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 copyright SYSTEM "copyright.xml">
6 <!ENTITY % local SYSTEM "local.ent">
8 <!ENTITY manref SYSTEM "manref.xml">
9 <!ENTITY gpl2 SYSTEM "gpl-2.0.xml">
10 <!ENTITY % idcommon SYSTEM "common/common.ent">
15 <title>Metaproxy - User's Guide and Reference</title>
18 <firstname>Adam</firstname><surname>Dickmeiss</surname>
21 <firstname>Marc</firstname><surname>Cromme</surname>
24 <firstname>Mike</firstname><surname>Taylor</surname>
27 <releaseinfo>&version;</releaseinfo>
29 <year>2005-2013</year>
30 <holder>Index Data</holder>
34 This manual is part of Metaproxy version &version;.
37 Metaproxy is a universal router, proxy and encapsulated
38 metasearcher for information retrieval protocols. It accepts,
39 processes, interprets and redirects requests from IR clients using
40 standard protocols such as the binary
41 <ulink url="&url.z39.50;">ANSI/NISO Z39.50</ulink>
42 and the information search and retrieval
43 web service <ulink url="&url.sru;">SRU</ulink>
44 as well as functioning as a limited
45 <ulink url="&url.http;">HTTP</ulink> server.
48 Metaproxy is configured by an XML file which
49 specifies how the software should function in terms of routes that
50 the request packets can take through the proxy, each step on a
51 route being an instantiation of a filter. Filters come in many
52 types, one for each operation: accepting Z39.50 packets, logging,
53 query transformation, multiplexing, etc. Further filter-types can
54 be added as loadable modules to extend Metaproxy functionality,
58 Metaproxy is covered by the GNU General Public License version 2.
63 <imagedata fileref="common/id.png" format="PNG"/>
66 <imagedata fileref="common/id.eps" format="EPS"/>
73 <chapter id="introduction">
74 <title>Introduction</title>
77 <ulink url="&url.metaproxy;">Metaproxy</ulink>
78 is a stand alone program that acts as a universal router, proxy and
79 encapsulated metasearcher for information retrieval protocols such
80 as <ulink url="&url.z39.50;">Z39.50</ulink> and
81 <ulink url="&url.sru;">SRU</ulink>.
82 To clients, it acts as a server of these protocols: it can be searched,
83 records can be retrieved from it, etc.
84 To servers, it acts as a client: it searches in them,
85 retrieves records from them, etc. it satisfies its clients'
86 requests by transforming them, multiplexing them, forwarding them
87 on to zero or more servers, merging the results, transforming
88 them, and delivering them back to the client. In addition, it
89 acts as a simple <ulink url="&url.http;">HTTP</ulink> server; support
90 for further protocols can be added in a modular fashion, through the
91 creation of new filters.
96 Fish, bananas, cold pyjamas,
97 Mutton, beef and trout!
98 - attributed to Cole Porter.
101 Metaproxy is a more capable alternative to
102 <ulink url="&url.yazproxy;">YAZ Proxy</ulink>,
103 being more powerful, flexible, configurable and extensible. Among
104 its many advantages over the older, more pedestrian work are
105 support for multiplexing (encapsulated metasearching), routing by
106 database name, authentication and authorization and serving local
107 files via HTTP. Equally significant, its modular architecture
108 facilitites the creation of pluggable modules implementing further
112 This manual will describe how to install Metaproxy
113 before giving an overview of its architecture, then discussing the
114 key concept of a filter in some depth and giving an overview of
115 the various filter types, then discussing the configuration file
116 format. After this come several optional chapters which may be
117 freely skipped: a detailed discussion of virtual databases and
118 multi-database searching, some notes on writing extensions
119 (additional filter types) and a high-level description of the
120 source code. Finally comes the reference guide, which contains
121 instructions for invoking the <command>metaproxy</command>
122 program, and detailed information on each type of filter,
127 <chapter id="installation">
128 <title>Installation</title>
130 Metaproxy depends on the following tools/libraries:
132 <varlistentry><term><ulink url="&url.yazplusplus;">YAZ++</ulink></term>
135 This is a C++ library based on <ulink url="&url.yaz;">YAZ</ulink>.
139 <varlistentry><term><ulink url="&url.libxslt;">Libxslt</ulink></term>
141 <para>This is an XSLT processor - based on
142 <ulink url="&url.libxml2;">Libxml2</ulink>. Both Libxml2 and
143 Libxslt must be installed with the development components
144 (header files, etc.) as well as the run-time libraries.
148 <varlistentry><term><ulink url="&url.boost;">Boost</ulink></term>
151 The popular C++ library. Initial versions of Metaproxy
152 was built with 1.32 but this is no longer supported.
153 Metaproxy is known to work with Boost version 1.33 through 1.46.
160 In order to compile Metaproxy a modern C++ compiler is
161 required. Boost, in particular, requires the C++ compiler
162 to facilitate the newest features. Refer to Boost
163 <ulink url="&url.boost.compilers.status;">Compiler Status</ulink>
164 for more information.
167 We have successfully built Metaproxy using the compilers
168 <ulink url="&url.gcc;">GCC</ulink> version 4.0 and
169 <ulink url="&url.vstudio;">Microsoft Visual Studio</ulink> 2003/2005/2008.
173 As an option, Metaproxy may also be compiled with
174 <ulink url="&url.usemarcon;">USEMARCON</ulink> support which allows for
175 MARC conversions for the <xref linkend="ref-record_transform"/> filter.
177 <section id="installation.unix">
178 <title>Installation on Unix (from Source)</title>
180 Here is a quick step-by-step guide on how to compile all the
181 tools that Metaproxy uses. Only few systems have none of the required
182 tools binary packages. If, for example, Libxml2/libxslt are already
183 installed as development packages use those (and omit compilation).
188 <ulink url="&url.usemarcon;">USEMARCON</ulink> is not available
189 as a package at the moment, so Metaproxy must be built from source
190 if that is to be used.
194 <section id="libxml2.fromsource">
195 <title>Libxml2/libxslt</title>
200 gunzip -c libxml2-version.tar.gz|tar xf -
208 gunzip -c libxslt-version.tar.gz|tar xf -
216 <section id="usemarcon">
217 <title>USEMARCON (optional)</title>
219 gunzip -c usemarcon317.tar.gz|tar xf -
228 <section id="yaz.fromsource">
229 <title>YAZ/YAZ++</title>
231 gunzip -c yaz-version.tar.gz|tar xf -
239 gunzip -c yazpp-version.tar.gz|tar xf -
248 <title id="boost.fromsource">Boost</title>
250 Metaproxy needs components thread and test from
254 gunzip -c boost-version.tar.gz|tar xf -
256 ./configure --with-libraries=thread,test,regex --with-toolset=gcc
262 However, under the hood bjam is used. You can invoke that with
265 ./bjam --toolset=gcc --with-thread --with-test --with-regex stage
268 Replace <literal>stage</literal> with <literal>clean</literal> /
269 <literal>install</literal> to perform clean and install respectively.
272 Add <literal>--prefix=DIR</literal> to install Boost in other
273 prefix than <literal>/usr/local</literal>.
276 <section id="metaproxy.fromsource">
277 <title>Metaproxy</title>
279 gunzip -c metaproxy-version.tar.gz|tar xf -
287 You may have to tell configure where Boost is installed by supplying
288 options <literal>--with-boost</literal> and <literal>--with-boost-toolset</literal>.
