1 <chapter id="record-model">
2 <!-- $Id: recordmodel.xml,v 1.14 2002-12-03 16:10:19 mike Exp $ -->
3 <title>The Record Model</title>
6 The Zebra system is designed to support a wide range of data management
7 applications. The system can be configured to handle virtually any
8 kind of structured data. Each record in the system is associated with
9 a <emphasis>record schema</emphasis> which lends context to the data
10 elements of the record.
11 Any number of record schemas can coexist in the system.
12 Although it may be wise to use only a single schema within
13 one database, the system poses no such restrictions.
17 The record model described in this chapter applies to the fundamental,
19 record type <literal>grs</literal>, introduced in
20 <xref linkend="record-types"/>.
22 FIXME - Need to describe the simple string-tag model, or at least
28 Records pass through three different states during processing in the
38 When records are accessed by the system, they are represented
39 in their local, or native format. This might be SGML or HTML files,
40 News or Mail archives, MARC records. If the system doesn't already
41 know how to read the type of data you need to store, you can set up an
42 input filter by preparing conversion rules based on regular
43 expressions and possibly augmented by a flexible scripting language
45 The input filter produces as output an internal representation,
53 When records are processed by the system, they are represented
54 in a tree-structure, constructed by tagged data elements hanging off a
55 root node. The tagged elements may contain data or yet more tagged
56 elements in a recursive structure. The system performs various
57 actions on this tree structure (indexing, element selection, schema
65 Before transmitting records to the client, they are first
66 converted from the internal structure to a form suitable for exchange
67 over the network - according to the Z39.50 standard.
75 <sect1 id="local-representation">
76 <title>Local Representation</title>
79 As mentioned earlier, Zebra places few restrictions on the type of
80 data that you can index and manage. Generally, whatever the form of
81 the data, it is parsed by an input filter specific to that format, and
82 turned into an internal structure that Zebra knows how to handle. This
83 process takes place whenever the record is accessed - for indexing and
88 The RecordType parameter in the <literal>zebra.cfg</literal> file, or
89 the <literal>-t</literal> option to the indexer tells Zebra how to
90 process input records.
91 Two basic types of processing are available - raw text and structured
92 data. Raw text is just that, and it is selected by providing the
93 argument <emphasis>text</emphasis> to Zebra. Structured records are
94 all handled internally using the basic mechanisms described in the
96 Zebra can read structured records in many different formats.
97 How this is done is governed by additional parameters after the
98 "grs" keyword, separated by "." characters.
102 Four basic subtypes to the <emphasis>grs</emphasis> type are
109 <term>grs.sgml</term>
112 This is the canonical input format —
113 described below. It is a simple SGML-like syntax.
118 <term>grs.regx.<emphasis>filter</emphasis></term>
121 This enables a user-supplied input
122 filter. The mechanisms of these filters are described below.
127 <term>grs.tcl.<emphasis>filter</emphasis></term>
130 Similar to grs.regx but using Tcl for rules.
135 <term>grs.marc.<emphasis>abstract syntax</emphasis></term>
138 This allows Zebra to read
139 records in the ISO2709 (MARC) encoding standard. In this case, the
140 last parameter <emphasis>abstract syntax</emphasis> names the
141 <literal>.abs</literal> file (see below)
142 which describes the specific MARC structure of the input record as
143 well as the indexing rules.
151 This filter reads XML records. Only one record per file
152 is supported. The filter is only available if Zebra/YAZ
153 is compiled with EXPAT support.
162 <title>Canonical Input Format</title>
165 Although input data can take any form, it is sometimes useful to
166 describe the record processing capabilities of the system in terms of
167 a single, canonical input format that gives access to the full
168 spectrum of structure and flexibility in the system. In Zebra, this
169 canonical format is an "SGML-like" syntax.
173 To use the canonical format specify <literal>grs.sgml</literal> as
178 Consider a record describing an information resource (such a record is
179 sometimes known as a <emphasis>locator record</emphasis>).
180 It might contain a field describing the distributor of the
181 information resource, which might in turn be partitioned into
182 various fields providing details about the distributor, like this:
188 <Distributor>
189 <Name> USGS/WRD </Name>
190 <Organization> USGS/WRD </Organization>
191 <Street-Address>
192 U.S. GEOLOGICAL SURVEY, 505 MARQUETTE, NW
193 </Street-Address>
194 <City> ALBUQUERQUE </City>
195 <State> NM </State>
196 <Zip-Code> 87102 </Zip-Code>
197 <Country> USA </Country>
198 <Telephone> (505) 766-5560 </Telephone>
199 </Distributor>
204 <!-- There is no indentation in the example above! -H
207 The indentation used above is used to illustrate how Zebra
208 interprets the mark-up. The indentation, in itself, has no
209 significance to the parser for the canonical input format, which
210 discards superfluous whitespace.
216 The keywords surrounded by <...> are
217 <emphasis>tags</emphasis>, while the sections of text
218 in between are the <emphasis>data elements</emphasis>.
219 A data element is characterized by its location in the tree
220 that is made up by the nested elements.
221 Each element is terminated by a closing tag - beginning
222 with <literal><</literal>/, and containing the same symbolic
223 tag-name as the corresponding opening tag.
224 The general closing tag - <literal></></literal> -
225 terminates the element started by the last opening tag. The
226 structuring of elements is significant.
227 The element <emphasis>Telephone</emphasis>,
228 for instance, may be indexed and presented to the client differently,
229 depending on whether it appears inside the
230 <emphasis>Distributor</emphasis> element, or some other,
231 structured data element such a <emphasis>Supplier</emphasis> element.
