Geografic-Information-System/lib/vector/vectorlib_files.dox

/*! \page vlibFormat Vector format description

by GRASS Development Team (http://grass.osgeo.org)

\par Table of contents

- \ref vlibDirectoryStructure
- \ref vlibDblnFileFormat
- \ref vlibCoorFileFormat
 - \ref vlibCoorFileHead
 - \ref vlibCoorFileBody

\section vlibDirectoryStructure Directory structure

Vector map is stored in a number of data files. Vector map directory
structure and file names were changed in GRASS 6 with respect to
previous GRASS versions. All vector files for one vector map are
stored in one directory:

\verbatim
$MAPSET/vector/vector_name/
\endverbatim

This directory contains these files:

- <b>coor</b> - binary file, coordinates [former dig/ file] (see \ref vlibCoorFileFormat)
- <b>topo</b> - binary file, topology [former dig_plus/ file] (see \ref vlibTopoFileFormat)
- <b>sidx</b> - binary file, spatial index (see \ref vlibSidxFileFormat)
- <b>cidx</b> - binary file, category index (see \ref vlibCidxFileFormat)
- <b>head</b> - text file, header information [former part of dig/ file] (see \ref vlibHeadFileFormat)
- <b>dbln</b> - text file, link(s) to attribute table(s) (see \ref vlibDblnFileFormat)
- <b>hist</b> - text file, vector map change history
- <b>frmt</b> - text file, format description (external formats only)
- <b>fidx</b> - binary file, feature index (OGR format only)

\section vlibHeadFileFormat Header file format specification

The header contains meta information, a description of the
vector map and many other information. The file is an unordered list
of key/value entries. The <i>key</i> is a string separated from
<i>value</i> by a colon and optional whitespace.

Keywords are:

- ORGANIZATION - organization that digitized the data
- DIGIT DATE - date the data was digitized
- DIGIT NAME - person who digitized the data
- MAP NAME - title of the original source map
- MAP DATE - date of the original source map
- MAP SCALE - scale of the original source map
- OTHER INFO - other comments about the map
- ZONE - zone of the map (e.g., UTM zone)
- MAP THRESH - digitizing threshold

This information holds \ref dig_head data structure.


\section vlibDblnFileFormat DB link file format specification

Each vector maps has its own DBMI settings stored in the
'$MAPSET/vector/vector_name/dbln' text file. For each pair <em>vector map +
layer</em>, all of <em>table, key column, database, driver</em> must be
defined in a new row. This definition must be written to
'$MAPSET/vector/vector_name/dbln' text file. Each row in the 'dbln'
file contains names separated by spaces in following order ([ ] -
optional):

\verbatim
map[@mapset] layer table [key [database [driver]]]
\endverbatim

If key, database or driver are omitted (on second and higher row only)
the last definition is used. When reading a vector map from another
mapset (if mapset is specified along with map name), definitions in
the related "dbln" file may overwrite the DBMI definition in the
current mapset. This means that the map-wise definition is always
"stronger".

Wild cards <b>*</b> and <b>?</b> may be used in map and mapset names.

Variables $GISDBASE, $LOCATION_NAME, $MAPSET, and $MAP may be used in
table, key, database and driver names (function
Vect_subst_var()). Note that $MAPSET is not the current mapset but
mapset of the map the rule is defined for.

Note that vector features in GRASS vector maps may have attributes in
different tables or may be without attributes. Boundaries form areas
but it may happen that some boundaries are not closed (such boundaries
would not appear in polygon layer).  Boundaries may have
attributes. All types may be mixed in one vector map.

The link to the table is permanent and it is stored in 'dbln' file in
vector directory. Tables are considered to be a part of the vector and
the command <tt>g.remove</tt>, for example, deletes linked tables of
the vector. Attributes must be joined with geometry.

Information about database links holds \ref dblinks data structure.

<b>Examples:</b>

Examples are written mostly for the DBF driver, where database is full
path to the directory with dbf files and table name is the name of dbf
file without .dbf extension:

\verbatim
* 1 mytable id $GISDBASE/$LOCATION_NAME/$MAPSET/vector/$MAP dbf
\endverbatim

This definition says that entities with category of layer 1 are linked
to dbf tables with names "mytable.dbf" saved in vector directories of
each map. The attribute column containing the category numbers is
called "id".

