mirrors/django
1
0
Fork 0
mirror of https://github.com/django/django.git synced 2024-12-29 12:36:08 +00:00
Code Issues 30 Releases Wiki Activity
9ec1ff7879
django/docs/ref/contrib/gis/gdal.txt

2028 lines
66 KiB
Plaintext
Raw Blame History

========
GDAL API
========
.. module:: django.contrib.gis.gdal
:synopsis: GeoDjango's high-level interface to the GDAL library.
`GDAL`__ stands for **Geospatial Data Abstraction Library**,
and is a veritable "Swiss army knife" of GIS data functionality. A subset
of GDAL is the `OGR`__ Simple Features Library, which specializes
in reading and writing vector geographic data in a variety of standard
formats.
GeoDjango provides a high-level Python interface for some of the
capabilities of OGR, including the reading and coordinate transformation
of vector spatial data and minimal support for GDAL's features with respect
to raster (image) data.
.. note::
Although the module is named ``gdal``, GeoDjango only supports some of the
capabilities of OGR and GDAL's raster features at this time.
__ https://gdal.org/
__ https://gdal.org/user/vector_data_model.html
Overview
========
.. _gdal_sample_data:
Sample Data
-----------
The GDAL/OGR tools described here are designed to help you read in
your geospatial data, in order for most of them to be useful you have
to have some data to work with. If you're starting out and don't yet
have any data of your own to use, GeoDjango tests contain a number of
data sets that you can use for testing. You can download them here:
.. code-block:: shell
$ wget https://raw.githubusercontent.com/django/django/main/tests/gis_tests/data/cities/cities.{shp,prj,shx,dbf}
$ wget https://raw.githubusercontent.com/django/django/main/tests/gis_tests/data/rasters/raster.tif
Vector Data Source Objects
==========================
``DataSource``
--------------
:class:`DataSource` is a wrapper for the OGR data source object that
supports reading data from a variety of OGR-supported geospatial file
formats and data sources using a consistent interface. Each
data source is represented by a :class:`DataSource` object which contains
one or more layers of data. Each layer, represented by a :class:`Layer`
object, contains some number of geographic features (:class:`Feature`),
information about the type of features contained in that layer (e.g.
points, polygons, etc.), as well as the names and types of any
additional fields (:class:`Field`) of data that may be associated with
each feature in that layer.
.. class:: DataSource(ds_input, encoding='utf-8')
The constructor for ``DataSource`` only requires one parameter: the path of
the file you want to read. However, OGR also supports a variety of more
complex data sources, including databases, that may be accessed by passing
a special name string instead of a path. For more information, see the
`OGR Vector Formats`__ documentation. The :attr:`name` property of a
``DataSource`` instance gives the OGR name of the underlying data source
that it is using.
The optional ``encoding`` parameter allows you to specify a non-standard
encoding of the strings in the source. This is typically useful when you
obtain ``DjangoUnicodeDecodeError`` exceptions while reading field values.
Once you've created your ``DataSource``, you can find out how many layers
of data it contains by accessing the :attr:`layer_count` property, or
(equivalently) by using the ``len()`` function. For information on
accessing the layers of data themselves, see the next section:
.. code-block:: pycon
>>> from django.contrib.gis.gdal import DataSource
>>> ds = DataSource("/path/to/your/cities.shp")
>>> ds.name
'/path/to/your/cities.shp'
>>> ds.layer_count # This file only contains one layer
1
.. attribute:: layer_count
Returns the number of layers in the data source.
.. attribute:: name
Returns the name of the data source.
__ https://gdal.org/drivers/vector/
``Layer``
---------
.. class:: Layer
``Layer`` is a wrapper for a layer of data in a ``DataSource`` object. You
never create a ``Layer`` object directly. Instead, you retrieve them from
a :class:`DataSource` object, which is essentially a standard Python
container of ``Layer`` objects. For example, you can access a specific
layer by its index (e.g. ``ds[0]`` to access the first layer), or you can
iterate over all the layers in the container in a ``for`` loop. The
``Layer`` itself acts as a container for geometric features.
Typically, all the features in a given layer have the same geometry type.
The :attr:`geom_type` property of a layer is an :class:`OGRGeomType` that
identifies the feature type. We can use it to print out some basic
information about each layer in a :class:`DataSource`:
.. code-block:: pycon
>>> for layer in ds:
... print('Layer "%s": %i %ss' % (layer.name, len(layer), layer.geom_type.name))
...
Layer "cities": 3 Points
The example output is from the cities data source, loaded above, which
evidently contains one layer, called ``"cities"``, which contains three
point features. For simplicity, the examples below assume that you've
stored that layer in the variable ``layer``:
.. code-block:: pycon
>>> layer = ds[0]
.. attribute:: name
Returns the name of this layer in the data source.