235 <title>Record Root</title>
238 The first tag in a record describes the root node of the tree that
239 makes up the total record. In the canonical input format, the root tag
240 should contain the name of the schema that lends context to the
241 elements of the record
242 (see <xref linkend="internal-representation"/>).
243 The following is a GILS record that
244 contains only a single element (strictly speaking, that makes it an
245 illegal GILS record, since the GILS profile includes several mandatory
246 elements - Zebra does not validate the contents of a record against
247 the Z39.50 profile, however - it merely attempts to match up elements
248 of a local representation with the given schema):
255 <title>Zen and the Art of Motorcycle Maintenance</title>
263 <sect3><!-- ### we shouldn't make such a big deal about this -->
264 <title>Variants</title>
267 Zebra allows you to provide individual data elements in a number of
268 <emphasis>variant forms</emphasis>. Examples of variant forms are
269 textual data elements which might appear in different languages, and
270 images which may appear in different formats or layouts.
271 The variant system in Zebra is essentially a representation of
272 the variant mechanism of Z39.50-1995.
276 The following is an example of a title element which occurs in two
284 <var lang lang "eng">
285 Zen and the Art of Motorcycle Maintenance</>
286 <var lang lang "dan">
287 Zen og Kunsten at Vedligeholde en Motorcykel</>
294 The syntax of the <emphasis>variant element</emphasis> is
295 <literal><var class type value></literal>.
296 The available values for the <emphasis>class</emphasis> and
297 <emphasis>type</emphasis> fields are given by the variant set
298 that is associated with the current schema
299 (see <xref linkend="variant-set"/>).
303 Variant elements are terminated by the general end-tag </>, by
304 the variant end-tag </var>, by the appearance of another variant
305 tag with the same <emphasis>class</emphasis> and
306 <emphasis>value</emphasis> settings, or by the
307 appearance of another, normal tag. In other words, the end-tags for
308 the variants used in the example above could have been omitted.
312 Variant elements can be nested. The element
319 <var lang lang "eng"><var body iana "text/plain">
320 Zen and the Art of Motorcycle Maintenance
327 Associates two variant components to the variant list for the title
332 Given the nesting rules described above, we could write
339 <var body iana "text/plain>
340 <var lang lang "eng">
341 Zen and the Art of Motorcycle Maintenance
342 <var lang lang "dan">
343 Zen og Kunsten at Vedligeholde en Motorcykel
350 The title element above comes in two variants. Both have the IANA body
351 type "text/plain", but one is in English, and the other in
352 Danish. The client, using the element selection mechanism of Z39.50,
353 can retrieve information about the available variant forms of data
354 elements, or it can select specific variants based on the requirements
363 <title>Input Filters</title>
366 In order to handle general input formats, Zebra allows the
367 operator to define filters which read individual records in their
368 native format and produce an internal representation that the system
373 Input filters are ASCII files, generally with the suffix
374 <literal>.flt</literal>.
375 The system looks for the files in the directories given in the
376 <emphasis>profilePath</emphasis> setting in the
377 <literal>zebra.cfg</literal> files.
378 The record type for the filter is
379 <literal>grs.regx.</literal><emphasis>filter-filename</emphasis>
380 (fundamental type <literal>grs</literal>, file read
381 type <literal>regx</literal>, argument
382 <emphasis>filter-filename</emphasis>).
386 Generally, an input filter consists of a sequence of rules, where each
387 rule consists of a sequence of expressions, followed by an action. The
388 expressions are evaluated against the contents of the input record,
389 and the actions normally contribute to the generation of an internal
390 representation of the record.
394 An expression can be either of the following:
404 The action associated with this expression is evaluated
405 exactly once in the lifetime of the application, before any records
406 are read. It can be used in conjunction with an action that
407 initializes tables or other resources that are used in the processing
416 Matches the beginning of the record. It can be used to
417 initialize variables, etc. Typically, the
418 <emphasis>BEGIN</emphasis> rule is also used
419 to establish the root node of the record.
427 Matches the end of the record - when all of the contents
428 of the record has been processed.
433 <term>/pattern/</term>
436 Matches a string of characters from the input record.
444 This keyword may only be used between two patterns.
445 It matches everything between (not including) those patterns.
453 The expression associated with this pattern is evaluated
454 once, before the application terminates. It can be used to release
455 system resources - typically ones allocated in the
456 <emphasis>INIT</emphasis> step.
464 An action is surrounded by curly braces ({...}), and
465 consists of a sequence of statements. Statements may be separated
466 by newlines or semicolons (;).
467 Within actions, the strings that matched the expressions
468 immediately preceding the action can be referred to as
473 The available statements are:
480 <term>begin <emphasis>type [parameter ... ]</emphasis></term>
484 data element. The type is one of the following:
491 Begin a new record. The following parameter should be the
492 name of the schema that describes the structure of the record, eg.
493 <literal>gils</literal> or <literal>wais</literal> (see below).
494 The <literal>begin record</literal> call should precede
495 any other use of the <emphasis>begin</emphasis> statement.
503 Begin a new tagged element. The parameter is the
504 name of the tag. If the tag is not matched anywhere in the tagsets
505 referenced by the current schema, it is treated as a local string
514 Begin a new node in a variant tree. The parameters are
515 <emphasis>class type value</emphasis>.
527 Create a data element. The concatenated arguments make
528 up the value of the data element.
