CDE/cde/doc/C/guides/man/man4/dtsrhanf.sgm

<!-- $XConsortium: dtsrhanf.sgm /main/6 1996/09/08 20:19:48 rws $ -->
<!-- (c) Copyright 1996 Digital Equipment Corporation. -->
<!-- (c) Copyright 1996 Hewlett-Packard Company. -->
<!-- (c) Copyright 1996 International Business Machines Corp. -->
<!-- (c) Copyright 1996 Sun Microsystems, Inc. -->
<!-- (c) Copyright 1996 Novell, Inc. -->
<!-- (c) Copyright 1996 FUJITSU LIMITED. -->
<!-- (c) Copyright 1996 Hitachi. -->
<![ %CDE.C.CDE; [<RefEntry Id="CDE.INFO.dtsrhanfile">]]>
<RefMeta>
<RefEntryTitle>dtsrhanfile</RefEntryTitle>
<ManVolNum>special file</ManVolNum>
</RefMeta>
<RefNameDiv>
<RefName>dtsrhanfile</RefName>
<RefPurpose>
Describes the format and syntax of DtSearch han files
</RefPurpose>
</RefNameDiv>
<RefSynopsisDiv>
<Synopsis>
<Symbol Role="Variable">filename</Symbol>.han
</Synopsis>
</RefSynopsisDiv>
<RefSect1>
<Title>DESCRIPTION</Title>
<Para>Han files are the user generated profile files for <command>dtsrhan</command>.
They identify fields in incoming text from which output fzk
file fields can be constructed. The data from han files
are loaded into memory by dtsrhan at initialization time.
<command>dtsrhan</command> and han files have not been internationalized;
han files may only contain ASCII characters.
</para>
<refsect2>
<Title>General Format</Title>
<para>All identifiers must begin with a letter, and must be composed entirely
of alphanumerics and/or the underscore.
</para>
<para>Observe the following points when using using "strings":
</para>
<itemizedlist>
<listitem>
<para>If an identifying string contains quotes, use a backslash
to create the quote. Example:
</para>
<programlisting>
this string \"contains\" quotes
</programlisting>
<para>would find the string <literal>this string "contains" quotes</literal>.
</para>
</listitem>
<listitem>
<para>The above point makes it necessary to use double backslashes to create
a single backslash. Example:
</para>
<programlisting>
this string has a \\ backslash
</programlisting>
<para>would find the string <literal>this string has a \ backslash</literal>.
</para>
</listitem>
<listitem>
<para>Actually, using the backslash in any string will cause the next
character to be included without exception. Thus, a string
with <literal>this is \a test</literal> will end up being
<literal>this is a test</literal>.
The backslash is ignored, and the next character is imbedded
in the string. This is only needed in the two cases described
above, but can be used for any purpose.
</para>
</listitem>
</itemizedlist>
</refsect2>
<refsect2>
<Title>Individual Line Syntax</Title>
<variablelist>
<varlistentry><term># ... | blank line</term>
<listitem>
<para>Han file comment. Any line beginning with a pound sign
in the first column, or any blank line, is discarded.
</para>
</listitem>
</varlistentry>
<varlistentry><term>line <emphasis>identifier</emphasis> = <emphasis>physical_line_number</emphasis></term>
<listitem>
<para>Defines a <literal>line</literal> with a physical line number in the record.
<emphasis>physical_line_number</emphasis> must be a number.
</para>
</listitem>
</varlistentry>
<varlistentry><term>line <emphasis>identifier</emphasis> = column_number,"<emphasis>string</emphasis>" [<emphasis>column_number</emphasis>,"<emphasis>string</emphasis>"] ...</term>
<listitem>
<para>Defines a <literal>line</literal> using a column number and a
'signature' string that should appear at that column.
<emphasis>column_number</emphasis> can be a number, or
<literal>*</literal> for 'any column'. "<emphasis>string</emphasis>"
should be a string that occurs on the line in question. It is possible
to define complex signatures using multiple clauses.
</para>
</listitem>
</varlistentry>
<varlistentry><term>field <emphasis>identifier</emphasis> = <emphasis>line_identifier</emphasis>,"<emphasis>string</emphasis>", <emphasis>offset</emphasis>, <emphasis>length</emphasis></term>
<listitem>
<para>Defines a <literal>field</literal> based on a declared line, a string
found on that line, the offset from the first letter of the string, and
the length of field.
</para>
<para><emphasis>line_identifier</emphasis> is an identifier declared with the
<literal>line</literal> directive (see above).
</para>
<para>"<emphasis>string</emphasis>" is a string for relative positioning,