.. code-block:: pycon
>>> layer.name
'cities'
.. attribute:: num_feat
Returns the number of features in the layer. Same as ``len(layer)``:
.. code-block:: pycon
>>> layer.num_feat
3
.. attribute:: geom_type
Returns the geometry type of the layer, as an :class:`OGRGeomType` object:
.. code-block:: pycon
>>> layer.geom_type.name
'Point'
.. attribute:: num_fields
Returns the number of fields in the layer, i.e the number of fields of
data associated with each feature in the layer:
.. code-block:: pycon
>>> layer.num_fields
4
.. attribute:: fields
Returns a list of the names of each of the fields in this layer:
.. code-block:: pycon
>>> layer.fields
['Name', 'Population', 'Density', 'Created']
.. attribute field_types
Returns a list of the data types of each of the fields in this layer. These
are subclasses of ``Field``, discussed below:
.. code-block:: pycon
>>> [ft.__name__ for ft in layer.field_types]
['OFTString', 'OFTReal', 'OFTReal', 'OFTDate']
.. attribute:: field_widths
Returns a list of the maximum field widths for each of the fields in this
layer:
.. code-block:: pycon
>>> layer.field_widths
[80, 11, 24, 10]
.. attribute:: field_precisions
Returns a list of the numeric precisions for each of the fields in this
layer. This is meaningless (and set to zero) for non-numeric fields:
.. code-block:: pycon
>>> layer.field_precisions
[0, 0, 15, 0]
.. attribute:: extent
Returns the spatial extent of this layer, as an :class:`Envelope` object:
.. code-block:: pycon
>>> layer.extent.tuple
(-104.609252, 29.763374, -95.23506, 38.971823)
.. attribute:: srs
Property that returns the :class:`SpatialReference` associated with this
layer:
.. code-block:: pycon
>>> print(layer.srs)
GEOGCS["GCS_WGS_1984",
DATUM["WGS_1984",
SPHEROID["WGS_1984",6378137,298.257223563]],
PRIMEM["Greenwich",0],
UNIT["Degree",0.017453292519943295]]
If the :class:`Layer` has no spatial reference information associated
with it, ``None`` is returned.
.. attribute:: spatial_filter
Property that may be used to retrieve or set a spatial filter for this
layer. A spatial filter can only be set with an :class:`OGRGeometry`
instance, a 4-tuple extent, or ``None``. When set with something other than
``None``, only features that intersect the filter will be returned when
iterating over the layer:
.. code-block:: pycon
>>> print(layer.spatial_filter)
None
>>> print(len(layer))
3
>>> [feat.get("Name") for feat in layer]
['Pueblo', 'Lawrence', 'Houston']
>>> ks_extent = (-102.051, 36.99, -94.59, 40.00) # Extent for state of Kansas
>>> layer.spatial_filter = ks_extent
>>> len(layer)
1
>>> [feat.get("Name") for feat in layer]
['Lawrence']
>>> layer.spatial_filter = None
>>> len(layer)
3
.. method:: get_fields()
A method that returns a list of the values of a given field for each
feature in the layer:
.. code-block:: pycon
>>> layer.get_fields("Name")
['Pueblo', 'Lawrence', 'Houston']
.. method:: get_geoms(geos=False)
A method that returns a list containing the geometry of each feature in the
layer. If the optional argument ``geos`` is set to ``True`` then the
geometries are converted to :class:`~django.contrib.gis.geos.GEOSGeometry`
objects. Otherwise, they are returned as :class:`OGRGeometry` objects:
.. code-block:: pycon
>>> [pt.tuple for pt in layer.get_geoms()]
[(-104.609252, 38.255001), (-95.23506, 38.971823), (-95.363151, 29.763374)]
.. method:: test_capability(capability)
Returns a boolean indicating whether this layer supports the given
capability (a string). Examples of valid capability strings include:
``'RandomRead'``, ``'SequentialWrite'``, ``'RandomWrite'``,
``'FastSpatialFilter'``, ``'FastFeatureCount'``, ``'FastGetExtent'``,
``'CreateField'``, ``'Transactions'``, ``'DeleteFeature'``, and
``'FastSetNextByIndex'``.
``Feature``
-----------
.. class:: Feature
``Feature`` wraps an OGR feature. You never create a ``Feature`` object
directly. Instead, you retrieve them from a :class:`Layer` object. Each
feature consists of a geometry and a set of fields containing additional
properties. The geometry of a field is accessible via its ``geom`` property,
which returns an :class:`OGRGeometry` object. A ``Feature`` behaves like a
standard Python container for its fields, which it returns as :class:`Field`
objects: you can access a field directly by its index or name, or you can
iterate over a feature's fields, e.g. in a ``for`` loop.
.. attribute:: geom
Returns the geometry for this feature, as an ``OGRGeometry`` object:
.. code-block:: pycon
>>> city.geom.tuple
(-104.609252, 38.255001)
.. attribute:: get
A method that returns the value of the given field (specified by name)
for this feature, **not** a ``Field`` wrapper object:
.. code-block:: pycon
>>> city.get("Population")
102121
.. attribute:: geom_type
Returns the type of geometry for this feature, as an :class:`OGRGeomType`
object. This will be the same for all features in a given layer and is
equivalent to the :attr:`Layer.geom_type` property of the :class:`Layer`
object the feature came from.
.. attribute:: num_fields
Returns the number of fields of data associated with the feature. This will
be the same for all features in a given layer and is equivalent to the
:attr:`Layer.num_fields` property of the :class:`Layer` object the feature
came from.
.. attribute:: fields
Returns a list of the names of the fields of data associated with the
feature. This will be the same for all features in a given layer and is
equivalent to the :attr:`Layer.fields` property of the :class:`Layer`
object the feature came from.