529 The option <literal>-text</literal> signals that
530 the layout (whitespace) of the data should be retained for
532 The option <literal>-element</literal>
533 <emphasis>tag</emphasis> wraps the data up in
534 the <emphasis>tag</emphasis>.
535 The use of the <literal>-element</literal> option is equivalent to
536 preceding the command with a <emphasis>begin
537 element</emphasis> command, and following
538 it with the <emphasis>end</emphasis> command.
543 <term>end <emphasis>[type]</emphasis></term>
546 Close a tagged element. If no parameter is given,
547 the last element on the stack is terminated.
548 The first parameter, if any, is a type name, similar
549 to the <emphasis>begin</emphasis> statement.
550 For the <emphasis>element</emphasis> type, a tag
551 name can be provided to terminate a specific tag.
559 The following input filter reads a Usenet news file, producing a
560 record in the WAIS schema. Note that the body of a news posting is
561 separated from the list of headers by a blank line (or rather a
562 sequence of two newline characters.
568 BEGIN { begin record wais }
570 /^From:/ BODY /$/ { data -element name $1 }
571 /^Subject:/ BODY /$/ { data -element title $1 }
572 /^Date:/ BODY /$/ { data -element lastModified $1 }
574 begin element bodyOfDisplay
575 begin variant body iana "text/plain"
584 If Zebra is compiled with support for Tcl (Tool Command Language)
585 enabled, the statements described above are supplemented with a complete
586 scripting environment, including control structures (conditional
587 expressions and loop constructs), and powerful string manipulation
588 mechanisms for modifying the elements of a record. Tcl is a popular
589 scripting environment, with several tutorials available both online
597 <sect1 id="internal-representation">
598 <title>Internal Representation</title>
601 When records are manipulated by the system, they're represented in a
602 tree-structure, with data elements at the leaf nodes, and tags or
603 variant components at the non-leaf nodes. The root-node identifies the
604 schema that lends context to the tagging and structuring of the
605 record. Imagine a simple record, consisting of a 'title' element and
613 TITLE "Zen and the Art of Motorcycle Maintenance"
614 AUTHOR "Robert Pirsig"
620 A slightly more complex record would have the author element consist
621 of two elements, a surname and a first name:
628 TITLE "Zen and the Art of Motorcycle Maintenance"
637 The root of the record will refer to the record schema that describes
638 the structuring of this particular record. The schema defines the
639 element tags (TITLE, FIRST-NAME, etc.) that may occur in the record, as
640 well as the structuring (SURNAME should appear below AUTHOR, etc.). In
641 addition, the schema establishes element set names that are used by
642 the client to request a subset of the elements of a given record. The
643 schema may also establish rules for converting the record to a
644 different schema, by stating, for each element, a mapping to a
649 <title>Tagged Elements</title>
652 A data element is characterized by its tag, and its position in the
653 structure of the record. For instance, while the tag "telephone
654 number" may be used different places in a record, we may need to
655 distinguish between these occurrences, both for searching and
656 presentation purposes. For instance, while the phone numbers for the
657 "customer" and the "service provider" are both
658 representatives for the same type of resource (a telephone number), it
659 is essential that they be kept separate. The record schema provides
660 the structure of the record, and names each data element (defined by
661 the sequence of tags - the tag path - by which the element can be
662 reached from the root of the record).
668 <title>Variants</title>
671 The children of a tag node may be either more tag nodes, a data node
672 (possibly accompanied by tag nodes),
673 or a tree of variant nodes. The children of variant nodes are either
674 more variant nodes or a data node (possibly accompanied by more
675 variant nodes). Each leaf node, which is normally a
676 data node, corresponds to a <emphasis>variant form</emphasis> of the
677 tagged element identified by the tag which parents the variant tree.
678 The following title element occurs in two different languages:
684 VARIANT LANG=ENG "War and Peace"
686 VARIANT LANG=DAN "Krig og Fred"
692 Which of the two elements are transmitted to the client by the server
693 depends on the specifications provided by the client, if any.
697 In practice, each variant node is associated with a triple of class,
698 type, value, corresponding to the variant mechanism of Z39.50.
704 <title>Data Elements</title>
707 Data nodes have no children (they are always leaf nodes in the record
712 FIXME! Documentation needs extension here about types of nodes - numerical,
713 textual, etc., plus the various types of inclusion notes.
721 <sect1 id="data-model">
722 <title>Configuring Your Data Model</title>
725 The following sections describe the configuration files that govern
726 the internal management of data records. The system searches for the files
727 in the directories specified by the <emphasis>profilePath</emphasis>
728 setting in the <literal>zebra.cfg</literal> file.
732 <title>The Abstract Syntax</title>
735 The abstract syntax definition (also known as an Abstract Record
736 Structure, or ARS) is the focal point of the
737 record schema description. For a given schema, the ABS file may state any
738 or all of the following:
742 FIXME - Need a diagram here, or a simple explanation how it all hangs together -H
751 The object identifier of the Z39.50 schema associated
752 with the ARS, so that it can be referred to by the client.
758 The attribute set (which can possibly be a compound of multiple
759 sets) which applies in the profile. This is used when indexing and
760 searching the records belonging to the given profile.
766 The tag set (again, this can consist of several different sets).
767 This is used when reading the records from a file, to recognize the
768 different tags, and when transmitting the record to the client -
769 mapping the tags to their numerical representation, if they are
776 The variant set which is used in the profile. This provides a
777 vocabulary for specifying the <emphasis>forms</emphasis> of
778 data that appear inside the records.