289 The former sets the PREFIX for Boost (same as --prefix for Boost above).
290 The latter the compiler toolset (eg. gcc34).
293 Pass <literal>--help</literal> to configure to get a list of
299 <section id="installation.debian">
300 <title>Installation on Debian GNU/Linux</title>
302 All dependencies for Metaproxy are available as
303 <ulink url="&url.debian;">Debian</ulink>
304 packages for the squeeze (stable in 2012) and wheezy (testing in 2012)
308 The procedures for Debian based systems, such as
309 <ulink url="&url.ubuntu;">Ubuntu</ulink> is probably similar
312 There is currently no official Debian package for YAZ++.
313 And the official Debian package for YAZ is probably too old.
314 But Index Data builds "new" versions of those for Debian (i386, amd64 only).
317 Update the <filename>/etc/apt/sources.list</filename>
318 to include the Index Data repository.
319 See YAZ' <ulink url="&url.yaz.download.debian;">Download Debian</ulink>
320 for more information.
323 apt-get install libxslt1-dev
324 apt-get install libyazpp4-dev
325 apt-get install libboost-dev
326 apt-get install libboost-thread-dev
327 apt-get install libboost-test-dev
328 apt-get install libboost-regex-dev
331 With these packages installed, the usual configure + make
332 procedure can be used for Metaproxy as outlined in
333 <xref linkend="installation.unix"/>.
337 <section id="installation.rpm">
338 <title>Installation on RPM based Linux Systems</title>
340 All external dependencies for Metaproxy are available as
341 RPM packages, either from your distribution site, or from the
342 <ulink url="http://fr.rpmfind.net/">RPMfind</ulink> site.
345 For example, an installation of the requires Boost C++ development
346 libraries on RedHat Fedora C4 and C5 can be done like this:
348 wget ftp://fr.rpmfind.net/wlinux/fedora/core/updates/testing/4/SRPMS/boost-1.33.0-3.fc4.src.rpm
349 sudo rpmbuild --buildroot src/ --rebuild -p fc4/boost-1.33.0-3.fc4.src.rpm
350 sudo rpm -U /usr/src/redhat/RPMS/i386/boost-*rpm
354 The <ulink url="&url.yaz;">YAZ</ulink> library is needed to
355 compile &metaproxy;, see there
356 for more information on available RPM packages.
359 There is currently no official RPM package for YAZ++.
360 See the <ulink url="&url.yazplusplus;">YAZ++</ulink> pages
361 for more information on a Unix tarball install.
364 With these packages installed, the usual configure + make
365 procedure can be used for Metaproxy as outlined in
366 <xref linkend="installation.unix"/>.
370 <section id="installation.windows">
371 <title>Installation on Windows</title>
373 Metaproxy can be compiled with Microsoft
374 <ulink url="&url.vstudio;">Visual Studio</ulink>.
375 Versions 2003 (C 7.1), 2005 (C 8.0) and 2008 (C 9.0) are known to work.
377 <section id="installation.windows.boost">
380 For Windows, it's easiest to get the precompiled Boost
381 package from <ulink url="&url.boost.windows.download;">here</ulink>.
382 Several versions of the Boost libraries may be selected when
383 installing Boost for windows. Please choose at least the
384 <emphasis>multithreaded</emphasis> (non-DLL) version because
385 the Metaproxy makefile uses that.
388 For more information about installing Boost refer to the
389 <ulink url="&url.boost.getting.started;">getting started</ulink>
394 <section id="installation.windows.libxslt">
395 <title>Libxslt</title>
397 <ulink url="&url.libxslt;">Libxslt</ulink> can be downloaded
399 <ulink url="&url.libxml2.download.win32;">here</ulink>.
402 Libxslt has other dependencies, but these can all be downloaded
403 from the same site. Get the following package:
404 iconv, zlib, libxml2, libxslt.
408 <section id="installation.windows.yaz">
411 <ulink url="&url.yaz;">YAZ</ulink> can be downloaded
413 <ulink url="&url.yaz.download.win32;">here</ulink>.
417 <section id="installation.windows.yazplusplus">
420 Get <ulink url="&url.yazplusplus;">YAZ++</ulink> as well.
421 Version 1.2.7 or later is required.
424 YAZ++ includes NMAKE makefiles, similar to those found in the
429 <section id="installation.windows.metaproxy">
430 <title>Metaproxy</title>
432 Metaproxy is shipped with NMAKE makefiles as well - similar
433 to those found in the YAZ++/YAZ packages. Adjust this Makefile
434 to point to the proper locations of Boost, Libxslt, Libxml2,
435 zlib, iconv, yaz and yazpp.
439 <varlistentry><term><literal>DEBUG</literal></term>
441 If set to 1, the software is
442 compiled with debugging libraries (code generation is
443 multi-threaded debug DLL).
444 If set to 0, the software is compiled with release libraries
445 (code generation is multi-threaded DLL).
450 <term><literal>BOOST</literal></term>
453 Boost install location
459 <term><literal>BOOST_VERSION</literal></term>
462 Boost version (replace . with _).
468 <term><literal>BOOST_TOOLSET</literal></term>
477 <term><literal>LIBXSLT_DIR</literal>,
478 <literal>LIBXML2_DIR</literal> ..</term>
481 Specify the locations of Libxslt, libiconv, libxml2 and
490 After successful compilation you'll find
491 <literal>metaproxy.exe</literal> in the
492 <literal>bin</literal> directory.
500 <chapter id="yazproxy-comparison">
501 <title>YAZ Proxy Comparison</title>
503 The table below lists facilities either supported by either
504 <ulink url="&url.yazproxy;">YAZ Proxy</ulink> or Metaproxy.
506 <table id="yazproxy-comparison-table">
507 <title>Metaproxy / YAZ Proxy comparison</title>
511 <entry>Facility</entry>
512 <entry>Metaproxy</entry>
513 <entry>YAZ Proxy</entry>
518 <entry>Z39.50 server</entry>
519 <entry>Using filter <literal>frontend_net</literal></entry>
520 <entry>Supported</entry>
523 <entry>SRU server</entry>
524 <entry>Supported with filter <literal>sru_z3950</literal></entry>
525 <entry>Supported</entry>
528 <entry>Z39.50 client</entry>
529 <entry>Supported with filter <literal>z3950_client</literal></entry>
530 <entry>Supported</entry>
533 <entry>SRU client</entry>
534 <entry>Unsupported</entry>
535 <entry>Unsupported</entry>
538 <entry>Connection reuse</entry>
539 <entry>Supported with filter <literal>session_shared</literal></entry>
540 <entry>Supported</entry>
543 <entry>Connection share</entry>
544 <entry>Supported with filter <literal>session_shared</literal></entry>
545 <entry>Unsupported</entry>
548 <entry>Result set reuse</entry>
549 <entry>Supported with filter <literal>session_shared</literal></entry>
550 <entry>Within one Z39.50 session / HTTP keep-alive</entry>
553 <entry>Record cache</entry>
554 <entry>Supported by filter <literal>session_shared</literal></entry>
555 <entry>Supported for last result set within one Z39.50/HTTP-keep alive session</entry>
558 <entry>Z39.50 Virtual database, i.e. select any Z39.50 target for database</entry>
559 <entry>Supported with filter <literal>virt_db</literal></entry>
560 <entry>Unsupported</entry>
563 <entry>SRU Virtual database, i.e. select any Z39.50 target for path</entry>
564 <entry>Supported with filter <literal>virt_db</literal>,
565 <literal>sru_z3950</literal></entry>
566 <entry>Supported</entry>
569 <entry>Multi target search</entry>
570 <entry>Supported with filter <literal>multi</literal> (round-robin)</entry>
571 <entry>Unsupported</entry>
574 <entry>Retrieval and search limits</entry>
575 <entry>Supported using filter <literal>limit</literal></entry>
576 <entry>Supported</entry>
579 <entry>Bandwidth limits</entry>
580 <entry>Supported using filter <literal>limit</literal></entry>
581 <entry>Supported</entry>
584 <entry>Connect limits</entry>
585 <entry>Supported by filter <literal>frontend_net</literal> (connect-max)</entry>
586 <entry>Supported</entry>
589 <entry>Retrieval sanity check and conversions</entry>
590 <entry>Supported using filter <literal>record_transform</literal></entry>
591 <entry>Supported</entry>
594 <entry>Query check</entry>
596 Supported by <literal>query_rewrite</literal> which may be check
597 a query and throw diagnostics (errors)
599 <entry>Supported</entry>
602 <entry>Query rewrite</entry>
603 <entry>Supported with <literal>query_rewrite</literal></entry>
604 <entry>Unsupported</entry>
607 <entry>Session invalidate for -1 hits</entry>
608 <entry>Unsupported</entry>
609 <entry>Supported</entry>
612 <entry>Architecture</entry>
613 <entry>Multi-threaded + select for networked modules such as
614 <literal>frontend_net</literal>)</entry>
615 <entry>Single-threaded using select</entry>
619 <entry>Extensability</entry>
620 <entry>Most functionality implemented as loadable modules</entry>
621 <entry>Unsupported and experimental</entry>
625 <entry><ulink url="&url.usemarcon;">USEMARCON</ulink></entry>
626 <entry>Supported with <literal>record_transform</literal></entry>
627 <entry>Supported</entry>
631 <entry>Portability</entry>
633 Requires YAZ, YAZ++ and modern C++ compiler supporting
634 <ulink url="&url.boost;">Boost</ulink>.
