Doc: relevance section refer to config description

[pazpar2-moved-to-github.git] / doc / book.xml

<?xml version="1.0" standalone="no"?>
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
    "http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd"
[
     <!ENTITY % local SYSTEM "local.ent">
     %local;
     <!ENTITY % entities SYSTEM "entities.ent">
     %entities;
     <!ENTITY % idcommon SYSTEM "common/common.ent">
     %idcommon;
]>
<book id="book">
 <bookinfo>
  <title>Pazpar2 - User's Guide and Reference</title>
  <author>
   <firstname>Sebastian</firstname><surname>Hammer</surname>
  </author>
  <author>
   <firstname>Adam</firstname><surname>Dickmeiss</surname>
  </author>
  <author>
   <firstname>Marc</firstname><surname>Cromme</surname>
  </author>
  <author>
   <firstname>Jakub</firstname><surname>Skoczen</surname>
  </author>
  <author>
   <firstname>Mike</firstname><surname>Taylor</surname>
  </author>
  <author>
   <firstname>Dennis</firstname><surname>Schafroth</surname>
  </author>
  <releaseinfo>&version;</releaseinfo>
  <copyright>
   <year>&copyright-year;</year>
   <holder>Index Data</holder>
  </copyright>
  <abstract>
   <simpara>
    Pazpar2 is a high-performance metasearch engine featuring
    merging, relevance ranking, record sorting,
    and faceted results.
    It is middleware: it has no user interface of its own, but can be
    configured and controlled by an XML-over-HTTP web-service to provide
    metasearching functionality behind any user interface.
   </simpara>
   <simpara>
    This document is a guide and reference to Pazpar2 version &version;.
   </simpara>
   <simpara>
    <inlinemediaobject>
     <imageobject>
      <imagedata fileref="common/id.png" format="PNG"/>
     </imageobject>
     <imageobject>
      <imagedata fileref="common/id.eps" format="EPS"/>
     </imageobject>
    </inlinemediaobject>
   </simpara>
  </abstract>
 </bookinfo>

 <chapter id="introduction">
  <title>Introduction</title>

  <section id="what.pazpar2.is">
   <title>What Pazpar2 is</title>
   <para>
    Pazpar2 is a stand-alone metasearch engine with a web-service API, designed
    to be used either from a browser-based client (JavaScript, Flash,
    Java applet,
    etc.), from server-side code, or any combination of the two.
    Pazpar2 is a highly optimized client designed to
    search many resources in parallel. It implements record merging,
    relevance-ranking and sorting by arbitrary data content, and facet
    analysis for browsing purposes. It is designed to be data-model
    independent, and is capable of working with MARC, DublinCore, or any
    other <ulink url="&url.xml;">XML</ulink>-structured response format
    -- <ulink url="&url.xslt;">XSLT</ulink> is used to normalize and extract
    data from retrieval records for display and analysis. It can be used
    against any server which supports the
    <ulink url="&url.z39.50;">Z39.50</ulink>, <ulink url="&url.sru;">SRU/SRW</ulink>
    or <ulink url="&url.solr;">SOLR</ulink> protocol. Proprietary
    backend modules can function as connectors between these standard
    protocols and any non-standard API, including web-site scraping, to
    support a large number of other protocols.
   </para>
   <para>
    Additional functionality such as
    user management and attractive displays are expected to be implemented by
    applications that use Pazpar2. Pazpar2 itself is user-interface independent.
    Its functionality is exposed through a simple XML-based web-service API,
    designed to be easy to use from an Ajax-enabled browser, Flash
    animation, Java applet, etc., or from a higher-level server-side language
    like PHP, Perl or Java. Because session information can be shared between
    browser-based logic and server-side scripting, there is tremendous
    flexibility in how you implement application-specific logic on top
    of Pazpar2.
   </para>
   <para>
    Once you launch a search in Pazpar2, the operation continues behind the
    scenes. Pazpar2 connects to servers, carries out searches, and
    retrieves, deduplicates, and stores results internally. Your application
    code may periodically inquire about the status of an ongoing operation,
    and ask to see records or result set facets. Results become
    available immediately, and it is easy to build end-user interfaces than
    feel extremely responsive, even when searching more than 100 servers
    concurrently.
   </para>
   <para>
    Pazpar2 is designed to be highly configurable. Incoming records are
    normalized to XML/UTF-8, and then further normalized using XSLT to a
    simple internal representation that is suitable for analysis. By
    providing XSLT stylesheets for different kinds of result records, you
    can configure Pazpar2 to work against different kinds of information
    retrieval servers. Finally, metadata is extracted in a configurable
    way from this internal record, to support display, merging, ranking,
    result set facets, and sorting. Pazpar2 is not bound to a specific model
    of metadata, such as DublinCore or MARC: by providing the right
    configuration, it can work with any combination of different kinds of data
    in support of many different applications.
   </para>
   <para>
    Pazpar2 is designed to be efficient and scalable. You can set it up to
    search several hundred targets in parallel, or you can use it to support
    hundreds of concurrent users. It is implemented with the same attention
    to performance and economy that we use in our indexing engines, so that
    you can focus on building your application without worrying about the
    details of metasearch logic. You can devote all of your attention to
    usability and let Pazpar2 do what it does best -- metasearch.
   </para>
   <para>
    Pazpar2 is our attempt to re-think the traditional paradigms for
    implementing and deploying metasearch logic, with an uncompromising
    approach to performance, and attempting to make maximum use of the
    capabilities of modern browsers. The demo user interface that
    accompanies the distribution is but one example. If you think of new
    ways of using Pazpar2, we hope you'll share them with us, and if we
    can provide assistance with regards to training, design, programming,
    integration with different backends, hosting, or support, please don't
    hesitate to contact us. If you'd like to see functionality in Pazpar2
    that is not there today, please don't hesitate to contact us. It may
    already be in our development pipeline, or there might be a
    possibility for you to help out by sponsoring development time or
    code. Either way, get in touch and we will give you straight answers.
   </para>
   <para>
    Enjoy!
   </para>
   <para>
    Pazpar2 is covered by the GNU General Public License (GPL) version 2.
    See <xref linkend="license"/> for further information.
   </para>
  </section>