.. attribute:: fid
Returns the feature identifier within the layer:
.. code-block:: pycon
>>> city.fid
0
.. attribute:: layer_name
Returns the name of the :class:`Layer` that the feature came from. This
will be the same for all features in a given layer:
.. code-block:: pycon
>>> city.layer_name
'cities'
.. attribute:: index
A method that returns the index of the given field name. This will be the
same for all features in a given layer:
.. code-block:: pycon
>>> city.index("Population")
1
``Field``
---------
.. class:: Field
.. attribute:: name
Returns the name of this field:
.. code-block:: pycon
>>> city["Name"].name
'Name'
.. attribute:: type
Returns the OGR type of this field, as an integer. The ``FIELD_CLASSES``
dictionary maps these values onto subclasses of ``Field``:
.. code-block:: pycon
>>> city["Density"].type
2
.. attribute:: type_name
Returns a string with the name of the data type of this field:
.. code-block:: pycon
>>> city["Name"].type_name
'String'
.. attribute:: value
Returns the value of this field. The ``Field`` class itself returns the
value as a string, but each subclass returns the value in the most
appropriate form:
.. code-block:: pycon
>>> city["Population"].value
102121
.. attribute:: width
Returns the width of this field:
.. code-block:: pycon
>>> city["Name"].width
80
.. attribute:: precision
Returns the numeric precision of this field. This is meaningless (and set
to zero) for non-numeric fields:
.. code-block:: pycon
>>> city["Density"].precision
15
.. method:: as_double()
Returns the value of the field as a double (float):
.. code-block:: pycon
>>> city["Density"].as_double()
874.7
.. method:: as_int()
Returns the value of the field as an integer:
.. code-block:: pycon
>>> city["Population"].as_int()
102121
.. method:: as_string()
Returns the value of the field as a string:
.. code-block:: pycon
>>> city["Name"].as_string()
'Pueblo'
.. method:: as_datetime()
Returns the value of the field as a tuple of date and time components:
.. code-block:: pycon
>>> city["Created"].as_datetime()
(c_long(1999), c_long(5), c_long(23), c_long(0), c_long(0), c_long(0), c_long(0))
``Driver``
----------
.. class:: Driver(dr_input)
The ``Driver`` class is used internally to wrap an OGR :class:`DataSource`
driver.
.. attribute:: driver_count
Returns the number of OGR vector drivers currently registered.
OGR Geometries
==============
``OGRGeometry``
---------------
:class:`OGRGeometry` objects share similar functionality with
:class:`~django.contrib.gis.geos.GEOSGeometry` objects and are thin wrappers
around OGR's internal geometry representation. Thus, they allow for more
efficient access to data when using :class:`DataSource`. Unlike its GEOS
counterpart, :class:`OGRGeometry` supports spatial reference systems and
coordinate transformation:
.. code-block:: pycon
>>> from django.contrib.gis.gdal import OGRGeometry
>>> polygon = OGRGeometry("POLYGON((0 0, 5 0, 5 5, 0 5))")
.. class:: OGRGeometry(geom_input, srs=None)
This object is a wrapper for the `OGR Geometry`__ class. These objects are
instantiated directly from the given ``geom_input`` parameter, which may be
a string containing WKT, HEX, GeoJSON, a ``buffer`` containing WKB data, or
an :class:`OGRGeomType` object. These objects are also returned from the
:class:`Feature.geom` attribute, when reading vector data from
:class:`Layer` (which is in turn a part of a :class:`DataSource`).
__ https://gdal.org/api/ogrgeometry_cpp.html#ogrgeometry-class
.. classmethod:: from_gml(gml_string)
Constructs an :class:`OGRGeometry` from the given GML string.
.. classmethod:: from_bbox(bbox)
Constructs a :class:`Polygon` from the given bounding-box (a 4-tuple).
.. method:: __len__()
Returns the number of points in a :class:`LineString`, the number of rings
in a :class:`Polygon`, or the number of geometries in a
:class:`GeometryCollection`. Not applicable to other geometry types.
.. method:: __iter__()
Iterates over the points in a :class:`LineString`, the rings in a
:class:`Polygon`, or the geometries in a :class:`GeometryCollection`.
Not applicable to other geometry types.
.. method:: __getitem__()
Returns the point at the specified index for a :class:`LineString`, the
interior ring at the specified index for a :class:`Polygon`, or the geometry
at the specified index in a :class:`GeometryCollection`. Not applicable to
other geometry types.
.. attribute:: dimension
Returns the number of coordinated dimensions of the geometry, i.e. 0
for points, 1 for lines, and so forth:
.. code-block:: pycon
>> polygon.dimension
2
.. attribute:: coord_dim
Returns or sets the coordinate dimension of this geometry. For example, the
value would be 2 for two-dimensional geometries.
.. attribute:: geom_count
Returns the number of elements in this geometry:
.. code-block:: pycon
>>> polygon.geom_count
1
.. attribute:: point_count
Returns the number of points used to describe this geometry:
.. code-block:: pycon
>>> polygon.point_count
4
.. attribute:: num_points
Alias for :attr:`point_count`.
.. attribute:: num_coords
Alias for :attr:`point_count`.
.. attribute:: geom_type
Returns the type of this geometry, as an :class:`OGRGeomType` object.
.. attribute:: geom_name
Returns the name of the type of this geometry:
.. code-block:: pycon
>>> polygon.geom_name
'POLYGON'
.. attribute:: area
Returns the area of this geometry, or 0 for geometries that do not contain
an area:
.. code-block:: pycon
>>> polygon.area
25.0
.. attribute:: envelope
Returns the envelope of this geometry, as an :class:`Envelope` object.
