1 <chapter id="record-model">
2 <!-- $Id: recordmodel.xml,v 1.3 2002-04-10 14:47:49 heikki 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 schema 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> as introduced in
20 <xref linkend="record-types"/>.
21 FIXME - Need to describe the simple string-tag model, or at least
26 Records pass through three different states during processing in the
36 When records are accessed by the system, they are represented
37 in their local, or native format. This might be SGML or HTML files,
38 News or Mail archives, MARC records. If the system doesn't already
39 know how to read the type of data you need to store, you can set up an
40 input filter by preparing conversion rules based on regular
41 expressions and possibly augmented by a flexible scripting language
43 The input filter produces as output an internal representation:
50 When records are processed by the system, they are represented
51 in a tree-structure, constructed by tagged data elements hanging off a
52 root node. The tagged elements may contain data or yet more tagged
53 elements in a recursive structure. The system performs various
54 actions on this tree structure (indexing, element selection, schema
62 Before transmitting records to the client, they are first
63 converted from the internal structure to a form suitable for exchange
64 over the network - according to the Z39.50 standard.
72 <sect1 id="local-representation">
73 <title>Local Representation</title>
76 As mentioned earlier, Zebra places few restrictions on the type of
77 data that you can index and manage. Generally, whatever the form of
78 the data, it is parsed by an input filter specific to that format, and
79 turned into an internal structure that Zebra knows how to handle. This
80 process takes place whenever the record is accessed - for indexing and
85 The RecordType parameter in the <literal>zebra.cfg</literal> file, or
86 the <literal>-t</literal> option to the indexer tells Zebra how to
87 process input records.
88 Two basic types of processing are available - raw text and structured
89 data. Raw text is just that, and it is selected by providing the
90 argument <emphasis>text</emphasis> to Zebra. Structured records are
91 all handled internally using the basic mechanisms described in the
93 Zebra can read structured records in many different formats.
94 How this is done is governed by additional parameters after the
95 "grs" keyboard, separated by "." characters.
99 Four basic subtypes to the <emphasis>grs</emphasis> type are
106 <term>grs.sgml</term>
109 This is the canonical input format —
110 described below. It is a simple SGML-like syntax.
115 <term>grs.regx.<emphasis>filter</emphasis></term>
118 This enables a user-supplied input
119 filter. The mechanisms of these filters are described below.
124 <term>grs.tcl.<emphasis>filter</emphasis></term>
127 Similar to grs.regx but using Tcl for rules.
132 <term>grs.marc.<emphasis>abstract syntax</emphasis></term>
135 This allows Zebra to read
136 records in the ISO2709 (MARC) encoding standard. In this case, the
137 last parameter <emphasis>abstract syntax</emphasis> names the
138 <literal>.abs</literal> file (see below)
139 which describes the specific MARC structure of the input record as
140 well as the indexing rules.
148 <title>Canonical Input Format</title>
151 Although input data can take any form, it is sometimes useful to
152 describe the record processing capabilities of the system in terms of
153 a single, canonical input format that gives access to the full
154 spectrum of structure and flexibility in the system. In Zebra, this
155 canonical format is an "SGML-like" syntax.
159 To use the canonical format specify <literal>grs.sgml</literal> as
164 Consider a record describing an information resource (such a record is
165 sometimes known as a <emphasis>locator record</emphasis>).
166 It might contain a field describing the distributor of the
167 information resource, which might in turn be partitioned into
168 various fields providing details about the distributor, like this:
174 <Distributor>
175 <Name> USGS/WRD </Name>
176 <Organization> USGS/WRD </Organization>
177 <Street-Address>
178 U.S. GEOLOGICAL SURVEY, 505 MARQUETTE, NW
179 </Street-Address>
180 <City> ALBUQUERQUE </City>
181 <State> NM </State>
182 <Zip-Code> 87102 </Zip-Code>
183 <Country> USA </Country>
184 <Telephone> (505) 766-5560 </Telephone>
185 </Distributor>
192 The indentation used above is used to illustrate how Zebra
193 interprets the mark-up. The indentation, in itself, has no
194 significance to the parser for the canonical input format, which
195 discards superfluous whitespace.
199 The keywords surrounded by <...> are
200 <emphasis>tags</emphasis>, while the sections of text
201 in between are the <emphasis>data elements</emphasis>.
202 A data element is characterized by its location in the tree
203 that is made up by the nested elements.
204 Each element is terminated by a closing tag - beginning
205 with <literal><</literal>/, and containing the same symbolic
206 tag-name as the corresponding opening tag.
207 The general closing tag - <literal><</literal>>/ -
208 terminates the element started by the last opening tag. The
209 structuring of elements is significant.
210 The element <emphasis>Telephone</emphasis>,
211 for instance, may be indexed and presented to the client differently,
212 depending on whether it appears inside the
213 <emphasis>Distributor</emphasis> element, or some other,
214 structured data element such a <emphasis>Supplier</emphasis> element.
218 <title>Record Root</title>
221 The first tag in a record describes the root node of the tree that
222 makes up the total record. In the canonical input format, the root tag
223 should contain the name of the schema that lends context to the
224 elements of the record
225 (see <xref linkend="internal-representation"/>).