  <section id="connectors">
   <title>Connectors to non-standard databases</title>
   <para>
    If you need to access commercial or open access resources that don't support
    Z39.50 or SRU, one approach would be to use a tool like <ulink
    url="&url.simpleserver;">SimpleServer</ulink> to build a
    gateway. An easier option is to use Index Data's <ulink
    url="&url.mkc;">MasterKey Connect</ulink>
    service, which will expose virtually <emphasis>any</emphasis> resource
    through Z39.50/SRU, dead easy to integrate with Pazpar2.
    The service is hosted, so all you have to do is to let us
    know which resources you are interested in, and we operate the gateways,
    or Connectors for you for a low annual charge.
    Types of resources supported include
    commercial databases, free online resources, and even local resources;
    almost anything that can be accessed through a web-facing user
    interface can be accessed in this way.
    Contact <email>info@indexdata.com</email> for more information.
    See <xref linkend="masterkey_connect"/> for an example.
   </para>
  </section>

  <section id="name">
   <title>A note on the name Pazpar2</title>
   <para>
    The name Pazpar2 derives from three sources.  One one hand, it is
    Index Data's second major piece of software that does parallel
    searching of Z39.50 targets.  On the other, it is a near-homophone
    of Passpartout, the ever-helpful servant in Jules Verne's novel
    Around the World in Eighty Days (who helpfully uses the language
    of his master).  Finally, "passe par tout" means something like
    "passes through anything" in French -- on other words, a universal
    solution, or if you like a MasterKey.
   </para>
  </section>
 </chapter>

 <chapter id="installation">
  <title>Installation</title>
  <para>
   The Pazpar2 package includes documentation as well
   as the Pazpar2 server. The package also includes a simple user
   interface called "test1", which consists of a single HTML page and a single
   JavaScript file to illustrate the use of Pazpar2.
  </para>
  <para>
   Pazpar2 depends on the following tools/libraries:
   <variablelist>
    <varlistentry><term><ulink url="&url.yaz;">YAZ</ulink></term>
    <listitem>
     <para>
      The popular Z39.50 toolkit for the C language.
      YAZ <emphasis>must</emphasis> be compiled with Libxml2/Libxslt support.
     </para>
    </listitem>
    </varlistentry>
    <varlistentry><term><ulink url="&url.icu;">International
    Components for Unicode (ICU)</ulink></term>
    <listitem>
     <para>
      ICU provides Unicode support for non-English languages with
      character sets outside the range of 7bit ASCII, like
      Greek, Russian, German and French. Pazpar2 uses the ICU
      Unicode character conversions, Unicode normalization, case
      folding and other fundamental operations needed in
      tokenization, normalization and ranking of records.
     </para>
     <para>
      Compiling, linking, and usage of the ICU libraries is optional,
      but strongly recommended for usage in an international
      environment.
     </para>
    </listitem>
    </varlistentry>
   </variablelist>
  </para>
  <para>
   In order to compile Pazpar2, a C compiler which supports C99 or later
   is required.
  </para>