.. attribute:: extent
Returns the envelope of this geometry as a 4-tuple, instead of as an
:class:`Envelope` object:
.. code-block:: pycon
>>> point.extent
(0.0, 0.0, 5.0, 5.0)
.. attribute:: srs
This property controls the spatial reference for this geometry, or
``None`` if no spatial reference system has been assigned to it.
If assigned, accessing this property returns a :class:`SpatialReference`
object. It may be set with another :class:`SpatialReference` object,
or any input that :class:`SpatialReference` accepts. Example:
.. code-block:: pycon
>>> city.geom.srs.name
'GCS_WGS_1984'
.. attribute:: srid
Returns or sets the spatial reference identifier corresponding to
:class:`SpatialReference` of this geometry. Returns ``None`` if
there is no spatial reference information associated with this
geometry, or if an SRID cannot be determined.
.. attribute:: geos
Returns a :class:`~django.contrib.gis.geos.GEOSGeometry` object
corresponding to this geometry.
.. attribute:: gml
Returns a string representation of this geometry in GML format:
.. code-block:: pycon
>>> OGRGeometry("POINT(1 2)").gml
'<gml:Point><gml:coordinates>1,2</gml:coordinates></gml:Point>'
.. attribute:: hex
Returns a string representation of this geometry in HEX WKB format:
.. code-block:: pycon
>>> OGRGeometry("POINT(1 2)").hex
'0101000000000000000000F03F0000000000000040'
.. attribute:: json
Returns a string representation of this geometry in JSON format:
.. code-block:: pycon
>>> OGRGeometry("POINT(1 2)").json
'{ "type": "Point", "coordinates": [ 1.000000, 2.000000 ] }'
.. attribute:: kml
Returns a string representation of this geometry in KML format.
.. attribute:: wkb_size
Returns the size of the WKB buffer needed to hold a WKB representation
of this geometry:
.. code-block:: pycon
>>> OGRGeometry("POINT(1 2)").wkb_size
21
.. attribute:: wkb
Returns a ``buffer`` containing a WKB representation of this geometry.
.. attribute:: wkt
Returns a string representation of this geometry in WKT format.
.. attribute:: ewkt
Returns the EWKT representation of this geometry.
.. method:: clone()
Returns a new :class:`OGRGeometry` clone of this geometry object.
.. method:: close_rings()
If there are any rings within this geometry that have not been closed,
this routine will do so by adding the starting point to the end:
.. code-block:: pycon
>>> triangle = OGRGeometry("LINEARRING (0 0,0 1,1 0)")
>>> triangle.close_rings()
>>> triangle.wkt
'LINEARRING (0 0,0 1,1 0,0 0)'
.. method:: transform(coord_trans, clone=False)
Transforms this geometry to a different spatial reference system. May take
a :class:`CoordTransform` object, a :class:`SpatialReference` object, or
any other input accepted by :class:`SpatialReference` (including spatial
reference WKT and PROJ strings, or an integer SRID).
By default nothing is returned and the geometry is transformed in-place.
However, if the ``clone`` keyword is set to ``True`` then a transformed
clone of this geometry is returned instead.
.. method:: intersects(other)
Returns ``True`` if this geometry intersects the other, otherwise returns
``False``.
.. method:: equals(other)
Returns ``True`` if this geometry is equivalent to the other, otherwise
returns ``False``.
.. method:: disjoint(other)
Returns ``True`` if this geometry is spatially disjoint to (i.e. does
not intersect) the other, otherwise returns ``False``.
.. method:: touches(other)
Returns ``True`` if this geometry touches the other, otherwise returns
``False``.
.. method:: crosses(other)
Returns ``True`` if this geometry crosses the other, otherwise returns
``False``.
.. method:: within(other)
Returns ``True`` if this geometry is contained within the other, otherwise
returns ``False``.
.. method:: contains(other)
Returns ``True`` if this geometry contains the other, otherwise returns
``False``.
.. method:: overlaps(other)
Returns ``True`` if this geometry overlaps the other, otherwise returns
``False``.
.. method:: boundary()
The boundary of this geometry, as a new :class:`OGRGeometry` object.
.. attribute:: convex_hull
The smallest convex polygon that contains this geometry, as a new
:class:`OGRGeometry` object.
.. method:: difference()
Returns the region consisting of the difference of this geometry and
the other, as a new :class:`OGRGeometry` object.
.. method:: intersection()
Returns the region consisting of the intersection of this geometry and
the other, as a new :class:`OGRGeometry` object.
.. method:: sym_difference()
Returns the region consisting of the symmetric difference of this
geometry and the other, as a new :class:`OGRGeometry` object.
.. method:: union()
Returns the region consisting of the union of this geometry and
the other, as a new :class:`OGRGeometry` object.
.. attribute:: tuple
Returns the coordinates of a point geometry as a tuple, the
coordinates of a line geometry as a tuple of tuples, and so forth:
.. code-block:: pycon
>>> OGRGeometry("POINT (1 2)").tuple
(1.0, 2.0)
>>> OGRGeometry("LINESTRING (1 2,3 4)").tuple
((1.0, 2.0), (3.0, 4.0))
.. attribute:: coords
An alias for :attr:`tuple`.