226 The following is a GILS record that
227 contains only a single element (strictly speaking, that makes it an
228 illegal GILS record, since the GILS profile includes several mandatory
229 elements - Zebra does not validate the contents of a record against
230 the Z39.50 profile, however - it merely attempts to match up elements
231 of a local representation with the given schema):
238 <title>Zen and the Art of Motorcycle Maintenance</title>
247 <title>Variants</title>
250 Zebra allows you to provide individual data elements in a number of
251 <emphasis>variant forms</emphasis>. Examples of variant forms are
252 textual data elements which might appear in different languages, and
253 images which may appear in different formats or layouts.
254 The variant system in Zebra is essentially a representation of
255 the variant mechanism of Z39.50-1995.
259 The following is an example of a title element which occurs in two
267 <var lang lang "eng">
268 Zen and the Art of Motorcycle Maintenance</>
269 <var lang lang "dan">
270 Zen og Kunsten at Vedligeholde en Motorcykel</>
277 The syntax of the <emphasis>variant element</emphasis> is
278 <literal><var class type value></literal>.
279 The available values for the <emphasis>class</emphasis> and
280 <emphasis>type</emphasis> fields are given by the variant set
281 that is associated with the current schema
282 (see <xref linkend="variant-set"/>).
286 Variant elements are terminated by the general end-tag </>, by
287 the variant end-tag </var>, by the appearance of another variant
288 tag with the same <emphasis>class</emphasis> and
289 <emphasis>value</emphasis> settings, or by the
290 appearance of another, normal tag. In other words, the end-tags for
291 the variants used in the example above could have been saved.
295 Variant elements can be nested. The element
302 <var lang lang "eng"><var body iana "text/plain">
303 Zen and the Art of Motorcycle Maintenance
310 Associates two variant components to the variant list for the title
315 Given the nesting rules described above, we could write
322 <var body iana "text/plain>
323 <var lang lang "eng">
324 Zen and the Art of Motorcycle Maintenance
325 <var lang lang "dan">
326 Zen og Kunsten at Vedligeholde en Motorcykel
333 The title element above comes in two variants. Both have the IANA body
334 type "text/plain", but one is in English, and the other in
335 Danish. The client, using the element selection mechanism of Z39.50,
336 can retrieve information about the available variant forms of data
337 elements, or it can select specific variants based on the requirements
346 <title>Input Filters</title>
349 In order to handle general input formats, Zebra allows the
350 operator to define filters which read individual records in their
351 native format and produce an internal representation that the system
356 Input filters are ASCII files, generally with the suffix
357 <literal>.flt</literal>.
358 The system looks for the files in the directories given in the
359 <emphasis>profilePath</emphasis> setting in the
360 <literal>zebra.cfg</literal> files.
361 The record type for the filter is
362 <literal>grs.regx.</literal><emphasis>filter-filename</emphasis>
363 (fundamental type <literal>grs</literal>, file read
364 type <literal>regx</literal>, argument
365 <emphasis>filter-filename</emphasis>).
369 Generally, an input filter consists of a sequence of rules, where each
370 rule consists of a sequence of expressions, followed by an action. The
371 expressions are evaluated against the contents of the input record,
372 and the actions normally contribute to the generation of an internal
373 representation of the record.
377 An expression can be either of the following:
387 The action associated with this expression is evaluated
388 exactly once in the lifetime of the application, before any records
389 are read. It can be used in conjunction with an action that
390 initializes tables or other resources that are used in the processing
399 Matches the beginning of the record. It can be used to
400 initialize variables, etc. Typically, the
401 <emphasis>BEGIN</emphasis> rule is also used
402 to establish the root node of the record.
410 Matches the end of the record - when all of the contents
411 of the record has been processed.
416 <term>/pattern/</term>
419 Matches a string of characters from the input record.
427 This keyword may only be used between two patterns.
428 It matches everything between (not including) those patterns.
436 The expression associated with this pattern is evaluated
437 once, before the application terminates. It can be used to release
438 system resources - typically ones allocated in the
439 <emphasis>INIT</emphasis> step.
447 An action is surrounded by curly braces ({...}), and
448 consists of a sequence of statements. Statements may be separated
449 by newlines or semicolons (;).
450 Within actions, the strings that matched the expressions
451 immediately preceding the action can be referred to as
452 $0, $1, $2, etc.
456 The available statements are:
463 <term>begin <emphasis>type [parameter ... ]</emphasis></term>
467 data element. The type is one of the following:
474 Begin a new record. The following parameter should be the
475 name of the schema that describes the structure of the record, eg.
476 <literal>gils</literal> or <literal>wais</literal> (see below).
477 The <literal>begin record</literal> call should precede
478 any other use of the <emphasis>begin</emphasis> statement.
486 Begin a new tagged element. The parameter is the
487 name of the tag. If the tag is not matched anywhere in the tagsets
488 referenced by the current schema, it is treated as a local string
497 Begin a new node in a variant tree. The parameters are
498 <emphasis>class type value</emphasis>.
510 Create a data element. The concatenated arguments make
511 up the value of the data element.