  <section id="installation.unix">
   <title>Installation from source on Unix (including Linux, MacOS, etc.)</title>
   <para>
    The latest source code for Pazpar2 is available from
    <ulink url="&url.pazpar2.download;"/>.
    Most Unix-based operating systems have the required
    tools available as binary packages.
    For example, if Libxml2/libXSLT libraries
    are already installed as development packages, use these.
   </para>

   <para>
    Ensure that the development libraries and header files are
    available on your system before compiling Pazpar2. For installation
    of YAZ, refer to the Installation chapter of the YAZ manual at
    <ulink url="&url.yaz.install;"/>.
   </para>
   <para>
    Once the dependencies are in place, Pazpar2 can be unpacked and
    installed as follows:
   </para>
   <screen>
    tar xzf pazpar2-VERSION.tar.gz
    cd pazpar2-VERSION
    ./configure
    make
    sudo make install
   </screen>
   <para>
    The <literal>make install</literal> will install manpages as well as the
    Pazpar2 server, <literal>pazpar2</literal>,
    in PREFIX<literal>/sbin</literal>.
    By default, PREFIX is <literal>/usr/local/</literal> . This can be
    changed with configure option <option>--prefix</option>.
   </para>
  </section>

  <section id="installation.win32">
   <title>Installation from source on Windows</title>
   <para>
    Pazpar2 can be built for Windows using
    <ulink url="&url.vstudio;">Microsoft Visual Studio</ulink>.
    The support files for building YAZ on Windows are located in the
    <filename>win</filename> directory. The compilation is performed
    using the <filename>win/makefile</filename> which is to be
    processed by the NMAKE utility part of Visual Studio.
   </para>
   <para>
    Ensure that the development libraries and header files are
    available on your system before compiling Pazpar2. For installation
    of YAZ, refer to
    the Installation chapter of the YAZ manual at
    <ulink url="&url.yaz.install;"/>.
    It is easiest if YAZ and Pazpar2 are unpacked in the same
    directory (side-by-side).
   </para>
   <para>
    The compilation is tuned by editing the makefile of Pazpar2.
    The process is similar to YAZ. Adjust the various directories
    <literal>YAZ_DIR</literal>, <literal>ZLIB_DIR</literal>, etc.,
    as required.
   </para>
   <para>
    Compile Pazpar2 by invoking <application>nmake</application> in
    the <filename>win</filename> directory.
    The resulting binaries of the build process are located in the
    <filename>bin</filename> of the Pazpar2 source
    tree - including the <filename>pazpar2.exe</filename> and necessary DLLs.
   </para>
   <para>
    The Windows version of Pazpar2 is a console application. It may
    be installed as a Windows Service by adding option
    <literal>-install</literal> for the pazpar2 program. This will
    register Pazpar2 as a service and use the other options provided
    in the same invocation. For example:
    <screen>
     cd \MyPazpar2\etc
     ..\bin\pazpar2 -install -f pazpar2.cfg -l pazpar2.log
    </screen>
    The Pazpar2 service may now be controlled via the Service Control
    Panel. It may be unregistered by passing the <literal>-remove</literal>
    option. Example:
    <screen>
     cd \MyPazpar2\etc
     ..\bin\pazpar2 -remove
    </screen>
   </para>
  </section>

  <section id="installation.test1">
   <title>Installation of test interfaces</title>
   <para>
    In this section we show how to make available the set of simple
    interfaces that are part of the Pazpar2 source package, and which
    demonstrate some ways to use Pazpar2.  (Note that Debian users can
    save time by just installing the package <literal>pazpar2-test1</literal>.)
   </para>
   <para>
    A web server, such as Apache, must be installed and running on the system.
   </para>