784 Element set names, which are a shorthand way for the client to
785 ask for a subset of the data elements contained in a record. Element
786 set names, in the retrieval module, are mapped to <emphasis>element
787 specifications</emphasis>, which contain information equivalent to the
788 <emphasis>Espec-1</emphasis> syntax of Z39.50.
794 Map tables, which may specify mappings to
795 <emphasis>other</emphasis> database profiles, if desired.
801 Possibly, a set of rules describing the mapping of elements to a
809 A list of element descriptions (this is the actual ARS of the
810 schema, in Z39.50 terms), which lists the ways in which the various
811 tags can be used and organized hierarchically.
820 Several of the entries above simply refer to other files, which
821 describe the given objects.
827 <title>The Configuration Files</title>
830 This section describes the syntax and use of the various tables which
831 are used by the retrieval module.
835 The number of different file types may appear daunting at first, but
836 each type corresponds fairly clearly to a single aspect of the Z39.50
837 retrieval facilities. Further, the average database administrator,
838 who is simply reusing an existing profile for which tables already
839 exist, shouldn't have to worry too much about the contents of these tables.
843 Generally, the files are simple ASCII files, which can be maintained
844 using any text editor. Blank lines, and lines beginning with a (#) are
845 ignored. Any characters on a line followed by a (#) are also ignored.
846 All other lines contain <emphasis>directives</emphasis>, which provide
847 some setting or value to the system.
848 Generally, settings are characterized by a single
849 keyword, identifying the setting, followed by a number of parameters.
850 Some settings are repeatable (r), while others may occur only once in a
851 file. Some settings are optional (o), while others again are
857 <sect2 id="abs-file">
858 <title>The Abstract Syntax (.abs) Files</title>
861 The name of this file type is slightly misleading in Z39.50 terms,
862 since, apart from the actual abstract syntax of the profile, it also
863 includes most of the other definitions that go into a database
868 When a record in the canonical, SGML-like format is read from a file
869 or from the database, the first tag of the file should reference the
870 profile that governs the layout of the record. If the first tag of the
871 record is, say, <literal><gils></literal>, the system will look
872 for the profile definition in the file <literal>gils.abs</literal>.
873 Profile definitions are cached, so they only have to be read once
874 during the lifespan of the current process.
878 When writing your own input filters, the
879 <emphasis>record-begin</emphasis> command
880 introduces the profile, and should always be called first thing when
881 introducing a new record.
885 The file may contain the following directives:
892 <term>name <replaceable>symbolic-name</replaceable></term>
895 (m) This provides a shorthand name or
896 description for the profile. Mostly useful for diagnostic purposes.
901 <term>reference <replaceable>OID-name</replaceable></term>
904 (m) The reference name of the OID for the profile.
905 The reference names can be found in the <emphasis>util</emphasis>
911 <term>attset <replaceable>filename</replaceable></term>
914 (m) The attribute set that is used for
915 indexing and searching records belonging to this profile.
920 <term>tagset <replaceable>filename</replaceable></term>
923 (o) The tag set (if any) that describe
924 that fields of the records.
929 <term>varset <replaceable>filename</replaceable></term>
932 (o) The variant set used in the profile.
937 <term>maptab <replaceable>filename</replaceable></term>
940 (o,r) This points to a
941 conversion table that might be used if the client asks for the record
942 in a different schema from the native one.
947 <term>marc <replaceable>filename</replaceable></term>
950 (o) Points to a file containing parameters
951 for representing the record contents in the ISO2709 syntax.
952 Read the description of the MARC representation facility below.
957 <term>esetname <replaceable>name filename</replaceable></term>
961 given element set name with an element selection file. If an (@) is
962 given in place of the filename, this corresponds to a null mapping for
963 the given element set name.
968 <term>any <replaceable>tags</replaceable></term>
971 (o) This directive specifies a list of attributes
972 which should be appended to the attribute list given for each
973 element. The effect is to make every single element in the abstract
974 syntax searchable by way of the given attributes. This directive
975 provides an efficient way of supporting free-text searching across all
976 elements. However, it does increase the size of the index
977 significantly. The attributes can be qualified with a structure, as in
978 the <replaceable>elm</replaceable> directive below.
983 <term>elm <replaceable>path name attributes</replaceable></term>
986 (o,r) Adds an element to the abstract record syntax of the schema.
987 The <replaceable>path</replaceable> follows the
988 syntax which is suggested by the Z39.50 document - that is, a sequence
989 of tags separated by slashes (/). Each tag is given as a
990 comma-separated pair of tag type and -value surrounded by parenthesis.
991 The <replaceable>name</replaceable> is the name of the element, and
992 the <replaceable>attributes</replaceable>
993 specifies which attributes to use when indexing the element in a
994 comma-separated list.
995 A ! in place of the attribute name is equivalent to
996 specifying an attribute name identical to the element name.
997 A - in place of the attribute name
998 specifies that no indexing is to take place for the given element.
999 The attributes can be qualified with <replaceable>field
1000 types</replaceable> to specify which
1001 character set should govern the indexing procedure for that field.
1002 The same data element may be indexed into several different
1003 fields, using different character set definitions.
1004 See the <xref linkend="field-structure-and-character-sets"/>.
1005 The default field type is <literal>w</literal> for
1006 <emphasis>word</emphasis>.
1012 <term>xelm <replaceable>xpath attributes</replaceable></term>
1015 Specifies indexing for record nodes given by
1016 <replaceable>xpath</replaceable>. Unlike directive
1017 elm, this directive allows you to index attribute
1018 contents. The <replaceable>xpath</replaceable> uses
1019 a syntax similar to XPath. The <replaceable>attributes</replaceable>
1020 have same syntax and meaning as directive elm, except that !