512 The option <literal>-text</literal> signals that
513 the layout (whitespace) of the data should be retained for
515 The option <literal>-element</literal>
516 <emphasis>tag</emphasis> wraps the data up in
517 the <emphasis>tag</emphasis>.
518 The use of the <literal>-element</literal> option is equivalent to
519 preceding the command with a <emphasis>begin
520 element</emphasis> command, and following
521 it with the <emphasis>end</emphasis> command.
526 <term>end <emphasis>[type]</emphasis></term>
529 Close a tagged element. If no parameter is given,
530 the last element on the stack is terminated.
531 The first parameter, if any, is a type name, similar
532 to the <emphasis>begin</emphasis> statement.
533 For the <emphasis>element</emphasis> type, a tag
534 name can be provided to terminate a specific tag.
542 The following input filter reads a Usenet news file, producing a
543 record in the WAIS schema. Note that the body of a news posting is
544 separated from the list of headers by a blank line (or rather a
545 sequence of two newline characters.
551 BEGIN { begin record wais }
553 /^From:/ BODY /$/ { data -element name $1 }
554 /^Subject:/ BODY /$/ { data -element title $1 }
555 /^Date:/ BODY /$/ { data -element lastModified $1 }
557 begin element bodyOfDisplay
558 begin variant body iana "text/plain"
567 If Zebra is compiled with support for Tcl (Tool Command Language)
568 enabled, the statements described above are supplemented with a complete
569 scripting environment, including control structures (conditional
570 expressions and loop constructs), and powerful string manipulation
571 mechanisms for modifying the elements of a record. Tcl is a popular
572 scripting environment, with several tutorials available both online
580 <sect1 id="internal-representation">
581 <title>Internal Representation</title>
584 When records are manipulated by the system, they're represented in a
585 tree-structure, with data elements at the leaf nodes, and tags or
586 variant components at the non-leaf nodes. The root-node identifies the
587 schema that lends context to the tagging and structuring of the
588 record. Imagine a simple record, consisting of a 'title' element and
595 TITLE "Zen and the Art of Motorcycle Maintenance"
597 AUTHOR "Robert Pirsig"
603 A slightly more complex record would have the author element consist
604 of two elements, a surname and a first name:
610 TITLE "Zen and the Art of Motorcycle Maintenance"
620 The root of the record will refer to the record schema that describes
621 the structuring of this particular record. The schema defines the
622 element tags (TITLE, FIRST-NAME, etc.) that may occur in the record, as
623 well as the structuring (SURNAME should appear below AUTHOR, etc.). In
624 addition, the schema establishes element set names that are used by
625 the client to request a subset of the elements of a given record. The
626 schema may also establish rules for converting the record to a
627 different schema, by stating, for each element, a mapping to a
632 <title>Tagged Elements</title>
635 A data element is characterized by its tag, and its position in the
636 structure of the record. For instance, while the tag "telephone
637 number" may be used different places in a record, we may need to
638 distinguish between these occurrences, both for searching and
639 presentation purposes. For instance, while the phone numbers for the
640 "customer" and the "service provider" are both
641 representatives for the same type of resource (a telephone number), it
642 is essential that they be kept separate. The record schema provides
643 the structure of the record, and names each data element (defined by
644 the sequence of tags - the tag path - by which the element can be
645 reached from the root of the record).
651 <title>Variants</title>
654 The children of a tag node may be either more tag nodes, a data node
655 (possibly accompanied by tag nodes),
656 or a tree of variant nodes. The children of variant nodes are either
657 more variant nodes or a data node (possibly accompanied by more
658 variant nodes). Each leaf node, which is normally a
659 data node, corresponds to a <emphasis>variant form</emphasis> of the
660 tagged element identified by the tag which parents the variant tree.
661 The following title element occurs in two different languages:
667 VARIANT LANG=ENG "War and Peace"
669 VARIANT LANG=DAN "Krig og Fred"
675 Which of the two elements are transmitted to the client by the server
676 depends on the specifications provided by the client, if any.
680 In practice, each variant node is associated with a triple of class,
681 type, value, corresponding to the variant mechanism of Z39.50.
687 <title>Data Elements</title>
690 Data nodes have no children (they are always leaf nodes in the record
696 FIXME! Documentation needs extension here about types of nodes - numerical,
697 textual, etc., plus the various types of inclusion notes.
705 <sect1 id="data-model">
706 <title>Configuring Your Data Model</title>
709 The following sections describe the configuration files that govern
710 the internal management of data records. The system searches for the files
711 in the directories specified by the <emphasis>profilePath</emphasis>
712 setting in the <literal>zebra.cfg</literal> file.
716 <title>The Abstract Syntax</title>
719 The abstract syntax definition (also known as an Abstract Record
720 Structure, or ARS) is the focal point of the
721 record schema description. For a given schema, the ABS file may state any
722 or all of the following:
726 FIXME - Need a diagram here, or a simple explanation how it all hangs together -H
735 The object identifier of the Z39.50 schema associated
736 with the ARS, so that it can be referred to by the client.
742 The attribute set (which can possibly be a compound of multiple
743 sets) which applies in the profile. This is used when indexing and
744 searching the records belonging to the given profile.