   <para>
    Start the Pazpar2 daemon using the 'in-source' binary of the Pazpar2
    daemon. On Unix the process is:
    <screen>
     cd etc
     cp pazpar2.cfg.dist pazpar2.cfg
     ../src/pazpar2 -f pazpar2.cfg
    </screen>
    And on Windows:
    <screen>
     cd etc
     copy pazpar2.cfg.dist pazpar2.cfg
     ..\bin\pazpar2 -f pazpar2.cfg
    </screen>
    This will start a Pazpar2 listener on port 9004. It will proxy
    HTTP requests to port 80 on localhost, which we assume will be the regular
    HTTP server on the system. Inspect and modify pazpar2.cfg as needed
    if this is to be changed. The pazpar2.cfg file includes settings from the
    file <filename>settings/edu.xml</filename>
    to use for searches.
   </para>

   <para>
    The test UIs are located in <literal>www</literal>. Ensure that this
    directory is available to the web server by copying
    <literal>www</literal> to the document root,
    using Apache's <literal>Alias</literal> directive, or
    creating a symbolic link: for example, on a Debian or Ubuntu
    system with Apache2 installed from the standard package, you might
    make the link as follows:
    <screen>
     cd .../pazpar2
     sudo ln -s `pwd`/www /var/www/pazpar2-demo
    </screen>
   </para>

   <para>
    This makes the test applications visible at
    <ulink url="http://localhost/pazpar2-demo/"/>
    but they can not be run successfully from that URL, as they submit
    search requests back to the server form which they were served,
    and Apache2 doesn't know how to handle them.  Instead, the test
    applications must be accessed from Pazpar2 itself, acting as a
    proxy to Apache2, at the URL
    <ulink url="http://localhost:9004/pazpar2-demo/"/>
   </para>

   <para>
    From here, the demo applications can be
    accessed: <literal>test1</literal>, <literal>test2</literal> and
    <literal>jsdemo</literal>
    are pure HTML+JavaScript setups, needing no server-side
    intelligence;
    <literal>demo</literal>
    requires PHP on the server.
   </para>
   <para>
    If you don't see the test interfaces, check whether they are available
    on port 80 (i.e. directly from the Apache2 server).  If not, the
    Apache configuration is incorrect.
   </para>
   <para>
    In order to use Apache as frontend for the interface on port 80
    for public access etc., refer to
    <xref linkend="installation.apache2proxy"/>.
   </para>
  </section>

  <section id="installation.debian">
   <title>Installation on Debian GNU/Linux and Ubuntu</title>
   <para>
    Index Data provides Debian and Ubuntu packages for Pazpar2.
    As of February 2010, these
    are prepared for Debian versions Etch, Lenny and Squeeze; and for
    Ubuntu versions 8.04 (hardy), 8.10 (intrepid), 9.04 (jaunty) and
    9.10 (karmic).  These packages are available at
    <ulink url="&url.pazpar2.download.debian;"/> and
    <ulink url="&url.pazpar2.download.ubuntu;"/>.
   </para>
  </section>

  <section id="installation.apache2proxy">
   <title>Apache 2 Proxy</title>
   <para>
    Apache 2 has a
    <ulink
        url="http://httpd.apache.org/docs/2.2/mod/mod_proxy.html">
     proxy module
    </ulink>
    which allows Pazpar2 to become a backend to an Apache 2
    based web service. The Apache 2 proxy must operate in the
    <emphasis>Reverse</emphasis> Proxy mode.
   </para>

   <para>
    On a Debian based Apache 2 system, the relevant modules can
    be enabled with:
    <screen>
     sudo a2enmod proxy_http proxy_balancer
    </screen>
   </para>

   <para>
    Traditionally Pazpar2 interprets URL paths with suffix
    <literal>/search.pz2</literal>.
    The
    <ulink
        url="http://httpd.apache.org/docs/2.2/mod/mod_proxy.html#proxypass">
    ProxyPass
    </ulink>
    directive of Apache must be used to map a URL path
    the the Pazpar2 server (listening port).
   </para>

   <note>
    <para>
     The ProxyPass directive takes a prefix rather than
     a suffix as URL path. It is important that the Java Script code
     uses the prefix given for it.
    </para>
   </note>

   <example id="installation.apache2proxy.example">
    <title>Apache 2 proxy configuration</title>
    <para>
     If Pazpar2 is running on port 8004 and the portal is using
     <filename>search.pz2</filename> inside portal in directory
     <filename>/myportal/</filename> we could use the following
     Apache 2 configuration:

     <screen><![CDATA[
      <IfModule mod_proxy.c>
       ProxyRequests Off

       <Proxy *>
        AddDefaultCharset off
        Order deny,allow
        Allow from all
       </Proxy>

       ProxyPass /myportal/search.pz2 http://localhost:8004/search.pz2
       ProxyVia Off
      </IfModule>
      ]]></screen>
    </para>
   </example>
  </section>

 </chapter>

 <chapter id="using">
  <title>Using Pazpar2</title>
  <para>
   This chapter provides a general introduction to the use and
   deployment of Pazpar2.
  </para>

  <section id="architecture">
   <title>Pazpar2 and your systems architecture</title>
   <para>
    Pazpar2 is designed to provide asynchronous, behind-the-scenes
    metasearching functionality to your application, exposing this
    functionality using a simple webservice API that can be accessed
    from any number of development environments. In particular, it is
    possible to combine Pazpar2 either with your server-side dynamic
    website scripting, with scripting or code running in the browser, or
    with any combination of the two. Pazpar2 is an excellent tool for
    building advanced, Ajax-based user interfaces for metasearch
    functionality, but it isn't a requirement -- you can choose to use
    Pazpar2 entirely as a backend to your regular server-side scripting.
    When you do use Pazpar2 in conjunction
    with browser scripting (JavaScript/Ajax, Flash, applets,
    etc.), there are    special considerations.
   </para>

   <para>
    Pazpar2 implements a simple but efficient HTTP server, and it is
    designed to interact directly with scripting running in the browser
    for the best possible performance, and to limit overhead when
    several browser clients generate numerous webservice requests.
    However, it is still desirable to use a conventional webserver,
    such as Apache, to serve up graphics, HTML documents, and
    server-side scripting. Because the security sandbox environment of
    most browser-side programming environments only allows communication
    with the server from which the enclosing HTML page or object
    originated, Pazpar2 is designed so that it can act as a transparent
    proxy in front of an existing webserver (see <xref
    linkend="pazpar2_conf"/> for details).
    In this mode, all regular
    HTTP requests are transparently passed through to your webserver,
    while Pazpar2 only intercepts search-related webservice requests.
   </para>

   <para>
    If you want to expose your combined service on port 80, you can
    either run your regular webserver on a different port, a different
    server, or a different IP address associated with the same server.
   </para>

   <para>
    Pazpar2 can also work behind
    a reverse Proxy. Refer to <xref linkend="installation.apache2proxy"/>)
    for more information.
    This allows your existing HTTP server to operate on port 80 as usual.
    Pazpar2 can be started on another (internal) port.
   </para>

   <para>
    Sometimes, it may be necessary to implement functionality on your
    regular webserver that makes use of search results, for example to
    implement data import functionality, emailing results, history
    lists, personal citation lists, interlibrary loan functionality,
    etc. Fortunately, it is simple to exchange information between
    Pazpar2, your browser scripting, and backend server-side scripting.
    You can send a session ID and possibly a record ID from your browser
    code to your server code, and from there use Pazpar2s webservice API
    to access result sets or individual records. You could even 'hide'
    all of Pazpar2s functionality between your own API implemented on
    the server-side, and access that from the browser or elsewhere. The
    possibilities are just about endless.
   </para>
  </section>

  <section id="data_model">
   <title>Your data model</title>
   <para>
    Pazpar2 does not have a preconceived model of what makes up a data
    model. There are no assumptions that records have specific fields or
    that they are organized in any particular way. The only assumption
    is that data comes packaged in a form that the software can work
    with (presently, that means XML or MARC), and that you can provide
    the necessary information to massage it into Pazpar2's internal
    record abstraction.
   </para>

   <para>
    Handling retrieval records in Pazpar2 is a two-step process. First,
    you decide which data elements of the source record you are
    interested in, and you specify any desired massaging or combining of
    elements using an XSLT stylesheet (MARC records are automatically
    normalized to <ulink url="&url.marcxml;">MARCXML</ulink> before this step).
    If desired, you can run multiple XSLT stylesheets in series to accomplish
    this, but the output of the last one should be a representation of the
    record in a schema that Pazpar2 understands.
   </para>