637 Requires YAZ and YAZ++.
638 STL is not required so pretty much any C++ compiler out there should work.
647 <chapter id="architecture">
648 <title>The Metaproxy Architecture</title>
650 The Metaproxy architecture is based on three concepts:
651 the <emphasis>package</emphasis>,
652 the <emphasis>route</emphasis>
653 and the <emphasis>filter</emphasis>.
657 <term>Packages</term>
660 A package is request or response, encoded in some protocol,
661 issued by a client, making its way through Metaproxy, send to or
662 received from a server, or sent back to the client.
665 The core of a package is the protocol unit - for example, a
666 Z39.50 Init Request or Search Response, or an SRU searchRetrieve
667 URL or Explain Response. In addition to this core, a package
668 also carries some extra information added and used by Metaproxy
672 In general, packages are doctored as they pass through
673 Metaproxy. For example, when the proxy performs authentication
674 and authorization on a Z39.50 Init request, it removes the
675 authentication credentials from the package so that they are not
676 passed onto the back-end server; and when search-response
677 packages are obtained from multiple servers, they are merged
678 into a single unified package that makes its way back to the
687 Packages make their way through routes, which can be thought of
688 as programs that operate on the package data-type. Each
689 incoming package initially makes its way through a default
690 route, but may be switched to a different route based on various
691 considerations. Routes are made up of sequences of filters (see
700 Filters provide the individual instructions within a route, and
701 effect the necessary transformations on packages. A particular
702 configuration of Metaproxy is essentially a set of filters,
703 described by configuration details and arranged in order in one
704 or more routes. There are many kinds of filter - about a dozen
705 at the time of writing with more appearing all the time - each
706 performing a specific function and configured by different
710 The word ``filter'' is sometimes used rather loosely, in two
711 different ways: it may be used to mean a particular
712 <emphasis>type</emphasis> of filter, as when we speak of ``the
713 auth_simple filter'' or ``the multi filter''; or it may be used
714 to be a specific <emphasis>instance</emphasis> of a filter
715 within a Metaproxy configuration. For example, a single
716 configuration will often contain multiple instances of the
717 <literal>z3950_client</literal> filter. In
718 operational terms, of these is a separate filter. In practice,
719 context always make it clear which sense of the word ``filter''
723 Extensibility of Metaproxy is primarily through the creation of
724 plugins that provide new filters. The filter API is small and
725 conceptually simple, but there are many details to master. See
727 <link linkend="filters">Filters</link>.
733 Since packages are created and handled by the system itself, and
734 routes are conceptually simple, most of the remainder of this
735 document concentrates on filters. After a brief overview of the
736 filter types follows, along with some thoughts on possible future
743 <chapter id="filters">
744 <title>Filters</title>
747 <section id="filters-introductory-notes">
748 <title>Introductory notes</title>
750 It's useful to think of Metaproxy as an interpreter providing a small
751 number of primitives and operations, but operating on a very
752 complex data type, namely the ``package''.
755 A package represents a Z39.50 or SRU/W request (whether for Init,
756 Search, Scan, etc.) together with information about where it came
757 from. Packages are created by front-end filters such as
758 <literal>frontend_net</literal> (see below), which reads them from
759 the network; other front-end filters are possible. They then pass
760 along a route consisting of a sequence of filters, each of which
761 transforms the package and may also have side-effects such as
762 generating logging. Eventually, the route will yield a response,
763 which is sent back to the origin.
766 There are many kinds of filter: some that are defined statically
767 as part of Metaproxy, and others may be provided by third parties
768 and dynamically loaded. They all conform to the same simple API
769 of essentially two methods: <function>configure()</function> is
770 called at startup time, and is passed an XML DOM tree representing that
771 part of the configuration file that pertains to this filter
772 instance: it is expected to walk that tree extracting relevant
773 information; and <function>process()</function> is called every
774 time the filter has to processes a package.
777 While all filters provide the same API, there are different modes
778 of functionality. Some filters are sources: they create
780 (<literal>frontend_net</literal>);
781 others are sinks: they consume packages and return a result
782 (<literal>backend_test</literal>,
783 <literal>bounce</literal>,
784 <literal>http_file</literal>,
785 <literal>z3950_client</literal>);
786 the others are true filters, that read, process and pass on the
787 packages they are fed
788 (<literal>auth_simple</literal>,
789 <literal>log</literal>,
790 <literal>multi</literal>,
791 <literal>query_rewrite</literal>,
792 <literal>record_transform</literal>,
793 <literal>session_shared</literal>,
794 <literal>sru_z3950</literal>,
795 <literal>template</literal>,
796 <literal>virt_db</literal>).
801 <section id="overview.filter.types">
802 <title>Overview of filter types</title>
804 We now briefly consider each of the types of filter supported by
805 the core Metaproxy binary. This overview is intended to give a
806 flavor of the available functionality; more detailed information
807 about each type of filter is included below in
808 <xref linkend="reference"/>.
811 The filters are here named by the string that is used as the
812 <literal>type</literal> attribute of a
813 <literal><filter></literal> element in the configuration
814 file to request them, with the name of the class that implements
815 them in parentheses. (The classname is not needed for normal
816 configuration and use of Metaproxy; it is useful only to
820 The filters are here listed in alphabetical order:
827 New virt_db-alike that does inteligent peer choice, explain merging,
828 adds FD&N to explain. Keeps init responses (like "virt_db Classic"),
829 makes routing choices based on local explain knowledge. Ref IDDI
832 Filter to convert Explain Classic to ZeeRex.
834 CQL2PQF (which needs augmented ZeeRex) - MARC for Talis.
838 Figure out what additional information we need in:
839 ZeeRex (check against D3.1)
840 Init request (e.g. loop detection)
841 Query package (e.g. number of hops)
842 Query response (e.g. record source)
846 <section id="auth_simple">
847 <title><literal>auth_simple</literal>
848 (mp::filter::AuthSimple)</title>
850 Simple authentication and authorization. The configuration
851 specifies the name of a file that is the user register, which
852 lists <varname>username</varname>:<varname>password</varname>
853 pairs, one per line, colon separated. When a session begins, it
854 is rejected unless username and passsword are supplied, and match
855 a pair in the register. The configuration file may also specific
856 the name of another file that is the target register: this lists
857 lists <varname>username</varname>:<varname>dbname</varname>,<varname>dbname</varname>...