750 The Tag set (again, this can consist of several different sets).
751 This is used when reading the records from a file, to recognize the
752 different tags, and when transmitting the record to the client -
753 mapping the tags to their numerical representation, if they are
760 The variant set which is used in the profile. This provides a
761 vocabulary for specifying the <emphasis>forms</emphasis> of data that appear inside
768 Element set names, which are a shorthand way for the client to
769 ask for a subset of the data elements contained in a record. Element
770 set names, in the retrieval module, are mapped to <emphasis>element
771 specifications</emphasis>, which contain information equivalent to the
772 <emphasis>Espec-1</emphasis> syntax of Z39.50.
778 Map tables, which may specify mappings to
779 <emphasis>other</emphasis> database profiles, if desired.
785 Possibly, a set of rules describing the mapping of elements to a
793 A list of element descriptions (this is the actual ARS of the
794 schema, in Z39.50 terms), which lists the ways in which the various
795 tags can be used and organized hierarchically.
804 Several of the entries above simply refer to other files, which
805 describe the given objects.
811 <title>The Configuration Files</title>
814 This section describes the syntax and use of the various tables which
815 are used by the retrieval module.
819 The number of different file types may appear daunting at first, but
820 each type corresponds fairly clearly to a single aspect of the Z39.50
821 retrieval facilities. Further, the average database administrator,
822 who is simply reusing an existing profile for which tables already
823 exist, shouldn't have to worry too much about the contents of these tables.
827 Generally, the files are simple ASCII files, which can be maintained
828 using any text editor. Blank lines, and lines beginning with a (#) are
829 ignored. Any characters on a line followed by a (#) are also ignored.
830 All other lines contain <emphasis>directives</emphasis>, which provide
831 some setting or value to the system.
832 Generally, settings are characterized by a single
833 keyword, identifying the setting, followed by a number of parameters.
834 Some settings are repeatable (r), while others may occur only once in a
835 file. Some settings are optional (o), while others again are
842 <title>The Abstract Syntax (.abs) Files</title>
845 The name of this file type is slightly misleading in Z39.50 terms,
846 since, apart from the actual abstract syntax of the profile, it also
847 includes most of the other definitions that go into a database
852 When a record in the canonical, SGML-like format is read from a file
853 or from the database, the first tag of the file should reference the
854 profile that governs the layout of the record. If the first tag of the
855 record is, say, <literal><gils></literal>, the system will look
856 for the profile definition in the file <literal>gils.abs</literal>.
857 Profile definitions are cached, so they only have to be read once
858 during the lifespan of the current process.
862 When writing your own input filters, the
863 <emphasis>record-begin</emphasis> command
864 introduces the profile, and should always be called first thing when
865 introducing a new record.
869 The file may contain the following directives:
876 <term>name <emphasis>symbolic-name</emphasis></term>
879 (m) This provides a shorthand name or
880 description for the profile. Mostly useful for diagnostic purposes.
885 <term>reference <emphasis>OID-name</emphasis></term>
888 (m) The reference name of the OID for the profile.
889 The reference names can be found in the <emphasis>util</emphasis>
890 module of <emphasis>YAZ</emphasis>.
895 <term>attset <emphasis>filename</emphasis></term>
898 (m) The attribute set that is used for
899 indexing and searching records belonging to this profile.
904 <term>tagset <emphasis>filename</emphasis></term>
907 (o) The tag set (if any) that describe
908 that fields of the records.
913 <term>varset <emphasis>filename</emphasis></term>
916 (o) The variant set used in the profile.
921 <term>maptab <emphasis>filename</emphasis></term>
924 (o,r) This points to a
925 conversion table that might be used if the client asks for the record
926 in a different schema from the native one.
928 </listitem></varlistentry>
930 <term>marc <emphasis>filename</emphasis></term>
933 (o) Points to a file containing parameters
934 for representing the record contents in the ISO2709 syntax. Read the
935 description of the MARC representation facility below.
937 </listitem></varlistentry>
939 <term>esetname <emphasis>name filename</emphasis></term>
943 given element set name with an element selection file. If an (@) is
944 given in place of the filename, this corresponds to a null mapping for
945 the given element set name.
947 </listitem></varlistentry>
949 <term>any <emphasis>tags</emphasis></term>
952 (o) This directive specifies a list of attributes
953 which should be appended to the attribute list given for each
954 element. The effect is to make every single element in the abstract
955 syntax searchable by way of the given attributes. This directive
956 provides an efficient way of supporting free-text searching across all
957 elements. However, it does increase the size of the index
958 significantly. The attributes can be qualified with a structure, as in
959 the <emphasis>elm</emphasis> directive below.
961 </listitem></varlistentry>
963 <term>elm <emphasis>path name attributes</emphasis></term>
966 (o,r) Adds an element to the abstract record syntax of the schema.
967 The <emphasis>path</emphasis> follows the
968 syntax which is suggested by the Z39.50 document - that is, a sequence
969 of tags separated by slashes (/). Each tag is given as a
970 comma-separated pair of tag type and -value surrounded by parenthesis.
971 The <emphasis>name</emphasis> is the name of the element, and
972 the <emphasis>attributes</emphasis>
973 specifies which attributes to use when indexing the element in a
974 comma-separated list.