1021 refers to the nodes selected by <replaceable>xpath</replaceable>.
1023 xelm / !:w default index
1024 xelm // !:w additional index
1025 xelm /gils/title/@att myatt:w index attribute @att in myatt
1026 xelm title/@att myatt:w same meaning.
1033 <term>encoding <replaceable>encodingname</replaceable></term>
1036 This directive specifies character encoding for external records.
1037 For records such as XML that specifies encoding within the
1038 file via a header this directive is ignored.
1039 If neither this directive is given, nor an encoding is set
1040 within external records, ISO-8859-1 encoding is assumed.
1045 <term>xpath <literal>enable</literal>/<literal>disable</literal></term>
1048 If this directive is followed by <literal>enable</literal>,
1049 then extra indexing is performed to allow for XPath-like queries.
1050 If this directive is not specified - equivalent to
1051 <literal>disable</literal> - no extra XPath-indexing is performed.
1058 <replaceable>systemTag</replaceable>
1059 <replaceable>actualTag</replaceable>
1063 Specifies what information, if any, Zebra should
1064 automatically include in retrieval records for the
1065 ``system fields'' that it supports.
1066 <replaceable>systemTag</replaceable> may
1067 be any of the following:
1070 <term><literal>rank</literal></term>
1072 An integer indicating the relevance-ranking score
1073 assigned to the record.
1077 <term><literal>sysno</literal></term>
1079 An automatically generated identifier for the record,
1080 unique within this database. It is represented by the
1081 <literal><localControlNumber></literal> element in
1082 XML and the <literal>(1,14)</literal> tag in GRS-1.
1086 <term><literal>size</literal></term>
1088 The size, in bytes, of the retrieved record.
1094 The <replaceable>actualTag</replaceable> parameter may be
1095 <literal>none</literal> to indicate that the named element
1096 should be omitted from retrieval records.
1105 The mechanism for controlling indexing is not adequate for
1106 complex databases, and will probably be moved into a separate
1107 configuration table eventually.
1112 The following is an excerpt from the abstract syntax file for the GILS
1120 reference GILS-schema
1125 maptab gils-usmarc.map
1129 esetname VARIANT gils-variant.est # for WAIS-compliance
1130 esetname B gils-b.est
1131 esetname G gils-g.est
1136 elm (1,14) localControlNumber Local-number
1137 elm (1,16) dateOfLastModification Date/time-last-modified
1138 elm (2,1) title w:!,p:!
1139 elm (4,1) controlIdentifier Identifier-standard
1140 elm (2,6) abstract Abstract
1141 elm (4,51) purpose !
1142 elm (4,52) originator -
1143 elm (4,53) accessConstraints !
1144 elm (4,54) useConstraints !
1145 elm (4,70) availability -
1146 elm (4,70)/(4,90) distributor -
1147 elm (4,70)/(4,90)/(2,7) distributorName !
1148 elm (4,70)/(4,90)/(2,10 distributorOrganization !
1149 elm (4,70)/(4,90)/(4,2) distributorStreetAddress !
1150 elm (4,70)/(4,90)/(4,3) distributorCity !
1157 <sect2 id="attset-files">
1158 <title>The Attribute Set (.att) Files</title>
1161 This file type describes the <replaceable>Use</replaceable> elements of
1163 It contains the following directives.
1169 <term>name <replaceable>symbolic-name</replaceable></term>
1172 (m) This provides a shorthand name or
1173 description for the attribute set.
1174 Mostly useful for diagnostic purposes.
1176 </listitem></varlistentry>
1178 <term>reference <replaceable>OID-name</replaceable></term>
1181 (m) The reference name of the OID for
1183 The reference names can be found in the <replaceable>util</replaceable>
1184 module of <replaceable>YAZ</replaceable>.
1186 </listitem></varlistentry>
1188 <term>include <replaceable>filename</replaceable></term>
1191 (o,r) This directive is used to
1192 include another attribute set as a part of the current one. This is
1193 used when a new attribute set is defined as an extension to another
1194 set. For instance, many new attribute sets are defined as extensions
1195 to the <replaceable>bib-1</replaceable> set.
1196 This is an important feature of the retrieval
1197 system of Z39.50, as it ensures the highest possible level of
1198 interoperability, as those access points of your database which are
1199 derived from the external set (say, bib-1) can be used even by clients
1200 who are unaware of the new set.
1202 </listitem></varlistentry>
1205 <replaceable>att-value att-name [local-value]</replaceable></term>
1209 repeatable directive introduces a new attribute to the set. The
1210 attribute value is stored in the index (unless a
1211 <replaceable>local-value</replaceable> is
1212 given, in which case this is stored). The name is used to refer to the
1213 attribute from the <replaceable>abstract syntax</replaceable>.
1215 </listitem></varlistentry>
1220 This is an excerpt from the GILS attribute set definition.
1221 Notice how the file describing the <emphasis>bib-1</emphasis>
1222 attribute set is referenced.
1229 reference GILS-attset
1232 att 2001 distributorName
1233 att 2002 indextermsControlled
1235 att 2004 accessConstraints
1236 att 2005 useConstraints
1244 <title>The Tag Set (.tag) Files</title>
1247 This file type defines the tagset of the profile, possibly by
1248 referencing other tag sets (most tag sets, for instance, will include
1249 tagsetG and tagsetM from the Z39.50 specification. The file may
1250 contain the following directives.