where a field will follow a string that may not always occur in the
same position on a line. If it is known that the field will always be
in the same position, an empty string("") may be used.
<emphasis>string</emphasis> must be enclosed in double quotes.
<emphasis>offset</emphasis> must be a number, identifying the offset
from the first character in the string. It starts at position 1, not 0,
and may be negative.
</para>
<para><emphasis>length</emphasis> represents the length of the field. It may
be a number, or it may be one of two special tokens:
</para>
<variablelist>
<varlistentry><term><literal>eow</literal></term>
<listitem>
<para>End of word. The field will begin at <emphasis>offset</emphasis> and
continue until the next white-space character.
</para>
</listitem>
</varlistentry>
<varlistentry><term><literal>eoln</literal></term>
<listitem>
<para>End of line. The field will begin at <emphasis>offset</emphasis> and
continue to the end of the line.
</para>
</listitem>
</varlistentry>
</variablelist>
<para>An identifier <emphasis>string</emphasis> beginning with 3 uppercase M's
("MMM...") will be considered an English month name string.
At run time, if the first 3 chars of the field's value
equal the first three chars of an English month name,
the value string will be translated to a two character
string of digits in the range "01" to "12".
For example, if field <emphasis>MMMmymonth</emphasis> had an original value of
"April ", it will be translated to "04" before use.
</para>
<para>In the case where a <literal>line</literal> identifier is associated with
multiple lines in a single document, the field value will
be determined from the last occurrence of the line within
the record.
</para>
</listitem>
</varlistentry>
<varlistentry><term>constant <emphasis>identifier</emphasis> = "<emphasis>string</emphasis>"</term>
<listitem>
<para>Defines a <literal>constant</literal> field that can be used in
abstracts and keys. The <emphasis>identifier</emphasis> is defined
exactly the same as a <literal>field</literal> identifier. The value
must be enclosed in double quotes.
</para>
</listitem>
</varlistentry>
<varlistentry><term>date = null | <emphasis>field_id</emphasis> [+ <emphasis>field_id</emphasis>] ...</term>
<listitem>
<para>Defines the document date for each document. It will
be converted into a correctly formatted fzk file date line.
</para>
<para><literal>null</literal> specifies undated documents. Undated documents
always qualify for searches irrespective of date
qualifiers in <function>DtSearchQuery</function>.
</para>
<para><emphasis>field_id</emphasis> is an identifier declared using the <literal>field</literal>
or <literal>constant</literal> directives (see above).
"MMM" fields are often useful for date assemblies.
</para>
<para>Multiple fields may be concatenated into a date.
</para>
<para>After concatenation, the assembled date must be of the following format:
<emphasis>YYYYMMDDhhmm</emphasis> (exactly 12 digits). For example,
<literal>199404171701</literal> is April 17, 1994 at 5:01 pm.
<literal>200405031000</literal> is May 3, 2004, at 10:00 am (10
o'oclock).
</para>
<para>Dates before 1900 or after 5995 are invalid.
</para>
<para>If <literal>date</literal> is not specified or is invalid, a generated date
based on the current date and time will be used, but an
invalid <literal>date</literal> will also generate an error message.
</para>
</listitem>
</varlistentry>
<varlistentry><term>key = <emphasis>field_id</emphasis> [+ <emphasis>field_id</emphasis>] ... | time | count</term>
<listitem>
<para>Defines the unique database key for each record in a fzk file.
</para>
<para><emphasis>field_id</emphasis> is a field identifier declared using the
<literal>field</literal> or <literal>constant</literal> directives.
</para>
<para>Multiple fields may be concatenated into a key.
</para>
<para><literal>time</literal> is a special keyword used to generate keys based
on the current run date and time, plus a sequential count suffix.
</para>
<para><literal>count</literal> is a special keyword used to generate keys
based on a sequential count of records.
</para>
</listitem>
</varlistentry>
<varlistentry><term>upper</term>
<listitem>
<para>Specifies that keys written by handel are to be entirely converted