975 A ! in place of the attribute name is equivalent to
976 specifying an attribute name identical to the element name.
977 A - in place of the attribute name
978 specifies that no indexing is to take place for the given element.
979 The attributes can be qualified with <emphasis>field
980 types</emphasis> to specify which
981 character set should govern the indexing procedure for that field.
982 The same data element may be indexed into several different
983 fields, using different character set definitions.
984 See the <xref linkend="field-structure-and-character-sets"/>.
985 The default field type is "w" for <emphasis>word</emphasis>.
987 </listitem></varlistentry>
993 The mechanism for controlling indexing is not adequate for
994 complex databases, and will probably be moved into a separate
995 configuration table eventually.
1000 The following is an excerpt from the abstract syntax file for the GILS
1008 reference GILS-schema
1013 maptab gils-usmarc.map
1017 esetname VARIANT gils-variant.est # for WAIS-compliance
1018 esetname B gils-b.est
1019 esetname G gils-g.est
1024 elm (1,14) localControlNumber Local-number
1025 elm (1,16) dateOfLastModification Date/time-last-modified
1026 elm (2,1) title w:!,p:!
1027 elm (4,1) controlIdentifier Identifier-standard
1028 elm (2,6) abstract Abstract
1029 elm (4,51) purpose !
1030 elm (4,52) originator -
1031 elm (4,53) accessConstraints !
1032 elm (4,54) useConstraints !
1033 elm (4,70) availability -
1034 elm (4,70)/(4,90) distributor -
1035 elm (4,70)/(4,90)/(2,7) distributorName !
1036 elm (4,70)/(4,90)/(2,10 distributorOrganization !
1037 elm (4,70)/(4,90)/(4,2) distributorStreetAddress !
1038 elm (4,70)/(4,90)/(4,3) distributorCity !
1045 <sect2 id="attset-files">
1046 <title>The Attribute Set (.att) Files</title>
1049 This file type describes the <emphasis>Use</emphasis> elements of
1051 It contains the following directives.
1057 <term>name <emphasis>symbolic-name</emphasis></term>
1060 (m) This provides a shorthand name or
1061 description for the attribute set.
1062 Mostly useful for diagnostic purposes.
1064 </listitem></varlistentry>
1066 <term>reference <emphasis>OID-name</emphasis></term>
1069 (m) The reference name of the OID for
1071 The reference names can be found in the <emphasis>util</emphasis>
1072 module of <emphasis>YAZ</emphasis>.
1074 </listitem></varlistentry>
1076 <term>include <emphasis>filename</emphasis></term>
1079 (o,r) This directive is used to
1080 include another attribute set as a part of the current one. This is
1081 used when a new attribute set is defined as an extension to another
1082 set. For instance, many new attribute sets are defined as extensions
1083 to the <emphasis>bib-1</emphasis> set.
1084 This is an important feature of the retrieval
1085 system of Z39.50, as it ensures the highest possible level of
1086 interoperability, as those access points of your database which are
1087 derived from the external set (say, bib-1) can be used even by clients
1088 who are unaware of the new set.
1090 </listitem></varlistentry>
1093 <emphasis>att-value att-name [local-value]</emphasis></term>
1097 repeatable directive introduces a new attribute to the set. The
1098 attribute value is stored in the index (unless a
1099 <emphasis>local-value</emphasis> is
1100 given, in which case this is stored). The name is used to refer to the
1101 attribute from the <emphasis>abstract syntax</emphasis>.
1103 </listitem></varlistentry>
1108 This is an excerpt from the GILS attribute set definition.
1109 Notice how the file describing the <emphasis>bib-1</emphasis>
1110 attribute set is referenced.
1117 reference GILS-attset
1120 att 2001 distributorName
1121 att 2002 indextermsControlled
1123 att 2004 accessConstraints
1124 att 2005 useConstraints
1132 <title>The Tag Set (.tag) Files</title>
1135 This file type defines the tagset of the profile, possibly by
1136 referencing other tag sets (most tag sets, for instance, will include
1137 tagsetG and tagsetM from the Z39.50 specification. The file may
1138 contain the following directives.
1145 <term>name <emphasis>symbolic-name</emphasis></term>
1148 (m) This provides a shorthand name or
1149 description for the tag set. Mostly useful for diagnostic purposes.
1151 </listitem></varlistentry>
1153 <term>reference <emphasis>OID-name</emphasis></term>
1156 (o) The reference name of the OID for the tag set.
1157 The reference names can be found in the <emphasis>util</emphasis>
1158 module of <emphasis>YAZ</emphasis>.
1159 The directive is optional, since not all tag sets
1160 are registered outside of their schema.
1162 </listitem></varlistentry>
1164 <term>type <emphasis>integer</emphasis></term>
1167 (m) The type number of the tagset within the schema
1168 profile (note: this specification really should belong to the .abs
1169 file. This will be fixed in a future release).
1171 </listitem></varlistentry>
1173 <term>include <emphasis>filename</emphasis></term>
1176 (o,r) This directive is used
1177 to include the definitions of other tag sets into the current one.
1179 </listitem></varlistentry>
1181 <term>tag <emphasis>number names type</emphasis></term>
1184 (o,r) Introduces a new tag to the set.