1257 <term>name <emphasis>symbolic-name</emphasis></term>
1260 (m) This provides a shorthand name or
1261 description for the tag set. Mostly useful for diagnostic purposes.
1263 </listitem></varlistentry>
1265 <term>reference <emphasis>OID-name</emphasis></term>
1268 (o) The reference name of the OID for the tag set.
1269 The reference names can be found in the <emphasis>util</emphasis>
1270 module of <emphasis>YAZ</emphasis>.
1271 The directive is optional, since not all tag sets
1272 are registered outside of their schema.
1274 </listitem></varlistentry>
1276 <term>type <emphasis>integer</emphasis></term>
1279 (m) The type number of the tagset within the schema
1280 profile (note: this specification really should belong to the .abs
1281 file. This will be fixed in a future release).
1283 </listitem></varlistentry>
1285 <term>include <emphasis>filename</emphasis></term>
1288 (o,r) This directive is used
1289 to include the definitions of other tag sets into the current one.
1291 </listitem></varlistentry>
1293 <term>tag <emphasis>number names type</emphasis></term>
1296 (o,r) Introduces a new tag to the set.
1297 The <emphasis>number</emphasis> is the tag number as used
1298 in the protocol (there is currently no mechanism for
1299 specifying string tags at this point, but this would be quick
1301 The <emphasis>names</emphasis> parameter is a list of names
1302 by which the tag should be recognized in the input file format.
1303 The names should be separated by slashes (/).
1304 The <emphasis>type</emphasis> is the recommended data type of
1306 It should be one of the following:
1372 </listitem></varlistentry>
1377 The following is an excerpt from the TagsetG definition file.
1388 tag 3 publicationPlace string
1389 tag 4 publicationDate string
1390 tag 5 documentId string
1391 tag 6 abstract string
1393 tag 8 date generalizedtime
1394 tag 9 bodyOfDisplay string
1395 tag 10 organization string
1401 <sect2 id="variant-set">
1402 <title>The Variant Set (.var) Files</title>
1405 The variant set file is a straightforward representation of the
1406 variant set definitions associated with the protocol. At present, only
1407 the <emphasis>Variant-1</emphasis> set is known.
1411 These are the directives allowed in the file.
1418 <term>name <emphasis>symbolic-name</emphasis></term>
1421 (m) This provides a shorthand name or
1422 description for the variant set. Mostly useful for diagnostic purposes.
1424 </listitem></varlistentry>
1426 <term>reference <emphasis>OID-name</emphasis></term>
1429 (o) The reference name of the OID for
1430 the variant set, if one is required. The reference names can be found
1431 in the <emphasis>util</emphasis> module of <emphasis>YAZ</emphasis>.
1433 </listitem></varlistentry>
1435 <term>class <emphasis>integer class-name</emphasis></term>
1438 (m,r) Introduces a new
1439 class to the variant set.
1441 </listitem></varlistentry>
1443 <term>type <emphasis>integer type-name datatype</emphasis></term>
1447 new type to the current class (the one introduced by the most recent
1448 <emphasis>class</emphasis> directive).
1449 The type names belong to the same name space as the one used
1450 in the tag set definition file.
1452 </listitem></varlistentry>
1457 The following is an excerpt from the file describing the variant set
1458 <emphasis>Variant-1</emphasis>.
1469 type 1 variantId octetstring
1474 type 2 z39.50 string
1483 <title>The Element Set (.est) Files</title>
1486 The element set specification files describe a selection of a subset
1487 of the elements of a database record. The element selection mechanism
1488 is equivalent to the one supplied by the <emphasis>Espec-1</emphasis>
1489 syntax of the Z39.50 specification.
1490 In fact, the internal representation of an element set
1491 specification is identical to the <emphasis>Espec-1</emphasis> structure,
1492 and we'll refer you to the description of that structure for most of
1493 the detailed semantics of the directives below.
1498 Not all of the Espec-1 functionality has been implemented yet.
1499 The fields that are mentioned below all work as expected, unless
1505 The directives available in the element set file are as follows:
1511 <term>defaultVariantSetId <emphasis>OID-name</emphasis></term>
1514 (o) If variants are used in
1515 the following, this should provide the name of the variantset used
1516 (it's not currently possible to specify a different set in the
1517 individual variant request). In almost all cases (certainly all
1518 profiles known to us), the name
1519 <literal>Variant-1</literal> should be given here.
1521 </listitem></varlistentry>
1523 <term>defaultVariantRequest <emphasis>variant-request</emphasis></term>
1527 provides a default variant request for
1528 use when the individual element requests (see below) do not contain a
1529 variant request. Variant requests consist of a blank-separated list of
1530 variant components. A variant compont is a comma-separated,
1531 parenthesized triple of variant class, type, and value (the two former
1532 values being represented as integers). The value can currently only be
1533 entered as a string (this will change to depend on the definition of
1534 the variant in question). The special value (@) is interpreted as a
1535 null value, however.
1537 </listitem></varlistentry>
1540 <emphasis>path ['variant' variant-request]</emphasis></term>
1543 (o,r) This corresponds to a simple element request
1544 in <emphasis>Espec-1</emphasis>.
1545 The path consists of a sequence of tag-selectors, where each of
1546 these can consist of either:
1553 A simple tag, consisting of a comma-separated type-value pair in
1554 parenthesis, possibly followed by a colon (:) followed by an
1555 occurrences-specification (see below). The tag-value can be a number
1556 or a string. If the first character is an apostrophe ('), this
1557 forces the value to be interpreted as a string, even if it
1558 appears to be numerical.