858 sets, one per line, with multiple database names separated by
859 commas. When a search is processed, it is rejected unless the
860 database to be searched is one of those listed as available to
865 <section id="backend_test">
866 <title><literal>backend_test</literal>
867 (mp::filter::Backend_test)</title>
869 A partial sink that provides dummy responses in the manner of the
870 <literal>yaz-ztest</literal> Z39.50 server. This is useful only
871 for testing. Seriously, you don't need this. Pretend you didn't
872 even read this section.
876 <section id="bounce">
877 <title><literal>bounce</literal>
878 (mp::filter::Bounce)</title>
880 A sink that swallows <emphasis>all packages</emphasis>,
881 and returns them almost unprocessed.
882 It never sends any package of any type further down the row, but
883 sets Z39.50 packages to Z_Close, and HTTP_Request packages to
884 HTTP_Response err code 400 packages, and adds a suitable bounce
886 The bounce filter is usually added at end of each filter chain route
887 to prevent infinite hanging of for example HTTP
888 requests packages when only the Z39.50 client partial sink
889 filter is found in the
894 <section id="cql_rpn">
895 <title><literal>cql_rpn</literal>
896 (mp::filter::CQLtoRPN)</title>
898 A query language transforming filter which catches Z39.50
899 <literal>searchRequest</literal>
900 packages containing <literal>CQL</literal> queries, transforms
901 those to <literal>RPN</literal> queries,
902 and sends the <literal>searchRequests</literal> on to the next
903 filters. It is among other things useful in a SRU context.
907 <section id="frontend_net">
908 <title><literal>frontend_net</literal>
909 (mp::filter::FrontendNet)</title>
911 A source that accepts Z39.50 connections from a port
912 specified in the configuration, reads protocol units, and
913 feeds them into the next filter in the route. When the result is
914 received, it is returned to the original origin.
918 <section id="http_file">
919 <title><literal>http_file</literal>
920 (mp::filter::HttpFile)</title>
922 A partial sink which swallows only
923 <literal>HTTP_Request</literal> packages, and
924 returns the contents of files from the local
925 filesystem in response to HTTP requests.
926 It lets Z39.50 packages and all other forthcoming package types
929 does mean that Metaproxy is also a Web-server in its spare time. So
930 far it does not contain either an email-reader or a Lisp
931 interpreter, but that day is surely coming.)
935 <section id="load_balance">
936 <title><literal>load_balance</literal>
937 (mp::filter::LoadBalance)</title>
939 Performs load balancing for incoming Z39.50 init requests.
940 It is used together with the <literal>virt_db</literal> filter,
941 but unlike the <literal>multi</literal> filter it does send an
942 entire session to only one of the virtual backends. The
943 <literal>load_balance</literal> filter is assuming that
944 all backend targets have equal content, and chooses the backend
945 with least load cost for a new session.
948 This filter is experimental and yet not mature for heavy load
956 <title><literal>log</literal>
957 (mp::filter::Log)</title>
959 Writes logging information to standard output, and passes on
960 the package unchanged. A log file name can be specified, as well
961 as multiple different logging formats.
966 <title><literal>multi</literal>
967 (mp::filter::Multi)</title>
969 Performs multi-database searching.
971 <link linkend="multidb">the extended discussion</link>
972 of virtual databases and multi-database searching below.
976 <section id="query_rewrite">
977 <title><literal>query_rewrite</literal>
978 (mp::filter::QueryRewrite)</title>
980 Rewrites Z39.50 <literal>Type-1</literal>
981 and <literal>Type-101</literal> (``<literal>RPN</literal>'')
983 three-step process: the query is transliterated from Z39.50
984 packet structures into an XML representation; that XML
985 representation is transformed by an XSLT stylesheet; and the
986 resulting XML is transliterated back into the Z39.50 packet
992 <section id="record_transform">
993 <title><literal>record_transform</literal>
994 (mp::filter::RecordTransform)</title>
996 This filter acts only on Z3950 present requests, and let all
997 other types of packages and requests pass untouched. It's use is
998 twofold: blocking Z3950 present requests, which the backend
999 server does not understand and can not honor, and transforming
1000 the present syntax and elementset name according to the rules
1001 specified, to fetch only existing record formats, and transform
1002 them on the fly to requested record syntaxes.
1006 <section id="session_shared">
1007 <title><literal>session_shared</literal>
1008 (mp::filter::SessionShared)</title>
1010 This filter implements global sharing of
1011 result sets (i.e. between threads and therefore between
1012 clients), yielding performance improvements by clever resource
1017 <section id="sru_z3950">
1018 <title><literal>sru_z3950</literal>
1019 (mp::filter::SRUtoZ3950)</title>
1021 This filter transforms valid
1022 SRU GET/POST/SOAP searchRetrieve requests to Z3950 init, search,
1023 and present requests, and wraps the
1024 received hit counts and XML records into suitable SRU response
1026 The <literal>sru_z3950</literal> filter processes also SRU
1027 GET/POST/SOAP explain requests, returning
1028 either the absolute minimum required by the standard, or a full
1029 pre-defined ZeeReX explain record.
1031 <ulink url="&url.zeerex.explain;">ZeeReX Explain</ulink>
1032 standard pages and the
1033 <ulink url="&url.sru.explain;">SRU Explain</ulink> pages
1034 for more information on the correct explain syntax.
1035 SRU scan requests are not supported yet.
1039 <section id="template">
1040 <title><literal>template</literal>
1041 (mp::filter::Template)</title>
1043 Does nothing at all, merely passing the packet on. (Maybe it
1044 should be called <literal>nop</literal> or
1045 <literal>passthrough</literal>?) This exists not to be used, but
1046 to be copied - to become the skeleton of new filters as they are
1047 written. As with <literal>backend_test</literal>, this is not
1048 intended for civilians.
1052 <section id="virt_db">
1053 <title><literal>virt_db</literal>
1054 (mp::filter::VirtualDB)</title>
1056 Performs virtual database selection: based on the name of the
1057 database in the search request, a server is selected, and its
1058 address added to the request in a <literal>VAL_PROXY</literal>
1059 otherInfo packet. It will subsequently be used by a
1060 <literal>z3950_client</literal> filter.
1062 <link linkend="multidb">the extended discussion</link>
1063 of virtual databases and multi-database searching below.
1067 <section id="z3950_client">
1068 <title><literal>z3950_client</literal>
1069 (mp::filter::Z3950Client)</title>
1071 A partial sink which swallows only Z39.50 packages.
1072 It performs Z39.50 searching and retrieval by proxying the
1073 packages that are passed to it. Init requests are sent to the
1074 address specified in the <literal>VAL_PROXY</literal> otherInfo
1075 attached to the request: this may have been specified by client,
1076 or generated by a <literal>virt_db</literal> filter earlier in
1077 the route. Subsequent requests are sent to the same address,
1078 which is remembered at Init time in a Session object.
1079 HTTP_Request packages and all other forthcoming package types
1080 are passed untouched.
1085 <section id="zeerex_explain">
1086 <title><literal>zeerex_explain</literal>
1087 (mp::filter::ZeerexExplain)</title>
1089 This filter acts as a sink for
1090 Z39.50 explain requests, returning a static ZeeReX
1091 Explain XML record from the config section. All other packages
1094 <ulink url="&url.zeerex.explain;">ZeeReX Explain</ulink>
1096 for more information on the correct explain syntax.