.. class:: Point
.. attribute:: x
Returns the X coordinate of this point:
.. code-block:: pycon
>>> OGRGeometry("POINT (1 2)").x
1.0
.. attribute:: y
Returns the Y coordinate of this point:
.. code-block:: pycon
>>> OGRGeometry("POINT (1 2)").y
2.0
.. attribute:: z
Returns the Z coordinate of this point, or ``None`` if the point does not
have a Z coordinate:
.. code-block:: pycon
>>> OGRGeometry("POINT (1 2 3)").z
3.0
.. class:: LineString
.. attribute:: x
Returns a list of X coordinates in this line:
.. code-block:: pycon
>>> OGRGeometry("LINESTRING (1 2,3 4)").x
[1.0, 3.0]
.. attribute:: y
Returns a list of Y coordinates in this line:
.. code-block:: pycon
>>> OGRGeometry("LINESTRING (1 2,3 4)").y
[2.0, 4.0]
.. attribute:: z
Returns a list of Z coordinates in this line, or ``None`` if the line does
not have Z coordinates:
.. code-block:: pycon
>>> OGRGeometry("LINESTRING (1 2 3,4 5 6)").z
[3.0, 6.0]
.. class:: Polygon
.. attribute:: shell
Returns the shell or exterior ring of this polygon, as a ``LinearRing``
geometry.
.. attribute:: exterior_ring
An alias for :attr:`shell`.
.. attribute:: centroid
Returns a :class:`Point` representing the centroid of this polygon.
.. class:: GeometryCollection
.. method:: add(geom)
Adds a geometry to this geometry collection. Not applicable to other
geometry types.
``OGRGeomType``
---------------
.. class:: OGRGeomType(type_input)
This class allows for the representation of an OGR geometry type
in any of several ways:
.. code-block:: pycon
>>> from django.contrib.gis.gdal import OGRGeomType
>>> gt1 = OGRGeomType(3) # Using an integer for the type
>>> gt2 = OGRGeomType("Polygon") # Using a string
>>> gt3 = OGRGeomType("POLYGON") # It's case-insensitive
>>> print(gt1 == 3, gt1 == "Polygon") # Equivalence works w/non-OGRGeomType objects
True True
.. attribute:: name
Returns a short-hand string form of the OGR Geometry type:
.. code-block:: pycon
>>> gt1.name
'Polygon'
.. attribute:: num
Returns the number corresponding to the OGR geometry type:
.. code-block:: pycon
>>> gt1.num
3
.. attribute:: django
Returns the Django field type (a subclass of GeometryField) to use for
storing this OGR type, or ``None`` if there is no appropriate Django type:
.. code-block:: pycon
>>> gt1.django
'PolygonField'
``Envelope``
------------
.. class:: Envelope(*args)
Represents an OGR Envelope structure that contains the minimum and maximum
X, Y coordinates for a rectangle bounding box. The naming of the variables
is compatible with the OGR Envelope C structure.
.. attribute:: min_x
The value of the minimum X coordinate.
.. attribute:: min_y
The value of the maximum X coordinate.
.. attribute:: max_x
The value of the minimum Y coordinate.
.. attribute:: max_y
The value of the maximum Y coordinate.
.. attribute:: ur
The upper-right coordinate, as a tuple.
.. attribute:: ll
The lower-left coordinate, as a tuple.
.. attribute:: tuple
A tuple representing the envelope.
.. attribute:: wkt
A string representing this envelope as a polygon in WKT format.
.. method:: expand_to_include(*args)
Coordinate System Objects
=========================
``SpatialReference``
--------------------
.. class:: SpatialReference(srs_input)
Spatial reference objects are initialized on the given ``srs_input``,
which may be one of the following:
* OGC Well Known Text (WKT) (a string)
* EPSG code (integer or string)
* PROJ string
* A shorthand string for well-known standards (``'WGS84'``, ``'WGS72'``,
``'NAD27'``, ``'NAD83'``)
Example:
.. code-block:: pycon
>>> wgs84 = SpatialReference("WGS84") # shorthand string
>>> wgs84 = SpatialReference(4326) # EPSG code
>>> wgs84 = SpatialReference("EPSG:4326") # EPSG string
>>> proj = "+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs "
>>> wgs84 = SpatialReference(proj) # PROJ string
>>> wgs84 = SpatialReference(
... """GEOGCS["WGS 84",
... DATUM["WGS_1984",
... SPHEROID["WGS 84",6378137,298.257223563,
... AUTHORITY["EPSG","7030"]],
... AUTHORITY["EPSG","6326"]],
... PRIMEM["Greenwich",0,
... AUTHORITY["EPSG","8901"]],
... UNIT["degree",0.01745329251994328,
... AUTHORITY["EPSG","9122"]],
... AUTHORITY["EPSG","4326"]]"""
... ) # OGC WKT
.. method:: __getitem__(target)
Returns the value of the given string attribute node, ``None`` if the node
doesn't exist. Can also take a tuple as a parameter, (target, child), where
child is the index of the attribute in the WKT. For example:
.. code-block:: pycon
>>> wkt = 'GEOGCS["WGS 84", DATUM["WGS_1984, ... AUTHORITY["EPSG","4326"]]'
>>> srs = SpatialReference(wkt) # could also use 'WGS84', or 4326
>>> print(srs["GEOGCS"])
WGS 84
>>> print(srs["DATUM"])
WGS_1984
>>> print(srs["AUTHORITY"])
EPSG
>>> print(srs["AUTHORITY", 1]) # The authority value
4326
>>> print(srs["TOWGS84", 4]) # the fourth value in this wkt
0
>>> print(srs["UNIT|AUTHORITY"]) # For the units authority, have to use the pipe symbol.
EPSG
>>> print(srs["UNIT|AUTHORITY", 1]) # The authority value for the units
9122
.. method:: attr_value(target, index=0)
The attribute value for the given target node (e.g. ``'PROJCS'``).