1564 A WildThing, represented as a question mark (?), possibly
1565 followed by a colon (:) followed by an occurrences
1566 specification (see below).
1572 A WildPath, represented as an asterisk (*). Note that the last
1573 element of the path should not be a wildPath (wildpaths don't
1574 work in this version).
1583 The occurrences-specification can be either the string
1584 <literal>all</literal>, the string <literal>last</literal>, or
1585 an explicit value-range. The value-range is represented as
1586 an integer (the starting point), possibly followed by a
1587 plus (+) and a second integer (the number of elements, default
1592 The variant-request has the same syntax as the defaultVariantRequest
1593 above. Note that it may sometimes be useful to give an empty variant
1594 request, simply to disable the default for a specific set of fields
1595 (we aren't certain if this is proper <emphasis>Espec-1</emphasis>,
1596 but it works in this implementation).
1598 </listitem></varlistentry>
1603 The following is an example of an element specification belonging to
1610 simpleelement (1,10)
1611 simpleelement (1,12)
1613 simpleelement (1,14)
1615 simpleelement (4,52)
1622 <sect2 id="schema-mapping">
1623 <title>The Schema Mapping (.map) Files</title>
1626 Sometimes, the client might want to receive a database record in
1627 a schema that differs from the native schema of the record. For
1628 instance, a client might only know how to process WAIS records, while
1629 the database record is represented in a more specific schema, such as
1630 GILS. In this module, a mapping of data to one of the MARC formats is
1631 also thought of as a schema mapping (mapping the elements of the
1632 record into fields consistent with the given MARC specification, prior
1633 to actually converting the data to the ISO2709). This use of the
1634 object identifier for USMARC as a schema identifier represents an
1635 overloading of the OID which might not be entirely proper. However,
1636 it represents the dual role of schema and record syntax which
1637 is assumed by the MARC family in Z39.50.
1641 <emphasis>NOTE: FIXME! The schema-mapping functions are so far limited to a
1642 straightforward mapping of elements. This should be extended with
1643 mechanisms for conversions of the element contents, and conditional
1644 mappings of elements based on the record contents.</emphasis>
1648 These are the directives of the schema mapping file format:
1655 <term>targetName <emphasis>name</emphasis></term>
1658 (m) A symbolic name for the target schema
1659 of the table. Useful mostly for diagnostic purposes.
1661 </listitem></varlistentry>
1663 <term>targetRef <emphasis>OID-name</emphasis></term>
1666 (m) An OID name for the target schema.
1667 This is used, for instance, by a server receiving a request to present
1668 a record in a different schema from the native one.
1669 The name, again, is found in the <emphasis>oid</emphasis>
1670 module of <emphasis>YAZ</emphasis>.
1672 </listitem></varlistentry>
1674 <term>map <emphasis>element-name target-path</emphasis></term>
1678 an element mapping rule to the table.
1680 </listitem></varlistentry>
1687 <title>The MARC (ISO2709) Representation (.mar) Files</title>
1690 This file provides rules for representing a record in the ISO2709
1691 format. The rules pertain mostly to the values of the constant-length
1692 header of the record.
1696 NOTE: FIXME! This will be described better. We're in the process of
1697 re-evaluating and most likely changing the way that MARC records are
1698 handled by the system.</emphasis>
1703 <sect2 id="field-structure-and-character-sets">
1704 <title>Field Structure and Character Sets
1708 In order to provide a flexible approach to national character set
1709 handling, Zebra allows the administrator to configure the set up the
1710 system to handle any 8-bit character set — including sets that
1711 require multi-octet diacritics or other multi-octet characters. The
1712 definition of a character set includes a specification of the
1713 permissible values, their sort order (this affects the display in the
1714 SCAN function), and relationships between upper- and lowercase
1715 characters. Finally, the definition includes the specification of
1716 space characters for the set.
1720 The operator can define different character sets for different fields,
1721 typical examples being standard text fields, numerical fields, and
1722 special-purpose fields such as WWW-style linkages (URx).
1726 The field types, and hence character sets, are associated with data
1727 elements by the .abs files (see above).
1728 The file <literal>default.idx</literal>
1729 provides the association between field type codes (as used in the .abs
1730 files) and the character map files (with the .chr suffix). The format
1731 of the .idx file is as follows
1738 <term>index <emphasis>field type code</emphasis></term>
1741 This directive introduces a new search index code.
1742 The argument is a one-character code to be used in the
1743 .abs files to select this particular index type. An index, roughly,
1744 corresponds to a particular structure attribute during search. Refer
1745 to <xref linkend="search"/>.
1747 </listitem></varlistentry>
1749 <term>sort <emphasis>field code type</emphasis></term>
1752 This directive introduces a
1753 sort index. The argument is a one-character code to be used in the
1754 .abs fie to select this particular index type. The corresponding
1755 use attribute must be used in the sort request to refer to this
1756 particular sort index. The corresponding character map (see below)
1757 is used in the sort process.
1759 </listitem></varlistentry>
1761 <term>completeness <emphasis>boolean</emphasis></term>
1764 This directive enables or disables complete field indexing.
1765 The value of the <emphasis>boolean</emphasis> should be 0
1766 (disable) or 1. If completeness is enabled, the index entry will
1767 contain the complete contents of the field (up to a limit), with words
1768 (non-space characters) separated by single space characters
1769 (normalized to " " on display). When completeness is
1770 disabled, each word is indexed as a separate entry. Complete subfield
1771 indexing is most useful for fields which are typically browsed (eg.