1185 The <emphasis>number</emphasis> is the tag number as used
1186 in the protocol (there is currently no mechanism for
1187 specifying string tags at this point, but this would be quick
1189 The <emphasis>names</emphasis> parameter is a list of names
1190 by which the tag should be recognized in the input file format.
1191 The names should be separated by slashes (/).
1192 The <emphasis>type</emphasis> is the recommended data type of
1194 It should be one of the following:
1260 </listitem></varlistentry>
1265 The following is an excerpt from the TagsetG definition file.
1276 tag 3 publicationPlace string
1277 tag 4 publicationDate string
1278 tag 5 documentId string
1279 tag 6 abstract string
1281 tag 8 date generalizedtime
1282 tag 9 bodyOfDisplay string
1283 tag 10 organization string
1289 <sect2 id="variant-set">
1290 <title>The Variant Set (.var) Files</title>
1293 The variant set file is a straightforward representation of the
1294 variant set definitions associated with the protocol. At present, only
1295 the <emphasis>Variant-1</emphasis> set is known.
1299 These are the directives allowed in the file.
1306 <term>name <emphasis>symbolic-name</emphasis></term>
1309 (m) This provides a shorthand name or
1310 description for the variant set. Mostly useful for diagnostic purposes.
1312 </listitem></varlistentry>
1314 <term>reference <emphasis>OID-name</emphasis></term>
1317 (o) The reference name of the OID for
1318 the variant set, if one is required. The reference names can be found
1319 in the <emphasis>util</emphasis> module of <emphasis>YAZ</emphasis>.
1321 </listitem></varlistentry>
1323 <term>class <emphasis>integer class-name</emphasis></term>
1326 (m,r) Introduces a new
1327 class to the variant set.
1329 </listitem></varlistentry>
1331 <term>type <emphasis>integer type-name datatype</emphasis></term>
1335 new type to the current class (the one introduced by the most recent
1336 <emphasis>class</emphasis> directive).
1337 The type names belong to the same name space as the one used
1338 in the tag set definition file.
1340 </listitem></varlistentry>
1345 The following is an excerpt from the file describing the variant set
1346 <emphasis>Variant-1</emphasis>.
1357 type 1 variantId octetstring
1362 type 2 z39.50 string
1371 <title>The Element Set (.est) Files</title>
1374 The element set specification files describe a selection of a subset
1375 of the elements of a database record. The element selection mechanism
1376 is equivalent to the one supplied by the <emphasis>Espec-1</emphasis>
1377 syntax of the Z39.50 specification.
1378 In fact, the internal representation of an element set
1379 specification is identical to the <emphasis>Espec-1</emphasis> structure,
1380 and we'll refer you to the description of that structure for most of
1381 the detailed semantics of the directives below.
1386 Not all of the Espec-1 functionality has been implemented yet.
1387 The fields that are mentioned below all work as expected, unless
1393 The directives available in the element set file are as follows:
1399 <term>defaultVariantSetId <emphasis>OID-name</emphasis></term>
1402 (o) If variants are used in
1403 the following, this should provide the name of the variantset used
1404 (it's not currently possible to specify a different set in the
1405 individual variant request). In almost all cases (certainly all
1406 profiles known to us), the name
1407 <literal>Variant-1</literal> should be given here.
1409 </listitem></varlistentry>
1411 <term>defaultVariantRequest <emphasis>variant-request</emphasis></term>
1415 provides a default variant request for
1416 use when the individual element requests (see below) do not contain a
1417 variant request. Variant requests consist of a blank-separated list of
1418 variant components. A variant compont is a comma-separated,
1419 parenthesized triple of variant class, type, and value (the two former
1420 values being represented as integers). The value can currently only be
1421 entered as a string (this will change to depend on the definition of
1422 the variant in question). The special value (@) is interpreted as a
1423 null value, however.
1425 </listitem></varlistentry>
1428 <emphasis>path ['variant' variant-request]</emphasis></term>
1431 (o,r) This corresponds to a simple element request
1432 in <emphasis>Espec-1</emphasis>.
1433 The path consists of a sequence of tag-selectors, where each of
1434 these can consist of either:
1441 A simple tag, consisting of a comma-separated type-value pair in
1442 parenthesis, possibly followed by a colon (:) followed by an
1443 occurrences-specification (see below). The tag-value can be a number
1444 or a string. If the first character is an apostrophe ('), this
1445 forces the value to be interpreted as a string, even if it
1446 appears to be numerical.
1452 A WildThing, represented as a question mark (?), possibly
1453 followed by a colon (:) followed by an occurrences
1454 specification (see below).
1460 A WildPath, represented as an asterisk (*). Note that the last
1461 element of the path should not be a wildPath (wildpaths don't
1462 work in this version).
1471 The occurrences-specification can be either the string
1472 <literal>all</literal>, the string <literal>last</literal>, or
1473 an explicit value-range. The value-range is represented as
1474 an integer (the starting point), possibly followed by a
1475 plus (+) and a second integer (the number of elements, default
1480 The variant-request has the same syntax as the defaultVariantRequest
1481 above. Note that it may sometimes be useful to give an empty variant
1482 request, simply to disable the default for a specific set of fields
1483 (we aren't certain if this is proper <emphasis>Espec-1</emphasis>,
1484 but it works in this implementation).
