
562 Handling retrieval records in Pazpar2 is a two-step process. First,
563 you decide which data elements of the source record you are
564 interested in, and you specify any desired massaging or combining of
565 elements using an XSLT stylesheet (MARC records are automatically
566 normalized to <ulink url="&url.marcxml;">MARCXML</ulink> before this step).
567 If desired, you can run multiple XSLT stylesheets in series to accomplish
568 this, but the output of the last one should be a representation of the
569 record in a schema that Pazpar2 understands.
573 The intermediate, internal representation of the record looks like
576 <record xmlns="http://www.indexdata.com/pazpar2/1.0"
577 mergekey="title The Shining author King, Stephen">
579 <metadata type="title" rank="2">The Shining</metadata>
581 <metadata type="author">King, Stephen</metadata>
583 <metadata type="kind">ebook</metadata>
584 <!-- ... and so on -->
588 As you can see, there isn't much to it. There are really only a few
589 important elements to this file.
593 Elements should belong to the namespace
594 <literal>http://www.indexdata.com/pazpar2/1.0</literal>.
595 If the root node contains the
596 attribute 'mergekey', then every record that generates the same
597 merge key (normalized for case differences, white space, and
598 truncation) will be joined into a cluster. In other words, you
599 decide how records are merged. If you don't include a merge key,
600 records are never merged. The 'metadata' elements provide the meat
601 of the elements -- the content. the 'type' attribute is used to
602 match each element against processing rules that determine what
603 happens to the data element next. The attribute, 'rank' specifies
604 specifies a multipler for ranking for this element.
608 The next processing step is the extraction of metadata from the
609 intermediate representation of the record. This is governed by the
610 'metadata' elements in the 'service' section of the configuration
611 file. See <xref linkend="config-server"/> for details. The metadata
612 in the retrieval record ultimately drives merging, sorting, ranking,
613 the extraction of browse facets, and display, all configurable.
617 Pazpar2 1.6.37 and later also allows already clustered records to
618 be ingested. Suppose a database already clusters for us and we would like
619 to keep that cluster for Pazpar2. In that case we can generate a
620 <literal>cluster</literal> wrapper element that holds individual
621 <literal>record</literal> elements.
624 Cluster record example:
626 <cluster xmlns="http://www.indexdata.com/pazpar2/1.0">
628 <metadata type="title" rank="2">The Shining</metadata>
629 <metadata type="author">King, Stephen</metadata>
630 <metadata type="kind">ebook</metadata>
633 <metadata type="title" rank="2">The Shining</metadata>
634 <metadata type="author">King, Stephen</metadata>
635 <metadata type="kind">audio</metadata>
642 <section id="client">
643 <title>Client development overview</title>
645 You can use Pazpar2 from any environment that allows you to use
646 webservices. The initial goal of the software was to support
647 Ajax-based applications, but there literally are no limits to what
648 you can do. You can use Pazpar2 from Javascript, Flash, Java, etc.,
649 on the browser side, and from any development environment on the
650 server side, and you can pass session tokens and record IDs freely
651 around between these environments to build sophisticated applications.
652 Use your imagination.
656 The webservice API of Pazpar2 is described in detail in <xref
657 linkend="pazpar2_protocol"/>.
661 In brief, you use the 'init' command to create a session, a
662 temporary workspace which carries information about the current
663 search. You start a new search using the 'search' command. Once the
664 search has been started, you can follow its progress using the
665 'stat', 'bytarget', 'termlist', or 'show' commands. Detailed records
666 can be fetched using the 'record' command.
672 <section id="unicode">
673 <title>Unicode Compliance</title>
675 Pazpar2 is Unicode compliant and language and locale aware but relies
676 on character encoding for the targets to be specified correctly if
677 the targets themselves are not UTF-8 based (most aren't).
678 Just a few bad behaving targets can spoil the search experience
679 considerably if for example Greek, Russian or otherwise non 7-bit ASCII
680 search terms are entered. In these cases some targets return
681 records irrelevant to the query, and the result screens will be
682 cluttered with noise.