\verbatim
* 1 $MAP id $GISDBASE/$LOCATION_NAME/$MAPSET/dbf dbf
\endverbatim 

Similar as above but all dbf files are in one directory dbf/ in mapset
and names of dbf files are $MAP.dbf

\verbatim
water* 1 rivers id /home/grass/dbf dbf
water* 2 lakes lakeid /home/guser/mydb
trans* 1 roads key basedb odbc
trans* 5 rails
\endverbatim

These definitions define more layers (called "field" in the API) for
one vector map i.e. in one vector map may be more features linked to
more attribute tables. Definitions on first 2 rows are applied for
example on maps water1, water2, ... so that more maps may share one
table.

\verbatim
water@PERMANENT 1 myrivers id /home/guser/mydbf dbf
\endverbatim

This definion overwrites the definition saved in PERMANENT/VAR and
links the water map from PERMANENT mapset to the user's table.

Modules should be written so that connections to databases for each
vector layer are independent. It should be possible to read attributes
of an input vector map from one database and write to some other and
even with some other driver (should not be a problem).

There are open questions, however. For one, how does one distinguish when
new tables should be written and when not? For example, definitions:

\verbatim
river 1 river id water odbc
river.backup* 1 NONE
\endverbatim

could be used to say that tables should not be copied for backups of
map river because table is stored in a reliable RDBMS.

\section vlibs Vector libraries

Besides internal library functions there are two main libraries:

- Vlib (Vector library), see \ref vlibIntro
- DGLib (Directed Graph Library), see \ref vlibDglib

For historical reasons, there are two internal libraries:

- diglib (with dig_*() functions), GRASS 3.x/4.x
- Vlib (with V1_*(), V2_*() and Vect_*() functions), since GRASS 4.x
  (except for the 5.7 interim version)

The vector library was introduced in GRASS 4.0 to hide internal vector
files' formats and structures.  In GRASS 6/7, everything is accessed via
Vect_*() functions, for example:

Old 4.x code:

\code
    xx = Map.Att[Map.Area[area_num].att].x;
\endcode

New 6.x/7.x functions:

\code
    centroid = Vect_get_area_centroid(Map, area_num);
    Vect_read_line(Map, line_p, NULL, centroid);
    Vect_line_get_point(line_p, 0, &xx, NULL, NULL);
\endcode

In GRASS 6/7, all internal, mostly non-topological vector functions are
hidden from the modules' API (mainly dig_*(), V1_*() and V2_*()
functions). All available Vect_*() functions are topological vector
functions.


The following include file contains definitions and structures
required by some of the routines in this library. The programmer
should therefore include this file in any code that uses the vector
library:

\code
#include <grass/vector.h>
\endcode

<i>Note: For details please read Blazek et al. 2002 (see below) as
well as the references in this document.</i>

\subsection vlibHistory Historical notes

The vector library in GRASS 4.0 changed significantly from the
<em>Digit Library</em> (diglib) used in GRASS 3.1. Below is an
overview of why the changes were made.

The Digit Library was a collage of subroutines created for developing
the map development programs. Few of these subroutines were actually
designed as a user access library. They required individuals to assume
too much responsibility and control over what happened to the data
file. Thus when it came time to change vector data file formats for
GRASS 4.0, many modules also required modification. The two different
access levels for 3.0 vector files provided very different ways of
calling the library; they offered little consistency for the user.

The Digit Library was originally designed to only have one file open
for read or write at a time. Although it was possible in some cases to
get around this, one restriction was the global head structure. Since
there was only one instance of this, there could only be one copy of
that information, and thus, only one open vector file.

The solution to these problems was to design a new user library as an
interface to the vector data files. This new library was designed to
provide a simple consistent interface, which hides as much of the
details of the data format as possible. It also could be extended for
future enhancements without the need to change existing programs.