1486 </listitem></varlistentry>
1491 The following is an example of an element specification belonging to
1498 simpleelement (1,10)
1499 simpleelement (1,12)
1501 simpleelement (1,14)
1503 simpleelement (4,52)
1510 <sect2 id="schema-mapping">
1511 <title>The Schema Mapping (.map) Files</title>
1514 Sometimes, the client might want to receive a database record in
1515 a schema that differs from the native schema of the record. For
1516 instance, a client might only know how to process WAIS records, while
1517 the database record is represented in a more specific schema, such as
1518 GILS. In this module, a mapping of data to one of the MARC formats is
1519 also thought of as a schema mapping (mapping the elements of the
1520 record into fields consistent with the given MARC specification, prior
1521 to actually converting the data to the ISO2709). This use of the
1522 object identifier for USMARC as a schema identifier represents an
1523 overloading of the OID which might not be entirely proper. However,
1524 it represents the dual role of schema and record syntax which
1525 is assumed by the MARC family in Z39.50.
1529 <emphasis>NOTE: FIXME! The schema-mapping functions are so far limited to a
1530 straightforward mapping of elements. This should be extended with
1531 mechanisms for conversions of the element contents, and conditional
1532 mappings of elements based on the record contents.</emphasis>
1536 These are the directives of the schema mapping file format:
1543 <term>targetName <emphasis>name</emphasis></term>
1546 (m) A symbolic name for the target schema
1547 of the table. Useful mostly for diagnostic purposes.
1549 </listitem></varlistentry>
1551 <term>targetRef <emphasis>OID-name</emphasis></term>
1554 (m) An OID name for the target schema.
1555 This is used, for instance, by a server receiving a request to present
1556 a record in a different schema from the native one.
1557 The name, again, is found in the <emphasis>oid</emphasis>
1558 module of <emphasis>YAZ</emphasis>.
1560 </listitem></varlistentry>
1562 <term>map <emphasis>element-name target-path</emphasis></term>
1566 an element mapping rule to the table.
1568 </listitem></varlistentry>
1575 <title>The MARC (ISO2709) Representation (.mar) Files</title>
1578 This file provides rules for representing a record in the ISO2709
1579 format. The rules pertain mostly to the values of the constant-length
1580 header of the record.
1584 <emphasis>NOTE: FIXME! This will be described better. We're in the process of
1585 re-evaluating and most likely changing the way that MARC records are
1586 handled by the system.</emphasis>
1591 <sect2 id="field-structure-and-character-sets">
1592 <title>Field Structure and Character Sets
1596 In order to provide a flexible approach to national character set
1597 handling, Zebra allows the administrator to configure the set up the
1598 system to handle any 8-bit character set — including sets that
1599 require multi-octet diacritics or other multi-octet characters. The
1600 definition of a character set includes a specification of the
1601 permissible values, their sort order (this affects the display in the
1602 SCAN function), and relationships between upper- and lowercase
1603 characters. Finally, the definition includes the specification of
1604 space characters for the set.
1608 The operator can define different character sets for different fields,
1609 typical examples being standard text fields, numerical fields, and
1610 special-purpose fields such as WWW-style linkages (URx).
1614 The field types, and hence character sets, are associated with data
1615 elements by the .abs files (see above).
1616 The file <literal>default.idx</literal>
1617 provides the association between field type codes (as used in the .abs
1618 files) and the character map files (with the .chr suffix). The format
1619 of the .idx file is as follows
1626 <term>index <emphasis>field type code</emphasis></term>
1629 This directive introduces a new search index code.
1630 The argument is a one-character code to be used in the
1631 .abs files to select this particular index type. An index, roughly,
1632 corresponds to a particular structure attribute during search. Refer
1633 to <xref linkend="search"/>.
1635 </listitem></varlistentry>
1637 <term>sort <emphasis>field code type</emphasis></term>
1640 This directive introduces a
1641 sort index. The argument is a one-character code to be used in the
1642 .abs fie to select this particular index type. The corresponding
1643 use attribute must be used in the sort request to refer to this
1644 particular sort index. The corresponding character map (see below)
1645 is used in the sort process.
1647 </listitem></varlistentry>
1649 <term>completeness <emphasis>boolean</emphasis></term>
1652 This directive enables or disables complete field indexing.
1653 The value of the <emphasis>boolean</emphasis> should be 0
1654 (disable) or 1. If completeness is enabled, the index entry will
1655 contain the complete contents of the field (up to a limit), with words
1656 (non-space characters) separated by single space characters
1657 (normalized to " " on display). When completeness is
1658 disabled, each word is indexed as a separate entry. Complete subfield
1659 indexing is most useful for fields which are typically browsed (eg.
1660 titles, authors, or subjects), or instances where a match on a
1661 complete subfield is essential (eg. exact title searching). For fields
1662 where completeness is disabled, the search engine will interpret a
1663 search containing space characters as a word proximity search.