The index keyword specifies an index of the child node to return.
.. method:: auth_name(target)
Returns the authority name for the given string target node.
.. method:: auth_code(target)
Returns the authority code for the given string target node.
.. method:: clone()
Returns a clone of this spatial reference object.
.. method:: identify_epsg()
This method inspects the WKT of this ``SpatialReference`` and will add EPSG
authority nodes where an EPSG identifier is applicable.
.. method:: from_esri()
Morphs this SpatialReference from ESRI's format to EPSG
.. method:: to_esri()
Morphs this SpatialReference to ESRI's format.
.. method:: validate()
Checks to see if the given spatial reference is valid, if not
an exception will be raised.
.. method:: import_epsg(epsg)
Import spatial reference from EPSG code.
.. method:: import_proj(proj)
Import spatial reference from PROJ string.
.. method:: import_user_input(user_input)
.. method:: import_wkt(wkt)
Import spatial reference from WKT.
.. method:: import_xml(xml)
Import spatial reference from XML.
.. attribute:: name
Returns the name of this Spatial Reference.
.. attribute:: srid
Returns the SRID of top-level authority, or ``None`` if undefined.
.. attribute:: linear_name
Returns the name of the linear units.
.. attribute:: linear_units
Returns the value of the linear units.
.. attribute:: angular_name
Returns the name of the angular units."
.. attribute:: angular_units
Returns the value of the angular units.
.. attribute:: units
Returns a 2-tuple of the units value and the units name and will
automatically determines whether to return the linear or angular units.
.. attribute:: ellipsoid
Returns a tuple of the ellipsoid parameters for this spatial reference:
(semimajor axis, semiminor axis, and inverse flattening).
.. attribute:: semi_major
Returns the semi major axis of the ellipsoid for this spatial reference.
.. attribute:: semi_minor
Returns the semi minor axis of the ellipsoid for this spatial reference.
.. attribute:: inverse_flattening
Returns the inverse flattening of the ellipsoid for this spatial reference.
.. attribute:: geographic
Returns ``True`` if this spatial reference is geographic (root node is
``GEOGCS``).
.. attribute:: local
Returns ``True`` if this spatial reference is local (root node is
``LOCAL_CS``).
.. attribute:: projected
Returns ``True`` if this spatial reference is a projected coordinate system
(root node is ``PROJCS``).
.. attribute:: wkt
Returns the WKT representation of this spatial reference.
.. attribute:: pretty_wkt
Returns the 'pretty' representation of the WKT.
.. attribute:: proj
Returns the PROJ representation for this spatial reference.
.. attribute:: proj4
Alias for :attr:`SpatialReference.proj`.
.. attribute:: xml
Returns the XML representation of this spatial reference.
``CoordTransform``
------------------
.. class:: CoordTransform(source, target)
Represents a coordinate system transform. It is initialized with two
:class:`SpatialReference`, representing the source and target coordinate
systems, respectively. These objects should be used when performing the same
coordinate transformation repeatedly on different geometries:
.. code-block:: pycon
>>> ct = CoordTransform(SpatialReference("WGS84"), SpatialReference("NAD83"))
>>> for feat in layer:
... geom = feat.geom # getting clone of feature geometry
... geom.transform(ct) # transforming
...
.. _raster-data-source-objects:
Raster Data Objects
===================
``GDALRaster``
----------------
:class:`GDALRaster` is a wrapper for the GDAL raster source object that
supports reading data from a variety of GDAL-supported geospatial file
formats and data sources using a consistent interface. Each
data source is represented by a :class:`GDALRaster` object which contains
one or more layers of data named bands. Each band, represented by a
:class:`GDALBand` object, contains georeferenced image data. For example, an RGB
image is represented as three bands: one for red, one for green, and one for
blue.
.. note::
For raster data there is no difference between a raster instance and its
data source. Unlike for the Geometry objects, :class:`GDALRaster` objects are
always a data source. Temporary rasters can be instantiated in memory
using the corresponding driver, but they will be of the same class as file-based
raster sources.
.. class:: GDALRaster(ds_input, write=False)
The constructor for ``GDALRaster`` accepts two parameters. The first
parameter defines the raster source, and the second parameter defines if a
raster should be opened in write mode. For newly-created rasters, the second
parameter is ignored and the new raster is always created in write mode.
The first parameter can take three forms: a string or
:class:`~pathlib.Path` representing a file path (filesystem or GDAL virtual
filesystem), a dictionary with values defining a new raster, or a bytes
object representing a raster file.
If the input is a file path, the raster is opened from there. If the input
is raw data in a dictionary, the parameters ``width``, ``height``, and
``srid`` are required. If the input is a bytes object, it will be opened
using a GDAL virtual filesystem.
For a detailed description of how to create rasters using dictionary input,
see :ref:`gdal-raster-ds-input`. For a detailed description of how to
create rasters in the virtual filesystem, see :ref:`gdal-raster-vsimem`.
The following example shows how rasters can be created from different input
sources (using the sample data from the GeoDjango tests; see also the
:ref:`gdal_sample_data` section).