   <para>
    The intermediate, internal representation of the record looks like
    this:
    <screen><![CDATA[
     <record xmlns="http://www.indexdata.com/pazpar2/1.0"
       mergekey="title The Shining author King, Stephen">

       <metadata type="title" rank="2">The Shining</metadata>

       <metadata type="author">King, Stephen</metadata>

       <metadata type="kind">ebook</metadata>
       <!-- ... and so on -->
     </record>
]]></screen>

    As you can see, there isn't much to it. There are really only a few
    important elements to this file.
   </para>

   <para>
    Elements should belong to the namespace
    <literal>http://www.indexdata.com/pazpar2/1.0</literal>.
    If the root node contains the
    attribute 'mergekey', then every record that generates the same
    merge key (normalized for case differences, white space, and
    truncation) will be joined into a cluster. In other words, you
    decide how records are merged. If you don't include a merge key,
    records are never merged. The 'metadata' elements provide the meat
    of the elements -- the content. the 'type' attribute is used to
    match each element against processing rules that determine what
    happens to the data element next. The attribute, 'rank' specifies
    specifies a multipler for ranking for this element.
   </para>

   <para>
    The next processing step is the extraction of metadata from the
    intermediate representation of the record. This is governed by the
    'metadata' elements in the 'service' section of the configuration
    file. See <xref linkend="config-server"/> for details. The metadata
    in the retrieval record ultimately drives merging, sorting, ranking,
    the extraction of browse facets, and display, all configurable.
   </para>

   <para>
    Pazpar2 1.6.37 and later also allows already clustered records to
    be ingested. Suppose a database already clusters for us and we would like
    to keep that cluster for Pazpar2. In that case we can generate a
    <literal>cluster</literal> wrapper element that holds individual
    <literal>record</literal> elements.
   </para>
   <para>
    Cluster record example:
    <screen><![CDATA[
     <cluster xmlns="http://www.indexdata.com/pazpar2/1.0">
       <record>
         <metadata type="title" rank="2">The Shining</metadata>
         <metadata type="author">King, Stephen</metadata>
         <metadata type="kind">ebook</metadata>
       </record>
       <record>
         <metadata type="title" rank="2">The Shining</metadata>
         <metadata type="author">King, Stephen</metadata>
         <metadata type="kind">audio</metadata>
       </record>
    </cluster>
     ]]></screen>
   </para>
  </section>

  <section id="client">
   <title>Client development overview</title>
   <para>
    You can use Pazpar2 from any environment that allows you to use
    webservices. The initial goal of the software was to support
    Ajax-based applications, but there literally are no limits to what
    you can do. You can use Pazpar2 from Javascript, Flash, Java, etc.,
    on the browser side, and from any development environment on the
    server side, and you can pass session tokens and record IDs freely
    around between these environments to build sophisticated applications.
    Use your imagination.
   </para>

   <para>
    The webservice API of Pazpar2 is described in detail in <xref
     linkend="pazpar2_protocol"/>.
   </para>

   <para>
    In brief, you use the 'init' command to create a session, a
    temporary workspace which carries information about the current
    search. You start a new search using the 'search' command. Once the
    search has been started, you can follow its progress using the
    'stat', 'bytarget', 'termlist', or 'show' commands. Detailed records
    can be fetched using the 'record' command. 
   </para>
  </section>

  &sect-ajaxdev;

  <section id="unicode">
   <title>Unicode Compliance</title>
   <para>
    Pazpar2 is Unicode compliant and language and locale aware but relies
    on character encoding for the targets to be specified correctly if
    the targets themselves are not UTF-8 based (most aren't).
    Just a few bad behaving targets can spoil the search experience
    considerably if for example Greek, Russian or otherwise non 7-bit ASCII
    search terms are entered. In these cases some targets return
    records irrelevant to the query, and the result screens will be
    cluttered with noise.
   </para>
   <para>
    While noise from misbehaving targets can not be removed, it can
    be reduced using truly Unicode based ranking. This is an
    option which is available to the system administrator if ICU
    support is compiled into Pazpar2, see
    <xref linkend="installation"/> for details.
   </para>
   <para>
    In addition, the ICU tokenization and normalization rules must
    be defined in the master configuration file described in
    <xref linkend="config-server"/>.
   </para>
  </section>