1100 This filter is not yet completed.
1109 <section id="future.directions">
1110 <title>Future directions</title>
1112 Some other filters that do not yet exist, but which would be
1113 useful, are briefly described. These may be added in future
1114 releases (or may be created by third parties, as loadable
1120 <term><literal>frontend_cli</literal> (source)</term>
1123 Command-line interface for generating requests.
1128 <term><literal>sru_client</literal> (sink)</term>
1131 SRU/GET and SRU/SOAP searching and retrieval.
1136 <term><literal>opensearch_client</literal> (sink)</term>
1139 A9 OpenSearch searching and retrieval.
1149 <chapter id="configuration">
1150 <title>Configuration: the Metaproxy configuration file format</title>
1153 <section id="configuration-introductory-notes">
1154 <title>Introductory notes</title>
1156 If Metaproxy is an interpreter providing operations on packages, then
1157 its configuration file can be thought of as a program for that
1158 interpreter. Configuration is by means of a single XML file, the name
1159 of which is supplied as the sole command-line argument to the
1160 <command>metaproxy</command> program. (See
1161 <xref linkend="reference"/> below for more information on invoking
1166 <section id="overview.xml.structure">
1167 <title>Overview of the config file XML structure</title>
1169 All elements and attributes are in the namespace
1170 <ulink url="http://indexdata.com/metaproxy"/>.
1171 This is most easily achieved by setting the default namespace on
1172 the top-level element, as here:
1175 <metaproxy xmlns="http://indexdata.com/metaproxy" version="1.0">
1178 The top-level element is <metaproxy>. This contains
1179 a <dlpath> element,
1180 a <start> element,
1181 a <filters> element and
1182 a <routes> element, in that order. <dlpath> and
1183 <filters> are optional; the other two are mandatory.
1184 All four are non-repeatable.
1187 The <dlpath;> element contains a text element which
1188 specifies the location of filter modules. This is only needed
1189 if Metaproxy must load 3rd party filters (most filters with Metaproxy
1190 are built into the Metaproxy application).
1193 The <start> element is empty, but carries a
1194 <literal>route</literal> attribute, whose value is the name of
1195 route at which to start running - analogous to the name of the
1196 start production in a formal grammar.
1199 If present, <filters> contains zero or more <filter>
1200 elements. Each filter carries a <literal>type</literal> attribute
1201 which specifies what kind of filter is being defined
1202 (<literal>frontend_net</literal>, <literal>log</literal>, etc.)
1203 and contain various elements that provide suitable configuration
1204 for a filter of its type. The filter-specific elements are
1206 <xref linkend="reference"/>.
1207 Filters defined in this part of the file must carry an
1208 <literal>id</literal> attribute so that they can be referenced
1212 <routes> contains one or more <route> elements, each
1213 of which must carry an <literal>id</literal> element. One of the
1214 routes must have the ID value that was specified as the start
1215 route in the <start> element's <literal>route</literal>
1216 attribute. Each route contains zero or more <filter>
1217 elements. These are of two types. They may be empty, but carry a
1218 <literal>refid</literal> attribute whose value is the same as the
1219 <literal>id</literal> of a filter previously defined in the
1220 <filters> section. Alternatively, a route within a filter
1221 may omit the <literal>refid</literal> attribute, but contain
1222 configuration elements similar to those used for filters defined
1223 in the <filters> section. (In other words, each filter in a
1224 route may be included either by reference or by physical
1230 <section id="example.configuration">
1231 <title>An example configuration</title>
1233 The following is a small, but complete, Metaproxy configuration
1234 file (included in the distribution as
1235 <literal>metaproxy/etc/config1.xml</literal>).
1236 This file defines a very simple configuration that simply proxies
1237 to whatever back-end server the client requests, but logs each
1238 request and response. This can be useful for debugging complex
1239 client-server dialogues.
1241 <screen><![CDATA[<?xml version="1.0"?>
1242 <metaproxy xmlns="http://indexdata.com/metaproxy" version="1.0">
1243 <dlpath>/usr/lib/metaproxy/modules</dlpath>
1244 <start route="start"/>
1246 <filter id="frontend" type="frontend_net">
1249 <filter id="backend" type="z3950_client">
1254 <filter refid="frontend"/>
1255 <filter type="log"/>
1256 <filter refid="backend"/>
1257 <filter type="bounce"/>
1263 It works by defining a single route, called
1264 <literal>start</literal>, which consists of a sequence of four
1265 filters. The first and last of these are included by reference:
1266 their <literal><filter></literal> elements have
1267 <literal>refid</literal> attributes that refer to filters defined
1268 within the prior <literal><filters></literal> section. The
1269 middle filter is included inline in the route.
1272 The four filters in the route are as follows: first, a
1273 <literal>frontend_net</literal> filter accepts Z39.50 requests
1274 from any host on port 9000; then these requests are passed through
1275 a <literal>log</literal> filter that emits a message for each
1276 request; they are then fed into a <literal>z3950_client</literal>
1277 filter, which forwards all Z39.50 requests to the client-specified
1278 back-end Z39.509 server. Those Z39.50 packages are returned by the
1279 <literal>z3950_client</literal> filter, with the response data
1280 filled by the external Z39.50 server targeted.
1281 All non-Z39.50 packages are passed through to the
1282 <literal>bounce</literal> filter, which definitely bounces
1283 everything, including fish, bananas, cold pyjamas,
1284 mutton, beef and trout packages.
1285 When the response arrives, it is handed
1286 back to the <literal>log</literal> filter, which emits another
1287 message; and then to the <literal>frontend_net</literal> filter,
1288 which returns the response to the client.
1292 <section id="config-file-modularity">
1293 <title>Config file modularity</title>
1295 Metaproxy XML configuration snippets can be reused by other
1296 filters using the <literal>XInclude</literal> standard, as seen in
1297 the <literal>/etc/config-sru-to-z3950.xml</literal> example SRU
1300 <filter id="sru" type="sru_z3950">
1301 <database name="Default">
1302 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
1303 href="explain.xml"/>
1310 <section id="config-file-syntax-check">
1311 <title>Config file syntax checking</title>
1313 The distribution contains RelaxNG Compact and XML syntax checking
1314 files, as well as XML Schema files. These are found in the
1317 xml/schema/metaproxy.rnc
1318 xml/schema/metaproxy.rng
1319 xml/schema/metaproxy.xsd
1321 and can be used to verify or debug the XML structure of
1322 configuration files. For example, using the utility
1323 <filename>xmllint</filename>, syntax checking is done like this:
1325 xmllint --noout --schema xml/schema/metaproxy.xsd etc/config-local.xml
1326 xmllint --noout --relaxng xml/schema/metaproxy.rng etc/config-local.xml
1328 (A recent version of <literal>libxml2</literal> is required, as
1329 support for XML Schemas is a relatively recent addition.)
1332 You can of course use any other RelaxNG or XML Schema compliant tool
1340 <chapter id="multidb">
1341 <title>Virtual databases and multi-database searching</title>
1344 <section id="multidb-introductory-notes">
1345 <title>Introductory notes</title>
1347 Two of Metaproxy's filters are concerned with multiple-database
1348 operations. Of these, <literal>virt_db</literal> can work alone
1349 to control the routing of searches to one of a number of servers,
1350 while <literal>multi</literal> can work together with
1351 <literal>virt_db</literal> to perform multi-database searching, merging
1352 the results into a unified result-set - ``metasearch in a box''.
1355 The interaction between
1356 these two filters is necessarily complex: it reflects the real,
1357 irreducible complexity of multi-database searching in a protocol such
1358 as Z39.50 that separates initialization from searching, and in
1359 which the database to be searched is not known at initialization
1363 It's possible to use these filters without understanding the
1364 details of their functioning and the interaction between them; the
1365 next two sections of this chapter are ``HOW-TO'' guides for doing
1366 just that. However, debugging complex configurations will require
1367 a deeper understanding, which the last two sections of this
1368 chapters attempt to provide.