1772 titles, authors, or subjects), or instances where a match on a
1773 complete subfield is essential (eg. exact title searching). For fields
1774 where completeness is disabled, the search engine will interpret a
1775 search containing space characters as a word proximity search.
1777 </listitem></varlistentry>
1779 <term>charmap <emphasis>filename</emphasis></term>
1782 This is the filename of the character
1783 map to be used for this index for field type.
1785 </listitem></varlistentry>
1790 The contents of the character map files are structured as follows:
1797 <term>lowercase <emphasis>value-set</emphasis></term>
1800 This directive introduces the basic value set of the field type.
1801 The format is an ordered list (without spaces) of the
1802 characters which may occur in "words" of the given type.
1803 The order of the entries in the list determines the
1804 sort order of the index. In addition to single characters, the
1805 following combinations are legal:
1813 Backslashes may be used to introduce three-digit octal, or
1814 two-digit hex representations of single characters
1815 (preceded by <literal>x</literal>).
1816 In addition, the combinations
1817 \\, \\r, \\n, \\t, \\s (space — remember that real
1818 space-characters may not occur in the value definition), and
1819 \\ are recognized, with their usual interpretation.
1825 Curly braces {} may be used to enclose ranges of single
1826 characters (possibly using the escape convention described in the
1827 preceding point), eg. {a-z} to introduce the
1828 standard range of ASCII characters.
1829 Note that the interpretation of such a range depends on
1830 the concrete representation in your local, physical character set.
1836 paranthesises () may be used to enclose multi-byte characters -
1837 eg. diacritics or special national combinations (eg. Spanish
1838 "ll"). When found in the input stream (or a search term),
1839 these characters are viewed and sorted as a single character, with a
1840 sorting value depending on the position of the group in the value
1848 </listitem></varlistentry>
1850 <term>uppercase <emphasis>value-set</emphasis></term>
1853 This directive introduces the
1854 upper-case equivalencis to the value set (if any). The number and
1855 order of the entries in the list should be the same as in the
1856 <literal>lowercase</literal> directive.
1858 </listitem></varlistentry>
1860 <term>space <emphasis>value-set</emphasis></term>
1863 This directive introduces the character
1864 which separate words in the input stream. Depending on the
1865 completeness mode of the field in question, these characters either
1866 terminate an index entry, or delimit individual "words" in
1867 the input stream. The order of the elements is not significant —
1868 otherwise the representation is the same as for the
1869 <literal>uppercase</literal> and <literal>lowercase</literal>
1872 </listitem></varlistentry>
1874 <term>map <emphasis>value-set</emphasis>
1875 <emphasis>target</emphasis></term>
1878 This directive introduces a
1879 mapping between each of the members of the value-set on the left to
1880 the character on the right. The character on the right must occur in
1881 the value set (the <literal>lowercase</literal> directive) of
1882 the character set, but
1883 it may be a paranthesis-enclosed multi-octet character. This directive
1884 may be used to map diacritics to their base characters, or to map
1885 HTML-style character-representations to their natural form, etc.
1887 </listitem></varlistentry>
1895 <sect1 id="formats">
1896 <title>Exchange Formats</title>
1899 Converting records from the internal structure to en exchange format
1900 is largely an automatic process. Currently, the following exchange
1901 formats are supported:
1908 GRS-1. The internal representation is based on GRS-1/XML, so the
1909 conversion here is straightforward. The system will create
1910 applied variant and supported variant lists as required, if a record
1911 contains variant information.
1917 XML. The internal representation is based on GRS-1/XML so
1918 the mapping is trivial. Note that XML schemas, preprocessing
1919 instructions and comments are not part of the internal representation
1920 and therefore will never be part of a generated XML record.
1921 Future versions of the Zebra will support that.
1927 SUTRS. Again, the mapping is fairly straightforward. Indentation
1928 is used to show the hierarchical structure of the record. All
1929 "GRS" type records support both the GRS-1 and SUTRS
1931 <!-- FIXME - What is SUTRS - should be expanded here -->
1937 ISO2709-based formats (USMARC, etc.). Only records with a
1938 two-level structure (corresponding to fields and subfields) can be
1939 directly mapped to ISO2709. For records with a different structuring
1940 (eg., GILS), the representation in a structure like USMARC involves a
1941 schema-mapping (see <xref linkend="schema-mapping"/>), to an
1942 "implied" USMARC schema (implied,
1943 because there is no formal schema which specifies the use of the
1944 USMARC fields outside of ISO2709). The resultant, two-level record is
1945 then mapped directly from the internal representation to ISO2709. See
1946 the GILS schema definition files for a detailed example of this
1953 Explain. This representation is only available for records
1954 belonging to the Explain schema.
1960 Summary. This ASN-1 based structure is only available for records
1961 belonging to the Summary schema - or schema which provide a mapping
1962 to this schema (see the description of the schema mapping facility
1969 SOIF. Support for this syntax is experimental, and is currently
1970 keyed to a private Index Data OID (1.2.840.10003.5.1000.81.2). All
1971 abstract syntaxes can be mapped to the SOIF format, although nested
1972 elements are represented by concatenation of the tag names at each
1974 <!-- FIXME - Is this used anywhere ? -H -->
1983 <!-- Keep this comment at the end of the file
1988 sgml-minimize-attributes:nil
1989 sgml-always-quote-attributes:t
1992 sgml-parent-document: "zebra.xml"
1993 sgml-local-catalogs: nil
1994 sgml-namecase-general:t