MFSMETA

Section: User Commands (1)
Updated: November 2000
Index Return to Main Contents

NAME

mfsmeta - XPAT MFS structure file builder

SYNOPSIS

mfsmeta [ -v ] [ -o " meta_structure_file" ] -D data_dictionary

DESCRIPTION

mfsmeta builds an MFS structure file for the MFS database specified by data_dictionary. The MFS structure file is used by multirgn and sgmlrgn to restrict the region build operations to specific areas of the MFS database's virtual text. The generated meta_structure_file name will be automatically appended ('.str') extension. Refer to the multirgn(1) and sgmlrgn(1) man pages for further details on how those programs use an MFS structure file.

OPTIONS

-v: Turn verbose mode on.
-o " meta_structure_file": Write the MFS structure information to the meta_structure_file. By default, mfsmeta uses the data dictionary name and append ('.str') extension at the end as the meta_structure_file name.

BACKGROUND

The MFS system creates a ``virtual text'' from all the files in the database. It does so by performing the following three steps on each file. First, it filters out the actual text from each file. Note that the files may be in several different word processor formats, so different filters may be needed for each different file type. Second, it wraps special tagged `meta fields' around each of those filtered texts. Third, it catenates all the `wrapped texts' together to form a single ``virtual text''. The index building programs then build all the indices on that virtual text. Refer to the mfs(5) man page for further details on how the MFS system works.

The portion of the virtual text that corresponds to each file in the database consists of three basic sections: The Meta-Header section, the Data section and the Meta-Trailer section. These sections are shown in the following diagram:


  <OTDoc><OTMeta>....</OTMeta><OTData>........</OTData></OTDoc>
  |--------- Meta-Header ------------|| Data ||- Meta-Trailer -|
  ^                                   ^       ^                ^
  start                               start   start            end
  header                              data    trailer          pos

The data in the Meta-Header and Meta-Trailer sections is highly structured and is uniform across all the files in the database. In contrast, the data in the Data sections may have widely varying structure (if any exists at all), especially in databases that consists of several different file types. For this reason, it is sometimes desirable to build the regions for the entire database in several passes, restricting the build operations for each pass to either the Meta-Header and Meta-Trailer sections or the Data sections of specific file types.

Note: The end position is usually one character after the </OTDoc> tag. This is because the virtual text for each file ends with a newline, which is considered to be the last character of the file.

MFS STRUCTURE FILE FORMAT


  <OTMetaStruct>
    <OTFile>            file name
    <OTDefaultDataTag>  default data tag name (default is 'OTData')
    <OTDisplayFmt>      display format
    <OTStart>           start header position
    <OTDataPos>         start data position
    <OTTrailPos>         start trailer position
    <OTEnd>             end position for this file
  </OTMetaStruct>

INTEGRITY CHECK

The first line of the MFS structure file is the MFS database name, enclosed by <FileMap> and </FileMap> tags. This field contains the same name as the FileMap field in the Data Dictionary. All region builders will check for the MFS database name before trying to use the meta structure file for the given MFS database.

This document was created by man2html, using the manual pages.
Time: 18:03:38 GMT, March 26, 2001