1373 <section id="multidb.virt_db">
1374 <title>Virtual databases with the <literal>virt_db</literal> filter</title>
1376 Working alone, the purpose of the
1377 <literal>virt_db</literal>
1378 filter is to route search requests to one of a selection of
1379 back-end databases. In this way, a single Z39.50 endpoint
1380 (running Metaproxy) can provide access to several different
1381 underlying services, including those that would otherwise be
1382 inaccessible due to firewalls. In many useful configurations, the
1383 back-end databases are local to the Metaproxy installation, but
1384 the software does not enforce this, and any valid Z39.50 servers
1385 may be used as back-ends.
1388 For example, a <literal>virt_db</literal>
1389 filter could be set up so that searches in the virtual database
1390 ``lc'' are forwarded to the Library of Congress bibliographic
1391 catalogue server, and searches in the virtual database ``marc''
1392 are forwarded to the toy database of MARC records that Index Data
1393 hosts for testing purposes. A <literal>virt_db</literal>
1394 configuration to make this switch would look like this:
1396 <screen><![CDATA[<filter type="virt_db">
1398 <database>lc</database>
1399 <target>z3950.loc.gov:7090/voyager</target>
1402 <database>marc</database>
1403 <target>indexdata.com/marc</target>
1405 </filter>]]></screen>
1407 As well as being useful in it own right, this filter also provides
1408 the foundation for multi-database searching.
1413 <section id="multidb.multi">
1414 <title>Multi-database search with the <literal>multi</literal> filter</title>
1416 To arrange for Metaproxy to broadcast searches to multiple back-end
1417 servers, the configuration needs to include two components: a
1418 <literal>virt_db</literal>
1419 filter that specifies multiple
1420 <literal><target></literal>
1421 elements, and a subsequent
1422 <literal>multi</literal>
1423 filter. Here, for example, is a complete configuration that
1424 broadcasts searches to both the Library of Congress catalogue and
1425 Index Data's tiny testing database of MARC records:
1427 <screen><![CDATA[<?xml version="1.0"?>
1428 <metaproxy xmlns="http://indexdata.com/metaproxy" version="1.0">
1429 <start route="start"/>
1432 <filter type="frontend_net">
1433 <threads>10</threads>
1436 <filter type="virt_db">
1438 <database>lc</database>
1439 <target>z3950.loc.gov:7090/voyager</target>
1442 <database>marc</database>
1443 <target>indexdata.com/marc</target>
1446 <database>all</database>
1447 <target>z3950.loc.gov:7090/voyager</target>
1448 <target>indexdata.com/marc</target>
1451 <filter type="multi"/>
1452 <filter type="z3950_client">
1453 <timeout>30</timeout>
1455 <filter type="bounce"/>
1458 </metaproxy>]]></screen>
1461 <literal>virt_db</literal>
1462 filter that specifies multiple
1463 <literal><target></literal>
1464 elements but without a subsequent
1465 <literal>multi</literal>
1466 filter yields surprising and undesirable results, as will be
1467 described below. Don't do that.)
1470 Metaproxy can be invoked with this configuration as follows:
1472 <screen>../src/metaproxy --config config-simple-multi.xml</screen>
1474 And thereafter, Z39.50 clients can connect to the running server
1475 (on port 9000, as specified in the configuration) and search in
1476 any of the databases
1477 <literal>lc</literal> (the Library of Congress catalogue),
1478 <literal>marc</literal> (Index Data's test database of MARC records)
1480 <literal>all</literal> (both of these). As an example, a session
1481 using the YAZ command-line client <literal>yaz-client</literal> is
1482 here included (edited for brevity and clarity):
1484 <screen><![CDATA[$ yaz-client @:9000
1488 Search was a success.
1489 Number of hits: 10000, setno 1
1493 Search was a success.
1494 Number of hits: 10, setno 3
1498 Search was a success.
1499 Number of hits: 10010, setno 4
1502 [marc]Record type: USmarc
1505 005 00000000000000.0
1506 008 910710c19910701nju 00010 eng
1510 100 10 $a Jack Collins
1511 245 10 $a How to program a computer
1517 [VOYAGER]Record type: USmarc
1519 005 20041229102447.0
1520 008 030910s2004 caua 000 0 eng
1521 035 $a (DLC) 2003112666
1522 906 $a 7 $b cbc $c orignew $d 4 $e epcn $f 20 $g y-gencatlg
1523 925 0 $a acquire $b 1 shelf copy $x policy default
1524 955 $a pc10 2003-09-10 $a pv12 2004-06-23 to SSCD; $h sj05 2004-11-30 $e sj05 2004-11-30 to Shelf.
1527 040 $a DLC $c DLC $d DLC
1528 050 00 $a MLCM 2004/03312 (G)
1529 245 10 $a 007, everything or nothing : $b Prima's official strategy guide / $c created by Kaizen Media Group.
1530 246 3 $a Double-O-seven, everything or nothing
1531 246 30 $a Prima's official strategy guide
1532 260 $a Roseville, CA : $b Prima Games, $c c2004.
1533 300 $a 161 p. : $b col. ill. ; $c 28 cm.
1534 500 $a "Platforms: Nintendo GameCube, Macintosh, PC, PlayStation 2 computer entertainment system, Xbox"--P. [4] of cover.
1535 650 0 $a Video games.
1536 710 2 $a Kaizen Media Group.
1537 856 42 $3 Publisher description $u http://www.loc.gov/catdir/description/random052/2003112666.html
1542 As can be seen, the first record in the result set is from the
1543 Index Data test database, and the second from the Library of
1544 Congress database. The result-set continues alternating records
1545 round-robin style until the point where one of the databases'
1546 records are exhausted.
1549 This example uses only two back-end databases; more may be used.
1550 There is no limitation imposed on the number of databases that may
1551 be metasearched in this way: issues of resource usage and
1552 administrative complexity dictate the practical limits.
1555 What happens when one of the databases doesn't respond? By default,
1556 the entire multi-database search fails, and the appropriate
1557 diagnostic is returned to the client. This is usually appropriate
1558 during development, when technicians need maximum information, but
1559 can be inconvenient in deployment, when users typically don't want
1560 to be bothered with problems of this kind and prefer just to get
1561 the records from the databases that are available. To obtain this
1562 latter behavior add an empty
1563 <literal><hideunavailable></literal>
1565 <literal>multi</literal> filter:
1567 <screen><![CDATA[ <filter type="multi">
1569 </filter>]]></screen>
1571 Under this regime, an error is reported to the client only if
1572 <emphasis>all</emphasis> the databases in a multi-database search
1578 <section id="multidb.what">
1579 <title>What's going on?</title>
1581 <title>Lark's vomit</title>
1583 This section goes into a level of technical detail that is
1584 probably not necessary in order to configure and use Metaproxy.
1585 It is provided only for those who like to know how things work.
1586 You should feel free to skip on to the next section if this one
1587 doesn't seem like fun.
1591 Hold on tight - this may get a little hairy.
1594 In the general course of things, a Z39.50 Init request may carry
1595 with it an otherInfo packet of type <literal>VAL_PROXY</literal>,
1596 whose value indicates the address of a Z39.50 server to which the
1597 ultimate connection is to be made. (This otherInfo packet is
1598 supported by YAZ-based Z39.50 clients and servers, but has not yet
1599 been ratified by the Maintenance Agency and so is not widely used
1600 in non-Index Data software. We're working on it.)
1601 The <literal>VAL_PROXY</literal> packet functions
1602 analogously to the absoluteURI-style Request-URI used with the GET
1603 method when a web browser asks a proxy to forward its request: see
1605 <ulink url="http://www.w3.org/Protocols/rfc2616/rfc2616-sec5.html#sec5.1.2"
1606 >Request-URI</ulink>
1608 <ulink url="http://www.w3.org/Protocols/rfc2616/rfc2616.html"
1609 >the HTTP 1.1 specification</ulink>.