to upper case. Without using this directive, mixed-case keys
are allowed.
</para>
</listitem>
</varlistentry>
<varlistentry><term>keychar = A | B | ...Z</term>
<listitem>
<para>Defines the character used to categorize keys for DtSearch. It
must be an uppercase ASCII alphabetic character.
</para>
</listitem>
</varlistentry>
<varlistentry><term>delimiter = <emphasis>line_identifier</emphasis>, bottom</term>
<listitem>
<para>Defines the end of text (ETX) delimiter that will separate records.
</para>
<para><emphasis>line_identifier</emphasis> is an identifier declared with the
<literal>line</literal> directive.
</para>
<para><literal>bottom</literal> is required. It specifies that the ETX will
occur at the bottom of each record. Top of record delimiters are not
supported.
</para>
</listitem>
</varlistentry>
<varlistentry><term>image = all | none</term>
<listitem>
<para>Defines whether the document image retrieved by
<function>DtSearchRetrieve</function> is to contain all or none of the
record, prior to application of <literal>imageinclude</literal> or
<literal>imageexclude</literal> directives later in the han file. It
defaults to <literal>all</literal>.
</para>
</listitem>
</varlistentry>
<varlistentry><term>imageinclude = <emphasis>line_identifier</emphasis> [- <emphasis>line_identifier</emphasis>]</term>
<listitem>
<para>Defines a line (or range of lines) to be included in the image.
<emphasis>line_identifier</emphasis> is an identifier declared with the
<literal>line</literal> directive.
</para>
</listitem>
</varlistentry>
<varlistentry><term>imageexclude = <emphasis>line_identifier</emphasis> [- <emphasis>line_identifier</emphasis>]</term>
<listitem>
<para>Defines a line (or range of lines) to be excluded from the image.
<emphasis>line_identifier</emphasis> is an identifier declared with the
<literal>line</literal> directive.
</para>
</listitem>
</varlistentry>
<varlistentry><term>abstract = field(s) <emphasis>field_identifier</emphasis> [+ <emphasis>field_identifier</emphasis>]...</term>
<listitem>
<para>Defines the abstract to be placed into the fzk file. It is created from
the concatenations of fields. <emphasis>field_identifier</emphasis> is
an identifier declared with the <literal>field</literal> directive.
</para>
</listitem>
</varlistentry>
<varlistentry><term>delblanklines = true | false</term>
<listitem>
<para>Determines if blank lines are to be removed from the record image or
not. It defaults to <literal>false</literal>.
</para>
</listitem>
</varlistentry>
</variablelist>
</refsect2>
<refsect2>
<Title>Example</Title>
<para>The sample han file shown here describes a text file containing a
concatenated set of man pages documents.
</para>
<programlisting>
# All records in the incoming text file are delimited by the same
# end of text convention as the default for an fzk file, namely
# a linefeed (control-L) on a line by itself ("\f\n").
# Define a line named "etx" with that description,
# and declare it to be the &lt;delimiter>.
# Note that there must be a real ASCII control-L character between
# the quotes in the line below.
line etx = *,"^L"
delimiter = etx, bottom

# The command name that the man page is describing is on the first line.
# To access it we need to define a line directive for line number 1.
line line1 = 1

# The name of the man page command begins in column 3 of line 1,
# and the length is variable.  So we define a field identifier
# named "command1" from column 3 to the end of the word.
field command1 = line1,"",3,eow

# We want each document abstract to have a constant prefix
# followed by the name of the command.
constant preabs = "Man Pages for "
abstract = fields preabs + command1

# We want all keys to be the name of the command, prefixed with
# the same identifying character, an uppercase M.
keychar = M
key = command1

# We want the each document date to be equivalent to the release
# date of the original man pages, which we choose here to hard code
# as November 1, 1994, at 1 o'clock in the afternoon.
constant datecons = "199411011300"
date = datecons
</programlisting>
</refsect2>
</refsect1>
<RefSect1>
<Title>SEE ALSO</Title>
<Para>&cdeman.dtsrhan;,
&cdeman.dtsrindex;,
&cdeman.dtsrfzkfiles;,
&cdeman.dtsrlangfiles;,
&cdeman.DtSearch;
</Para>
</RefSect1>
</RefEntry>