Documentation Source Text

  of the full-text search engine is added to the build automatically.
}

COMPILE_OPTION {SQLITE_ENABLE_FTS5} {
  When this option is defined in the [amalgamation], versions 5
  of the full-text search engine ([fts5]) is added to the build automatically.
}

COMPILE_OPTION {SQLITE_ENABLE_GEOPOLY} {
  When this option is defined in the [amalgamation], the [Geopoly extension]
  is included in the build.
}

COMPILE_OPTION {SQLITE_ENABLE_ICU} {
  This option causes the 
  [http://www.icu-project.org/ | International Components for Unicode]
  or "ICU" extension to SQLite to be added to the build.  
}

<title>The Geopoly Interface To The SQLite R*Tree Module</title>
<tcl>hd_keywords {geopoly} {Geopoly module} {Geopoly extension}</tcl>
<table_of_contents>

<h1>Overview</h1>

<p>
The Geopoly module is an alternative interface to the [R-Tree extension] that uses
the [http://geojson.org | GeoJSON] notation
([https://tools.ietf.org/html/rfc7946 | RFC-7946]) to describe two-dimensional
polygons.  Geopoly includes functions for detecting when one polygon is
contained within or overlaps with another, for computing the
area enclosed by a polygon, for doing linear transformations of polygons,
for rendering polygons as
[https://en.wikipedia.org/wiki/Scalable_Vector_Graphics | SVG], and other
similar operations.

<p>
The source code for Geopoly is included in the [amalgamation] but is not
included in the library unless the [-DSQLITE_ENABLE_GEOPOLY] compile-time option
is used.

<p>
Geopoly operates on "simple" polygons - that is, polygons for which
the boundary does not intersect itself.  Geopoly thus extends the capabilities
of the [R-Tree extension] which can only deal with rectangular areas.
On the other hand, the [R-Tree extension] is
able to handle between 1 and 5 coordinate dimensions, whereas Geopoly is restricted
to 2-dimensional shapes only.

with the origin at the lower left, whereas SVG uses a left-handed coordinate
system with the origin at the upper left.  The geopoly_svg() routine makes no
attempt to transform the coordinate system, so the displayed images are shown
in mirror image and rotated.  If that is undesirable, the geopoly_xform() routine
can be used to transform the output from cartesian to SVG coordinates prior to
passing the polygons into geopoly_svg().

<tcl>hd_fragment gbbox geopoly_bbox geopoly_group_bbox</tcl>
<h2>The geopoly_bbox(P) and geopoly_group_bbox(P) Functions</h2>

<p>
If P is a polygon, then geopoly_bbox(P) returns a new polygon that is
the smallest (axis-aligned) rectangle completely enclosing P.
If P is not a polygon, geopoly_bbox(P) returns NULL.

<p>
The geopoly_group_bbox(P) function is an aggregate version of geopoly_bbox(P).
The geopoly_group_bbox(P) function returns the smallest rectangle that will
enclose all P values seen during aggregation.

<tcl>hd_fragment gpoint geopoly_constains_point</tcl>
<h2>The geopoly_contains_point(P,X,Y) Function</h2>

<p>
If P is a polygon, then geopoly_contains_point(P,X,Y) returns true if and only
if the coordinate X,Y is inside or on the boundary of the polygon P.
If P is not a polygon, geopoly_contains_point(P,X,Y) returns NULL.

<tcl>hd_fragment xform geopoly_xform</tcl>
<h2>The geopoly_xform(P,A,B,C,D,E,F) Function</h2>

<p>
The geopoly_xform(P,A,B,C,D,E,F) function returns a new polygon that is an
affine transformation of the polygon P and where the transformation
is defined by values A,B,C,D,E,F. If P is not a valid polygon, this
routine returns NULL.

<p>
The transformation converts each vertex of the polygon according to the
following formula:

to compute polygon decoding, overlap, and containment.

<h2>Binary Encoding of Polygons</h2>

<p>
Geopoly stores all polygons internally using a binary format.  A binary
polygon consists of a 4-byte header following by an array of coordinate
pairs in which each dimension of each coordinate is a 32-bit floating point
number.

<p>
The first byte of the header is a flag byte.  The least significant bit
of the flag byte determines whether the coordinate pairs that follow the
header are stored big-endian or little-endian.  A value of 0 for the least
significant bit means big-endian and a value of 1 means little endian.