1612 Within Metaproxy, Search requests that are part of the same
1613 session as an Init request that carries a
1614 <literal>VAL_PROXY</literal> otherInfo are also annotated with the
1615 same information. The role of the <literal>virt_db</literal>
1616 filter is to rewrite this otherInfo packet dependent on the
1617 virtual database that the client wants to search.
1620 When Metaproxy receives a Z39.50 Init request from a client, it
1621 doesn't immediately forward that request to the back-end server.
1622 Why not? Because it doesn't know <emphasis>which</emphasis>
1623 back-end server to forward it to until the client sends a Search
1624 request that specifies the database that it wants to search in.
1625 Instead, it just treasures the Init request up in its heart; and,
1626 later, the first time the client does a search on one of the
1627 specified virtual databases, a connection is forged to the
1628 appropriate server and the Init request is forwarded to it. If,
1629 later in the session, the same client searches in a different
1630 virtual database, then a connection is forged to the server that
1631 hosts it, and the same cached Init request is forwarded there,
1635 All of this clever Init-delaying is done by the
1636 <literal>frontend_net</literal> filter. The
1637 <literal>virt_db</literal> filter knows nothing about it; in
1638 fact, because the Init request that is received from the client
1639 doesn't get forwarded until a Search request is received, the
1640 <literal>virt_db</literal> filter (and the
1641 <literal>z3950_client</literal> filter behind it) doesn't even get
1642 invoked at Init time. The <emphasis>only</emphasis> thing that a
1643 <literal>virt_db</literal> filter ever does is rewrite the
1644 <literal>VAL_PROXY</literal> otherInfo in the requests that pass
1648 It is possible for a <literal>virt_db</literal> filter to contain
1650 <literal><target></literal>
1651 elements. What does this mean? Only that the filter will add
1652 multiple <literal>VAL_PROXY</literal> otherInfo packets to the
1653 Search requests that pass through it. That's because the virtual
1654 DB filter is dumb, and does exactly what it's told - no more, no
1656 If a Search request with multiple <literal>VAL_PROXY</literal>
1657 otherInfo packets reaches a <literal>z3950_client</literal>
1658 filter, this is an error. That filter doesn't know how to deal
1659 with multiple targets, so it will either just pick one and search
1660 in it, or (better) fail with an error message.
1663 The <literal>multi</literal> filter comes to the rescue! This is
1664 the only filter that knows how to deal with multiple
1665 <literal>VAL_PROXY</literal> otherInfo packets, and it does so by
1666 making multiple copies of the entire Search request: one for each
1667 <literal>VAL_PROXY</literal>. Each of these new copies is then
1668 passed down through the remaining filters in the route. (The
1669 copies are handled in parallel though the
1670 spawning of new threads.) Since the copies each have only one
1671 <literal>VAL_PROXY</literal> otherInfo, they can be handled by the
1672 <literal>z3950_client</literal> filter, which happily deals with
1673 each one individually. When the results of the individual
1674 searches come back up to the <literal>multi</literal> filter, it
1675 merges them into a single Search response, which is what
1676 eventually makes it back to the client.
1681 <imagedata fileref="multi.pdf" format="PDF" scale="50"/>
1684 <imagedata fileref="multi.png" format="PNG"/>
1687 <!-- Fall back if none of the images can be used -->
1689 [Here there should be a diagram showing the progress of
1690 packages through the filters during a simple virtual-database
1691 search and a multi-database search, but is seems that your
1692 tool chain has not been able to include the diagram in this
1697 <para>A picture is worth a thousand words (but only five hundred on 64-bit architectures)</para>
1704 <chapter id="sru-server">
1705 <title>Combined SRU webservice and Z39.50 server configuration</title>
1707 Metaproxy can act as
1708 <ulink url="&url.sru;">SRU</ulink> and
1709 web service server, which translates web service requests to
1710 <ulink url="&url.z39.50;">ANSI/NISO Z39.50</ulink> packages and
1711 sends them off to common available targets.
1714 A typical setup for this operation needs a filter route including the
1718 <table id="sru-server-table-config" frame="top">
1719 <title>SRU/Z39.50 Server Filter Route Configuration</title>
1723 <entry>Filter</entry>
1724 <entry>Importance</entry>
1725 <entry>Purpose</entry>
1731 <entry><literal>frontend_net</literal></entry>
1732 <entry>required</entry>
1733 <entry>Accepting HTTP connections and passing them to following
1734 filters. Since this filter also accepts Z39.50 connections, the
1735 server works as SRU and Z39.50 server on the same port.</entry>
1738 <entry><literal>sru_z3950</literal></entry>
1739 <entry>required</entry>
1740 <entry>Accepting SRU GET/POST/SOAP explain and
1741 searchRetrieve requests for the the configured databases.
1742 Explain requests are directly served from the static XML configuration.
1743 SearchRetrieve requests are
1744 transformed to Z39.50 search and present packages.
1745 All other HTTP and Z39.50 packages are passed unaltered.</entry>
1748 <entry><literal>http_file</literal></entry>
1749 <entry>optional</entry>
1750 <entry>Serving HTTP requests from the filesystem. This is only
1751 needed if the server should serve XSLT stylesheets, static HTML
1752 files or Java Script for thin browser based clients.
1753 Z39.50 packages are passed unaltered.</entry>
1756 <entry><literal>cql_rpn</literal></entry>
1757 <entry>required</entry>
1758 <entry>Usually, Z39.50 servers do not talk CQL, hence the
1759 translation of the CQL query language to RPN is mandatory in
1760 most cases. Affects only Z39.50 search packages.</entry>
1763 <entry><literal>record_transform</literal></entry>
1764 <entry>optional</entry>
1765 <entry>Some Z39.50 backend targets can not present XML record
1766 syntaxes in common wanted element sets. using this filter, one
1767 can transform binary MARC records to MARCXML records, and
1768 further transform those to any needed XML schema/format by XSLT
1769 transformations. Changes only Z39.50 present packages.</entry>
1772 <entry><literal>session_shared</literal></entry>
1773 <entry>optional</entry>
1774 <entry>The stateless nature of web services requires frequent
1775 re-searching of the same targets for display of paged result set
1776 records. This might be an unacceptable burden for the accessed
1777 backend Z39.50 targets, and this mosule can be added for
1778 efficient backend target resource pooling.</entry>
1781 <entry><literal>z3950_client</literal></entry>
1782 <entry>required</entry>
1783 <entry>Finally, a Z39.50 package sink is needed in the filter
1784 chain to provide the response packages. The Z39.50 client module
1785 is used to access external targets over the network, but any
1786 coming local Z39.50 package sink could be used instead of.</entry>
1789 <entry><literal>bounce</literal></entry>
1790 <entry>required</entry>
1791 <entry>Any Metaproxy package arriving here did not do so by
1792 purpose, and is bounced back with connection closure. this
1793 prevents inifinite package hanging inside the SRU server.</entry>
1799 A typical minimal example <ulink url="&url.sru;">SRU</ulink>
1800 server configuration file is found in the tarball distribution at
1801 <literal>etc/config-sru-to-z3950.xml</literal>.
1804 Off course, any other metaproxy modules can be integrated into a
1805 SRU server solution, including, but not limited to, load balancing,
1806 multiple target querying
1807 (see <xref linkend="multidb"/>), and complex RPN query rewrites.
1814 <chapter id="extensions">
1815 <title>Writing extensions for Metaproxy</title>
1816 <para>### To be written</para>
1822 <chapter id="classes">
1823 <title>Classes in the Metaproxy source code</title>
1826 <section id="classes-introductory-notes">
1827 <title>Introductory notes</title>
1829 <emphasis>Stop! Do not read this!</emphasis>
1830 You won't enjoy it at all. You should just skip ahead to
1831 <xref linkend="reference"/>,
1833 <!-- The remainder of this paragraph is lifted verbatim from
1834 Douglas Adams' _Hitch Hiker's Guide to the Galaxy_, chapter 8 -->
1835 you things you really need to know, like the fact that the
1836 fabulously beautiful planet Bethselamin is now so worried about
1837 the cumulative erosion by ten billion visiting tourists a year
1838 that any net imbalance between the amount you eat and the amount
1839 you excrete whilst on the planet is surgically removed from your
1840 bodyweight when you leave: so every time you go to the lavatory it
1841 is vitally important to get a receipt.