1665 </listitem></varlistentry>
1667 <term>charmap <emphasis>filename</emphasis></term>
1670 This is the filename of the character
1671 map to be used for this index for field type.
1673 </listitem></varlistentry>
1678 The contents of the character map files are structured as follows:
1685 <term>lowercase <emphasis>value-set</emphasis></term>
1688 This directive introduces the basic value set of the field type.
1689 The format is an ordered list (without spaces) of the
1690 characters which may occur in "words" of the given type.
1691 The order of the entries in the list determines the
1692 sort order of the index. In addition to single characters, the
1693 following combinations are legal:
1701 Backslashes may be used to introduce three-digit octal, or
1702 two-digit hex representations of single characters
1703 (preceded by <literal>x</literal>).
1704 In addition, the combinations
1705 \\, \\r, \\n, \\t, \\s (space — remember that real
1706 space-characters may not occur in the value definition), and
1707 \\ are recognized, with their usual interpretation.
1713 Curly braces {} may be used to enclose ranges of single
1714 characters (possibly using the escape convention described in the
1715 preceding point), eg. {a-z} to introduce the
1716 standard range of ASCII characters.
1717 Note that the interpretation of such a range depends on
1718 the concrete representation in your local, physical character set.
1724 paranthesises () may be used to enclose multi-byte characters -
1725 eg. diacritics or special national combinations (eg. Spanish
1726 "ll"). When found in the input stream (or a search term),
1727 these characters are viewed and sorted as a single character, with a
1728 sorting value depending on the position of the group in the value
1736 </listitem></varlistentry>
1738 <term>uppercase <emphasis>value-set</emphasis></term>
1741 This directive introduces the
1742 upper-case equivalencis to the value set (if any). The number and
1743 order of the entries in the list should be the same as in the
1744 <literal>lowercase</literal> directive.
1746 </listitem></varlistentry>
1748 <term>space <emphasis>value-set</emphasis></term>
1751 This directive introduces the character
1752 which separate words in the input stream. Depending on the
1753 completeness mode of the field in question, these characters either
1754 terminate an index entry, or delimit individual "words" in
1755 the input stream. The order of the elements is not significant —
1756 otherwise the representation is the same as for the
1757 <literal>uppercase</literal> and <literal>lowercase</literal>
1760 </listitem></varlistentry>
1762 <term>map <emphasis>value-set</emphasis>
1763 <emphasis>target</emphasis></term>
1766 This directive introduces a
1767 mapping between each of the members of the value-set on the left to
1768 the character on the right. The character on the right must occur in
1769 the value set (the <literal>lowercase</literal> directive) of
1770 the character set, but
1771 it may be a paranthesis-enclosed multi-octet character. This directive
1772 may be used to map diacritics to their base characters, or to map
1773 HTML-style character-representations to their natural form, etc.
1775 </listitem></varlistentry>
1783 <sect1 id="formats">
1784 <title>Exchange Formats</title>
1787 Converting records from the internal structure to en exchange format
1788 is largely an automatic process. Currently, the following exchange
1789 formats are supported:
1796 GRS-1. The internal representation is based on GRS-1/XML, so the
1797 conversion here is straightforward. The system will create
1798 applied variant and supported variant lists as required, if a record
1799 contains variant information.
1805 XML. The internal representation is based on GRS-1/XML so
1806 the mapping is trivial. Note that XML schemas, preprocessing
1807 instructions and comments are not part of the internal representation
1808 and therefore will never be part of a generated XML record.
1809 Future versions of the Zebra will support that.
1815 SUTRS. Again, the mapping is fairly straightforward. Indentation
1816 is used to show the hierarchical structure of the record. All
1817 "GRS" type records support both the GRS-1 and SUTRS
1819 FIXME - What is SUTRS - should be expanded here
1825 ISO2709-based formats (USMARC, etc.). Only records with a
1826 two-level structure (corresponding to fields and subfields) can be
1827 directly mapped to ISO2709. For records with a different structuring
1828 (eg., GILS), the representation in a structure like USMARC involves a
1829 schema-mapping (see <xref linkend="schema-mapping"/>), to an
1830 "implied" USMARC schema (implied,
1831 because there is no formal schema which specifies the use of the
1832 USMARC fields outside of ISO2709). The resultant, two-level record is
1833 then mapped directly from the internal representation to ISO2709. See
1834 the GILS schema definition files for a detailed example of this
1841 Explain. This representation is only available for records
1842 belonging to the Explain schema.
1848 Summary. This ASN-1 based structure is only available for records
1849 belonging to the Summary schema - or schema which provide a mapping
1850 to this schema (see the description of the schema mapping facility
1857 SOIF. Support for this syntax is experimental, and is currently
1858 keyed to a private Index Data OID (1.2.840.10003.5.1000.81.2). All
1859 abstract syntaxes can be mapped to the SOIF format, although nested
1860 elements are represented by concatenation of the tag names at each
1862 FIXME - Is this used anywhere ? What is SOIF anyway? -H
1871 <!-- Keep this comment at the end of the file
1876 sgml-minimize-attributes:nil
1877 sgml-always-quote-attributes:t
1880 sgml-parent-document: "zebra.xml"
1881 sgml-local-catalogs: nil
1882 sgml-namecase-general:t