  <section id="load_balancing">
   <title>Load balancing</title>
   <para>
     Just like any web server, Pazpar2, can be load balanced by a standard
     hardware or software load balancer as long as the session stickiness
     is ensured. If you are already running the Apache2 web server in front
     of Pazpar2 and use the apache mod_proxy module to 'relay' client
     requests to Pazpar2, this set up can be easily extended to include
     load balancing capabilites.
     To do so you need to enable the
     <ulink url="http://httpd.apache.org/docs/2.2/mod/mod_proxy_balancer.html">
      mod_proxy_balance
     </ulink>
     module in your Apache2 installation.
   </para>

   <para>
    On a Debian based Apache 2 system, the relevant modules can
    be enabled with:
    <screen>
     sudo a2enmod proxy_http
    </screen>
   </para>

   <para>
    The mod_proxy_balancer can pass all 'sessionsticky' requests to the
    same backend worker as long as the requests are marked with the
    originating worker's ID (called 'route'). If the Pazpar2 serverID is
    configured (by setting an 'id' attribute on the 'server' element in
    the Pazpar2 configuration file) Pazpar2 will append it to the
    'session' element returned during the 'init' in a mod_proxy_balancer
    compatible manner.
    Since the 'session' is then re-sent by the client (for all pazpar2
    request besides 'init'), the balancer can use the marker to pass
    the request to the right route. To do so the balancer needs to be
    configured to inspect the 'session' parameter.
   </para>

   <example id="load_balancing.example">
    <title>Apache 2 load balancing configuration</title>
    <para>
     Having 4 Pazpar2 instances running on the same host, port range of
     8004-8007 and serverIDs of: pz1, pz2, pz3 and pz4 respectively we
     could use the following Apache 2 configuration to expose a single
     pazpar2 'endpoint' on a standard
     (<filename>/pazpar2/search.pz2</filename>) location:

     <screen><![CDATA[
       <Proxy *>
         AddDefaultCharset off
         Order deny,allow
         Allow from all
       </Proxy>
       ProxyVia Off

       # 'route' has to match the configured pazpar2 server ID
       <Proxy balancer://pz2cluster>
         BalancerMember http://localhost:8004 route=pz1
         BalancerMember http://localhost:8005 route=pz2
         BalancerMember http://localhost:8006 route=pz3
         BalancerMember http://localhost:8007 route=pz4
       </Proxy>

       # route is resent in the 'session' param which has the form:
       # 'sessid.serverid', understandable by the mod_proxy_load_balancer
       # this is not going to work if the client tampers with the 'session' param
       ProxyPass /pazpar2/search.pz2 balancer://pz2cluster lbmethod=byrequests stickysession=session nofailover=On
     ]]></screen>

     The 'ProxyPass' line sets up a reverse proxy for request
     ‘/pazpar2/search.pz2’ and delegates all requests to the load balancer
     (virtual worker) with name ‘pz2cluster’.
     Sticky sessions are enabled and implemented using the ‘session’ parameter.
     The ‘Proxy’ section lists all the servers (real workers) which the
     load balancer can use.
    </para>

   </example>

  </section>

  <section id="relevance_ranking">
   <title>Relevance ranking</title>
   <para>
    Pazpar2 uses a variant of the fterm frequency–inverse document frequency
    (Tf-idf) ranking algorithm.
   </para>
   <para>
    The Tf-part is straightforward to calculate and is based on the
    documents that Pazpar2 fetches. The idf-part, however, is more tricky
    since the corpus at hand is ONLY the relevant documents and not
    irrelevant ones. Pazpar2 does not have the full corpus -- only the
    documents that match a particular search.
   </para>
   <para>
    Computatation of the Tf-part is based on the normalized documents.
    The length, the position and terms are thus normalized at this point.
    Also the computation if performed for each document received from the
    target - before merging takes place. The result of a TF-compuation is
    added to the TF-total of a cluster. Thus, if a document occurs twice,
    then the TF-part is doubled. That, however, can be adjusted, because the
    TF-part may be divided by the number of documents in a cluster.
   </para>
   <para>
    The algorithm used by Pazpar2 has two phases. In phase one
    Pazpar2 computes a tf-array .. This is being done as records are
    fetched form the database. In this case, the rank weigth
    <literal>w</literal>, the and rank tweaks <literal>lead</literal>,
    <literal>follow</literal> and <literal>length</literal>.