1844 This chapter contains documentation of the Metaproxy source code, and is
1845 of interest only to maintainers and developers. If you need to
1846 change Metaproxy's behavior or write a new filter, then you will most
1847 likely find this chapter helpful. Otherwise it's a waste of your
1848 good time. Seriously: go and watch a film or something.
1849 <citetitle>This is Spinal Tap</citetitle> is particularly good.
1852 Still here? OK, let's continue.
1855 In general, classes seem to be named big-endianly, so that
1856 <literal>FactoryFilter</literal> is not a filter that filters
1857 factories, but a factory that produces filters; and
1858 <literal>FactoryStatic</literal> is a factory for the statically
1859 registered filters (as opposed to those that are dynamically
1864 <section id="individual.classes">
1865 <title>Individual classes</title>
1867 The classes making up the Metaproxy application are here listed by
1868 class-name, with the names of the source files that define them in
1872 <section id="class-FactoryFilter">
1873 <title><literal>mp::FactoryFilter</literal>
1874 (<filename>factory_filter.cpp</filename>)</title>
1876 A factory class that exists primarily to provide the
1877 <literal>create()</literal> method, which takes the name of a
1878 filter class as its argument and returns a new filter of that
1879 type. To enable this, the factory must first be populated by
1880 calling <literal>add_creator()</literal> for static filters (this
1881 is done by the <literal>FactoryStatic</literal> class, see below)
1882 and <literal>add_creator_dyn()</literal> for filters loaded
1887 <section id="class-FactoryStatic">
1888 <title><literal>mp::FactoryStatic</literal>
1889 (<filename>factory_static.cpp</filename>)</title>
1891 A subclass of <literal>FactoryFilter</literal> which is
1892 responsible for registering all the statically defined filter
1893 types. It does this by knowing about all those filters'
1894 structures, which are listed in its constructor. Merely
1895 instantiating this class registers all the static classes. It is
1896 for the benefit of this class that <literal>struct
1897 metaproxy_1_filter_struct</literal> exists, and that all the filter
1898 classes provide a static object of that type.
1902 <section id="class-filter-Base">
1903 <title><literal>mp::filter::Base</literal>
1904 (<filename>filter.cpp</filename>)</title>
1906 The virtual base class of all filters. The filter API is, on the
1907 surface at least, extremely simple: two methods.
1908 <literal>configure()</literal> is passed an XML DOM tree representing
1909 that part of the configuration file that pertains to this filter
1910 instance, and is expected to walk that tree extracting relevant
1911 information. And <literal>process()</literal> processes a
1912 package (see below). That surface simplicity is a bit
1913 misleading, as <literal>process()</literal> needs to know a lot
1914 about the <literal>Package</literal> class in order to do
1919 <section id="class-AuthSimple">
1920 <title><literal>mp::filter::AuthSimple</literal>,
1921 <literal>Backend_test</literal>, etc.
1922 (<filename>filter_auth_simple.cpp</filename>,
1923 <filename>filter_backend_test.cpp</filename>, etc.)</title>
1925 Individual filters. Each of these is implemented by a header and
1926 a source file, named <filename>filter_*.hpp</filename> and
1927 <filename>filter_*.cpp</filename> respectively. All the header
1928 files should be pretty much identical, in that they declare the
1929 class, including a private <literal>Rep</literal> class and a
1930 member pointer to it, and the two public methods.
1933 The source file for each filter needs to supply:
1938 A definition of the private <literal>Rep</literal> class.
1943 Some boilerplate constructors and destructors.
1948 A <literal>configure()</literal> method that uses the
1949 appropriate XML fragment.
1954 Most important, the <literal>process()</literal> method that
1955 does all the actual work.
1961 <section id="class-Package">
1962 <title><literal>mp::Package</literal>
1963 (<filename>package.cpp</filename>)</title>
1965 Represents a package on its way through the series of filters
1966 that make up a route. This is essentially a Z39.50 or SRU APDU
1967 together with information about where it came from, which is
1968 modified as it passes through the various filters.
1972 <section id="class-Pipe">
1973 <title><literal>mp::Pipe</literal>
1974 (<filename>pipe.cpp</filename>)</title>
1976 This class provides a compatibility layer so that we have an IPC
1977 mechanism that works the same under Unix and Windows. It's not
1978 particularly exciting.
1982 <section id="class-RouterChain">
1983 <title><literal>mp::RouterChain</literal>
1984 (<filename>router_chain.cpp</filename>)</title>
1990 <section id="class-RouterFleXML">
1991 <title><literal>mp::RouterFleXML</literal>
1992 (<filename>router_flexml.cpp</filename>)</title>
1998 <section id="class-Session">
1999 <title><literal>mp::Session</literal>
2000 (<filename>session.cpp</filename>)</title>
2006 <section id="class-ThreadPoolSocketObserver">
2007 <title><literal>mp::ThreadPoolSocketObserver</literal>
2008 (<filename>thread_pool_observer.cpp</filename>)</title>
2014 <section id="class-util">
2015 <title><literal>mp::util</literal>
2016 (<filename>util.cpp</filename>)</title>
2018 A namespace of various small utility functions and classes,
2019 collected together for convenience. Most importantly, includes
2020 the <literal>mp::util::odr</literal> class, a wrapper for YAZ's
2025 <section id="class-xml">
2026 <title><literal>mp::xml</literal>
2027 (<filename>xmlutil.cpp</filename>)</title>
2029 A namespace of various XML utility functions and classes,
2030 collected together for convenience.
2036 <section id="other.source.files">
2037 <title>Other Source Files</title>
2039 In addition to the Metaproxy source files that define the classes
2040 described above, there are a few additional files which are
2041 briefly described here:
2045 <term><literal>metaproxy_prog.cpp</literal></term>
2048 The main function of the <command>metaproxy</command> program.
2053 <term><literal>ex_router_flexml.cpp</literal></term>
2056 Identical to <literal>metaproxy_prog.cpp</literal>: it's not clear why.
2061 <term><literal>test_*.cpp</literal></term>
2064 Unit-tests for various modules.
2070 ### Still to be described:
2071 <literal>ex_filter_frontend_net.cpp</literal>,
2072 <literal>filter_dl.cpp</literal>,
2073 <literal>plainfile.cpp</literal>,
2074 <literal>tstdl.cpp</literal>.
2080 <reference id="reference">
2081 <title>Reference</title>
2082 <partintro id="reference-introduction">
2084 The material in this chapter is drawn directly from the individual
2085 manual entries. In particular, the Metaproxy invocation section is
2086 available using <command>man metaproxy</command>, and the section
2087 on each individual filter is available using the name of the filter
2088 as the argument to the <command>man</command> command.
2094 <appendix id="license">
2095 <title>License</title>
2100 Metaproxy is free software; you can redistribute it and/or modify it under
2101 the terms of the GNU General Public License as published by the Free
2102 Software Foundation; either version 2, or (at your option) any later
2107 Metaproxy is distributed in the hope that it will be useful, but WITHOUT ANY
2108 WARRANTY; without even the implied warranty of MERCHANTABILITY or
2109 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
2114 You should have received a copy of the GNU General Public License
2115 along with Metaproxy; see the file LICENSE. If not, write to the
2116 Free Software Foundation,
2117 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
2125 <!-- Keep this comment at the end of the file
2128 nxml-child-indent: 1