685 While noise from misbehaving targets can not be removed, it can
686 be reduced using truly Unicode based ranking. This is an
687 option which is available to the system administrator if ICU
688 support is compiled into YAZ, see
689 <xref linkend="installation"/> for details.
692 In addition, the ICU tokenization and normalization rules must
693 be defined in the master configuration file described in
694 <xref linkend="config-server"/>.
698 <section id="load_balancing">
699 <title>Load balancing</title>
701 Just like any web server, Pazpar2, can be load balanced by a standard
702 hardware or software load balancer as long as the session stickiness
703 is ensured. If you are already running the Apache2 web server in front
704 of Pazpar2 and use the apache mod_proxy module to 'relay' client
705 requests to Pazpar2, this set up can be easily extended to include
706 load balancing capabilites.
707 To do so you need to enable the
708 <ulink url="http://httpd.apache.org/docs/2.2/mod/mod_proxy_balancer.html">
711 module in your Apache2 installation.
715 On a Debian based Apache 2 system, the relevant modules can
718 sudo a2enmod proxy_http
723 The mod_proxy_balancer can pass all 'sessionsticky' requests to the
724 same backend worker as long as the requests are marked with the
725 originating worker's ID (called 'route'). If the Pazpar2 serverID is
726 configured (by setting an 'id' attribute on the 'server' element in
727 the Pazpar2 configuration file) Pazpar2 will append it to the
728 'session' element returned during the 'init' in a mod_proxy_balancer
730 Since the 'session' is then re-sent by the client (for all pazpar2
731 request besides 'init'), the balancer can use the marker to pass
732 the request to the right route. To do so the balancer needs to be
733 configured to inspect the 'session' parameter.
736 <example id="load_balancing.example">
737 <title>Apache 2 load balancing configuration</title>
739 Having 4 Pazpar2 instances running on the same host, port range of
740 8004-8007 and serverIDs of: pz1, pz2, pz3 and pz4 respectively we
741 could use the following Apache 2 configuration to expose a single
742 pazpar2 'endpoint' on a standard
743 (<filename>/pazpar2/search.pz2</filename>) location:
747 AddDefaultCharset off
753 # 'route' has to match the configured pazpar2 server ID
754 <Proxy balancer://pz2cluster>
755 BalancerMember http://localhost:8004 route=pz1
756 BalancerMember http://localhost:8005 route=pz2
757 BalancerMember http://localhost:8006 route=pz3
758 BalancerMember http://localhost:8007 route=pz4
761 # route is resent in the 'session' param which has the form:
762 # 'sessid.serverid', understandable by the mod_proxy_load_balancer
763 # this is not going to work if the client tampers with the 'session' param
764 ProxyPass /pazpar2/search.pz2 balancer://pz2cluster lbmethod=byrequests stickysession=session nofailover=On
767 The 'ProxyPass' line sets up a reverse proxy for request
768 ‘/pazpar2/search.pz2’ and delegates all requests to the load balancer
769 (virtual worker) with name ‘pz2cluster’.
770 Sticky sessions are enabled and implemented using the ‘session’ parameter.
771 The ‘Proxy’ section lists all the servers (real workers) which the
772 load balancer can use.
779 <section id="relevance_ranking">
780 <title>Relevance ranking</title>
782 Pazpar2 uses a variant of the fterm frequency–inverse document frequency
783 (Tf-idf) ranking algorithm.
786 The Tf-part is straightforward to calculate and is based on the
787 documents that Pazpar2 fetches. The idf-part, however, is more tricky
788 since the corpus at hand is ONLY the relevant documents and not
789 irrelevant ones. Pazpar2 does not have the full corpus -- only the
790 documents that match a particular search.