>>> from django.contrib.gis.gdal import GDALRaster
>>> rst = GDALRaster('/path/to/your/raster.tif', write=False)
>>> rst.name
'/path/to/your/raster.tif'
>>> rst.width, rst.height # This file has 163 x 174 pixels
(163, 174)
>>> rst = GDALRaster({ # Creates an in-memory raster
... 'srid': 4326,
... 'width': 4,
... 'height': 4,
... 'datatype': 1,
... 'bands': [{
... 'data': (2, 3),
... 'offset': (1, 1),
... 'size': (2, 2),
... 'shape': (2, 1),
... 'nodata_value': 5,
... }]
... })
>>> rst.srs.srid
4326
>>> rst.width, rst.height
(4, 4)
>>> rst.bands[0].data()
array([[5, 5, 5, 5],
[5, 2, 3, 5],
[5, 2, 3, 5],
[5, 5, 5, 5]], dtype=uint8)
>>> rst_file = open('/path/to/your/raster.tif', 'rb')
>>> rst_bytes = rst_file.read()
>>> rst = GDALRaster(rst_bytes)
>>> rst.is_vsi_based
True
>>> rst.name # Stored in a random path in the vsimem filesystem.
'/vsimem/da300bdb-129d-49a8-b336-e410a9428dad'
.. versionchanged:: 4.2
Support for :class:`pathlib.Path` ``ds_input`` was added.
.. attribute:: name
The name of the source which is equivalent to the input file path or the name
provided upon instantiation.
>>> GDALRaster({'width': 10, 'height': 10, 'name': 'myraster', 'srid': 4326}).name
'myraster'
.. attribute:: driver
The name of the GDAL driver used to handle the input file. For ``GDALRaster``\s created
from a file, the driver type is detected automatically. The creation of rasters from
scratch is an in-memory raster by default (``'MEM'``), but can be
altered as needed. For instance, use ``GTiff`` for a ``GeoTiff`` file.
For a list of file types, see also the `GDAL Raster Formats`__ list.
__ https://gdal.org/drivers/raster/
An in-memory raster is created through the following example:
>>> GDALRaster({'width': 10, 'height': 10, 'srid': 4326}).driver.name
'MEM'
A file based GeoTiff raster is created through the following example:
>>> import tempfile
>>> rstfile = tempfile.NamedTemporaryFile(suffix='.tif')
>>> rst = GDALRaster({'driver': 'GTiff', 'name': rstfile.name, 'srid': 4326,
... 'width': 255, 'height': 255, 'nr_of_bands': 1})
>>> rst.name
'/tmp/tmp7x9H4J.tif' # The exact filename will be different on your computer
>>> rst.driver.name
'GTiff'
.. attribute:: width
The width of the source in pixels (X-axis).
>>> GDALRaster({'width': 10, 'height': 20, 'srid': 4326}).width
10
.. attribute:: height
The height of the source in pixels (Y-axis).
>>> GDALRaster({'width': 10, 'height': 20, 'srid': 4326}).height
20
.. attribute:: srs
The spatial reference system of the raster, as a
:class:`SpatialReference` instance. The SRS can be changed by
setting it to an other :class:`SpatialReference` or providing any input
that is accepted by the :class:`SpatialReference` constructor.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.srs.srid
4326
>>> rst.srs = 3086
>>> rst.srs.srid
3086
.. attribute:: srid
The Spatial Reference System Identifier (SRID) of the raster. This
property is a shortcut to getting or setting the SRID through the
:attr:`srs` attribute.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.srid
4326
>>> rst.srid = 3086
>>> rst.srid
3086
>>> rst.srs.srid # This is equivalent
3086
.. attribute:: geotransform
The affine transformation matrix used to georeference the source, as a
tuple of six coefficients which map pixel/line coordinates into
georeferenced space using the following relationship::
Xgeo = GT(0) + Xpixel * GT(1) + Yline * GT(2)
Ygeo = GT(3) + Xpixel * GT(4) + Yline * GT(5)
The same values can be retrieved by accessing the :attr:`origin`
(indices 0 and 3), :attr:`scale` (indices 1 and 5) and :attr:`skew`
(indices 2 and 4) properties.
The default is ``[0.0, 1.0, 0.0, 0.0, 0.0, -1.0]``.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.geotransform
[0.0, 1.0, 0.0, 0.0, 0.0, -1.0]
.. attribute:: origin
Coordinates of the top left origin of the raster in the spatial
reference system of the source, as a point object with ``x`` and ``y``
members.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.origin
[0.0, 0.0]
>>> rst.origin.x = 1
>>> rst.origin
[1.0, 0.0]
.. attribute:: scale
Pixel width and height used for georeferencing the raster, as a point
object with ``x`` and ``y`` members. See :attr:`geotransform` for more
information.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.scale
[1.0, -1.0]
>>> rst.scale.x = 2
>>> rst.scale
[2.0, -1.0]
.. attribute:: skew
Skew coefficients used to georeference the raster, as a point object
with ``x`` and ``y`` members. In case of north up images, these
coefficients are both ``0``.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.skew
[0.0, 0.0]
>>> rst.skew.x = 3
>>> rst.skew
[3.0, 0.0]
.. attribute:: extent
Extent (boundary values) of the raster source, as a 4-tuple
``(xmin, ymin, xmax, ymax)`` in the spatial reference system of the
source.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.extent
(0.0, -20.0, 10.0, 0.0)
>>> rst.origin.x = 100
>>> rst.extent
(100.0, -20.0, 110.0, 0.0)
.. attribute:: bands
List of all bands of the source, as :class:`GDALBand` instances.