   </para>
   <screen><![CDATA[
    tf[1,2,..N] = 0;
    foreach document in a cluster
       foreach field
          w[1,2,..N] = 0;
          for i = 1, .. N:  (each term)
             foreach pos (where term i occurs in field)
                // w is configured weight for field
                // pos is position of term in field
                w[i] += w / (1 + log2(1+lead*pos))
                if (d > 0)
                    w[i] += w[i] * follow / (1+log2(d)
          // length: length of field (number of terms that is)
          if (length strategy is "linear")
             tf[i] += w[i] / length;
          else if (length strategy is "log")
             tf[i] += w[i] / log2(length);
          else if (length strategy is "none")
             tf[i] += w[i];
          ]]></screen>
   <para>
    In phase two, the idf-array is computed and the final score
    is computed. This is done for each cluster as part of each show command.
    The rank tweak <literal>cluster</literal> is in use here.
   </para>
   <screen><![CDATA[
    // dococcur[i]: number of records where term occurs
    // doctotal: number of records
    for i = 1, .., N (each term)
      if (dococcur[i] > 0)
         idf[i] = log(1 + doctotal / dococcur[i])
      else
         idf[i] = 0;

    relevance = 0;
    for i = 1, .., N: (each term)
       if (cluster is "yes")
          tf[i] = tf[i] / cluster_size;
       relevance += 100000 * tf[i] / idf[i];
       ]]></screen>
   <para>
    For controlling the ranking parameters, refer to the
    <link linkend="service-rank">rank</link> element of the 
    service definition.
    Refer to the <link linkend="metadata-rank">rank</link> attribute
    of the metadata element for how to control ranking for individual
    metadata fields.
   </para>
  </section> <!-- relevance_ranking -->

  <section id="masterkey_connect">
   <title>Pazpar2 and MasterKey Connect</title>
   <para>
    MasterKey Connect is a hosted connector, or gateway, service that exposes
    whatever searchable resources you need. Since the service exposes all
    resources using Z39.50 (or SRU), it is easy to set up Pazpar2 to use the
    service. In particular, since all connectors expose basically the same core
    behavior, it is a good use of Pazpar2's mechanism for managing default
    behaviors across similar databases.
   </para>
   <para>
    After installation of Pazpar2, the directory
    <filename>/etc/pazpar2/settings/mkc</filename> (location may
    vary depending on installation preferences) contains an example setup that
    searches two different resources through a MasterKey Connect demo account.
    The file mkc.xml contains default parameters that will work for all
    MasterKey Connect resources (if you decide to become a customer of the
    service, you will substitute your own account credentials for
    the guest/guest). The other files contain specific information about
    a couple of demonstration resources.
   </para>

   <para>
    To play with the demo, just create a symlink from
    <filename>/etc/pazpar2/services-enabled/default.xml</filename>
    to <filename>/etc/pazpar2/services-available/mkc.xml</filename>.
    And restart Pazpar2. You should now be able to search the two demo
    resources using JSDemo or any user interface of your choice.
    If you are interested in learning more about MasterKey Connect, or to
    try out the service for free against your favorite online resource, just
    contact us at <email>info@indexdata.com</email>.
   </para>
  </section>

 </chapter> <!-- Using Pazpar2 -->

 <reference id="reference">
  <title>Reference</title>
  <partintro id="reference-introduction">
   <para>
    The material in this chapter is drawn directly from the individual
    manual entries.
   </para>
  </partintro>
  &manref;
 </reference>

 <appendix id="license">
  <title>License</title>

 <para>
  Pazpar2,
  Copyright &#xa9; &copyright-year; Index Data.
 </para>

 <para>
  Pazpar2 is free software; you can redistribute it and/or modify it under
  the terms of the GNU General Public License as published by the Free
  Software Foundation; either version 2, or (at your option) any later
  version.
 </para>

 <para>
  Pazpar2 is distributed in the hope that it will be useful, but WITHOUT ANY
  WARRANTY; without even the implied warranty of MERCHANTABILITY or
  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  for more details.
 </para>

 <para>
  You should have received a copy of the GNU General Public License
  along with Pazpar2; see the file LICENSE.  If not, write to the
  Free Software Foundation,
  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 </para>

 </appendix>

 &gpl2;

</book>

<!-- Keep this comment at the end of the file
Local variables:
mode: nxml
nxml-child-indent: 1
End:
-->