793 Computatation of the Tf-part is based on the normalized documents.
794 The length, the position and terms are thus normalized at this point.
795 Also the computation if performed for each document received from the
796 target - before merging takes place. The result of a TF-compuation is
797 added to the TF-total of a cluster. Thus, if a document occurs twice,
798 then the TF-part is doubled. That, however, can be adjusted, because the
799 TF-part may be divided by the number of documents in a cluster.
802 The algorithm used by Pazpar2 has two phases. In phase one
803 Pazpar2 computes a tf-array .. This is being done as records are
804 fetched form the database. In this case, the rank weigth
805 <literal>w</literal>, the and rank tweaks <literal>lead</literal>,
806 <literal>follow</literal> and <literal>length</literal>.
811 foreach document in a cluster
814 for i = 1, .. N: (each term)
815 foreach pos (where term i occurs in field)
816 // w is configured weight for field
817 // pos is position of term in field
818 w[i] += w / (1 + log2(1+lead*pos))
820 w[i] += w[i] * follow / (1+log2(d)
821 // length: length of field (number of terms that is)
822 if (length strategy is "linear")
823 tf[i] += w[i] / length;
824 else if (length strategy is "log")
825 tf[i] += w[i] / log2(length);
826 else if (length strategy is "none")
830 In phase two, the idf-array is computed and the final score
831 is computed. This is done for each cluster as part of each show command.
832 The rank tweak <literal>cluster</literal> is in use here.
835 // dococcur[i]: number of records where term occurs
836 // doctotal: number of records
837 for i = 1, .., N (each term)
839 idf[i] = log(1 + doctotal / dococcur[i])
844 for i = 1, .., N: (each term)
845 if (cluster is "yes")
846 tf[i] = tf[i] / cluster_size;
847 relevance += 100000 * tf[i] / idf[i];
850 For controlling the ranking parameters, refer to the
851 <link linkend="service-rank">rank</link> element of the
853 Refer to the <link linkend="metadata-rank">rank</link> attribute
854 of the metadata element for how to control ranking for individual
857 </section> <!-- relevance_ranking -->
859 <section id="masterkey_connect">
860 <title>Pazpar2 and MasterKey Connect</title>
862 MasterKey Connect is a hosted connector, or gateway, service that exposes
863 whatever searchable resources you need. Since the service exposes all
864 resources using Z39.50 (or SRU), it is easy to set up Pazpar2 to use the
865 service. In particular, since all connectors expose basically the same core
866 behavior, it is a good use of Pazpar2's mechanism for managing default
867 behaviors across similar databases.
870 After installation of Pazpar2, the directory
871 <filename>/etc/pazpar2/settings/mkc</filename> (location may
872 vary depending on installation preferences) contains an example setup that
873 searches two different resources through a MasterKey Connect demo account.
874 The file mkc.xml contains default parameters that will work for all
875 MasterKey Connect resources (if you decide to become a customer of the
876 service, you will substitute your own account credentials for
877 the guest/guest). The other files contain specific information about
878 a couple of demonstration resources.
882 To play with the demo, just create a symlink from
883 <filename>/etc/pazpar2/services-enabled/default.xml</filename>
884 to <filename>/etc/pazpar2/services-available/mkc.xml</filename>.
885 And restart Pazpar2. You should now be able to search the two demo
886 resources using JSDemo or any user interface of your choice.
887 If you are interested in learning more about MasterKey Connect, or to
888 try out the service for free against your favorite online resource, just
889 contact us at <email>info@indexdata.com</email>.
893 </chapter> <!-- Using Pazpar2 -->
895 <reference id="reference">
896 <title>Reference</title>
897 <partintro id="reference-introduction">
899 The material in this chapter is drawn directly from the individual
906 <appendix id="license">
907 <title>License</title>
911 Copyright © ©right-year; Index Data.
915 Pazpar2 is free software; you can redistribute it and/or modify it under
916 the terms of the GNU General Public License as published by the Free
917 Software Foundation; either version 2, or (at your option) any later
922 Pazpar2 is distributed in the hope that it will be useful, but WITHOUT ANY
923 WARRANTY; without even the implied warranty of MERCHANTABILITY or
924 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
929 You should have received a copy of the GNU General Public License
930 along with Pazpar2; see the file LICENSE. If not, write to the
931 Free Software Foundation,
932 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
941 <!-- Keep this comment at the end of the file