The new vector library in GRASS 4 provided routines for opening,
closing, reading, and writing vector files, as well as several support
functions. The Digit Library has been replaced, so that all existing
modules was converted to use the new library. Those routines that
existed in the Digit Library and were not affected by these changes
continue to exist in unmodified form, and were included in the vector
library. Most of the commonly used routines have been discarded, and
replaced by the new vector routines.

Instead the global head structure was used own local version of
it. The structure that replaced structure head is structure \ref
dig_head. There were still two levels of interface to the vector files
(future releases may include more). Level one provided access only to
arc (i.e. polyline) information and to the type of line (AREA, LINE,
DOT). Level two provided access to polygons (areas), attributes, and
network topology.


\subsection vlibCoorFileFormat Coor file format specification

In the coor file the following is stored: 'line' (element) type,
number of attributes and layer number for each category. Coordinates
in binary file are stored as double (8 bytes). See \ref Coor_info data
structure.

\subsubsection vlibCoorFileHead Header

<table border="1" style="border-collapse: collapse" cellpadding="5">
<tr><td><b>Name</b></td><td><b>Type</b></td><td><b>Number</b></td><td><b>Description</b></td></tr>
<tr><td>Version_Major</td> <td>C</td> <td>1</td> <td>file version (major)</td></tr>
<tr><td>Version_Minor</td> <td>C</td> <td>1</td> <td>file version (minor)</td></tr>
<tr><td>Back_Major</td>    <td>C</td> <td>1</td> <td>supported from GRASS version (major)</td></tr>
<tr><td>Back_Minor</td>    <td>C</td> <td>1</td> <td>supported from GRASS version (minor)</td></tr>
<tr><td>byte_order</td>    <td>C</td> <td>1</td> <td>little or big endian flag</td></tr>
<tr><td>head_size</td>     <td>L</td> <td>1</td> <td>header size of coor file</td></tr>
<tr><td>with_z</td>        <td>C</td> <td>1</td> <td>2D or 3D flag; zero for 2D</td></tr>
<tr><td>size</td>          <td>L</td> <td>1</td> <td>coor file size</td></tr>
</table>

\section vlibCoorFileBody Body

The body consists of line records:

<table border="1" style="border-collapse: collapse" cellpadding="5">
<tr><td><b>Name</b></td><td><b>Type</b></td><td><b>Number</b></td><td><b>Description</b></td></tr>
<tr><td>record header</td><td>C</td><td>1</td><td>
 - 0. bit: 1 - alive, 0 - dead line
 - 1. bit: 1 - categories, 0 - no categories
 - 2.-3. bit: type - one of: GV_POINT, GV_LINE, GV_BOUNDARY, GV_CENTROID, GV_FACE, GV_KERNEL
 - 4.-7. bit: reserved, not used 
</td></tr>

<tr><td>ncats</td><td>I</td><td>1</td><td>number of categories 
    (written only if categories exist) </td></tr>

<tr><td>field</td><td>I</td><td>ncats</td><td>field identifier,
    distinguishes between more categories append to one feature (written
    only if categories exist; field is called "layer" at user
    level)</td></tr>

<tr><td>cat</td><td>I</td><td>ncats</td><td>category value (written
    only if categories exist)</td></tr>

<tr><td>ncoor</td><td>I</td><td>1</td><td>written for GV_LINES and GV_BOUNDARIES 
    only</td></tr>

<tr><td>x</td><td>D</td><td>ncoor</td><td>x coordinate</td></tr>

<tr><td>y</td><td>D</td><td>ncoor</td><td>y coordinate</td></tr>

<tr><td>z</td><td>D</td><td>ncoor</td><td>z coordinate; present if
    with_z in head is set to 1</td></tr> </table>

Types used in coor file:

<table border="1" style="border-collapse: collapse" cellpadding="5">
<tr><td><b>Type</b></td><td><b>Name</b></td><td><b>Size in Bytes</b></td></tr>
<tr><td>D</td><td>Double</td><td>8</td></tr>
<tr><td>L</td><td>Long  </td><td>4</td></tr>
<tr><td>I</td><td>Int   </td><td>4</td></tr>
<tr><td>S</td><td>Short </td><td>4</td></tr>
<tr><td>C</td><td>Char  </td><td>1</td></tr>
</table>



*/