>>> rst = GDALRaster({"width": 1, "height": 2, 'srid': 4326,
... "bands": [{"data": [0, 1]}, {"data": [2, 3]}]})
>>> len(rst.bands)
2
>>> rst.bands[1].data()
array([[ 2., 3.]], dtype=float32)
.. method:: warp(ds_input, resampling='NearestNeighbour', max_error=0.0)
Returns a warped version of this raster.
The warping parameters can be specified through the ``ds_input``
argument. The use of ``ds_input`` is analogous to the corresponding
argument of the class constructor. It is a dictionary with the
characteristics of the target raster. Allowed dictionary key values are
width, height, SRID, origin, scale, skew, datatype, driver, and name
(filename).
By default, the warp functions keeps most parameters equal to the
values of the original source raster, so only parameters that should be
changed need to be specified. Note that this includes the driver, so
for file-based rasters the warp function will create a new raster on
disk.
The only parameter that is set differently from the source raster is the
name. The default value of the raster name is the name of the source
raster appended with ``'_copy' + source_driver_name``. For file-based
rasters it is recommended to provide the file path of the target raster.
The resampling algorithm used for warping can be specified with the
``resampling`` argument. The default is ``NearestNeighbor``, and the
other allowed values are ``Bilinear``, ``Cubic``, ``CubicSpline``,
``Lanczos``, ``Average``, and ``Mode``.
The ``max_error`` argument can be used to specify the maximum error
measured in input pixels that is allowed in approximating the
transformation. The default is 0.0 for exact calculations.
For users familiar with ``GDAL``, this function has a similar
functionality to the ``gdalwarp`` command-line utility.
For example, the warp function can be used for aggregating a raster to
the double of its original pixel scale:
>>> rst = GDALRaster({
... "width": 6, "height": 6, "srid": 3086,
... "origin": [500000, 400000],
... "scale": [100, -100],
... "bands": [{"data": range(36), "nodata_value": 99}]
... })
>>> target = rst.warp({"scale": [200, -200], "width": 3, "height": 3})
>>> target.bands[0].data()
array([[ 7., 9., 11.],
[ 19., 21., 23.],
[ 31., 33., 35.]], dtype=float32)
.. method:: transform(srs, driver=None, name=None, resampling='NearestNeighbour', max_error=0.0)
Transforms this raster to a different spatial reference system
(``srs``), which may be a :class:`SpatialReference` object, or any
other input accepted by :class:`SpatialReference` (including spatial
reference WKT and PROJ strings, or an integer SRID).
It calculates the bounds and scale of the current raster in the new
spatial reference system and warps the raster using the
:attr:`~GDALRaster.warp` function.
By default, the driver of the source raster is used and the name of the
raster is the original name appended with
``'_copy' + source_driver_name``. A different driver or name can be
specified with the ``driver`` and ``name`` arguments.
The default resampling algorithm is ``NearestNeighbour`` but can be
changed using the ``resampling`` argument. The default maximum allowed
error for resampling is 0.0 and can be changed using the ``max_error``
argument. Consult the :attr:`~GDALRaster.warp` documentation for detail
on those arguments.
>>> rst = GDALRaster({
... "width": 6, "height": 6, "srid": 3086,
... "origin": [500000, 400000],
... "scale": [100, -100],
... "bands": [{"data": range(36), "nodata_value": 99}]
... })
>>> target_srs = SpatialReference(4326)
>>> target = rst.transform(target_srs)
>>> target.origin
[-82.98492744885776, 27.601924753080144]
.. attribute:: info
Returns a string with a summary of the raster. This is equivalent to
the `gdalinfo`__ command line utility.
__ https://gdal.org/programs/gdalinfo.html
.. attribute:: metadata
The metadata of this raster, represented as a nested dictionary. The
first-level key is the metadata domain. The second-level contains the
metadata item names and values from each domain.
To set or update a metadata item, pass the corresponding metadata item
to the method using the nested structure described above. Only keys
that are in the specified dictionary are updated; the rest of the
metadata remains unchanged.
To remove a metadata item, use ``None`` as the metadata value.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.metadata
{}
>>> rst.metadata = {'DEFAULT': {'OWNER': 'Django', 'VERSION': '1.0'}}
>>> rst.metadata
{'DEFAULT': {'OWNER': 'Django', 'VERSION': '1.0'}}
>>> rst.metadata = {'DEFAULT': {'OWNER': None, 'VERSION': '2.0'}}
>>> rst.metadata
{'DEFAULT': {'VERSION': '2.0'}}
.. attribute:: vsi_buffer
A ``bytes`` representation of this raster. Returns ``None`` for rasters
that are not stored in GDAL's virtual filesystem.
.. attribute:: is_vsi_based
A boolean indicating if this raster is stored in GDAL's virtual
filesystem.
``GDALBand``
------------
.. class:: GDALBand
``GDALBand`` instances are not created explicitly, but rather obtained
from a :class:`GDALRaster` object, through its :attr:`~GDALRaster.bands`
attribute. The GDALBands contain the actual pixel values of the raster.
.. attribute:: description
The name or description of the band, if any.
.. attribute:: width
The width of the band in pixels (X-axis).
.. attribute:: height
The height of the band in pixels (Y-axis).
.. attribute:: pixel_count
The total number of pixels in this band. Is equal to ``width * height``.
.. method:: statistics(refresh=False, approximate=False)
Compute statistics on the pixel values of this band. The return value
is a tuple with the following structure:
``(minimum, maximum, mean, standard deviation)``.
If the ``approximate`` argument is set to ``True``, the statistics may
be computed based on overviews or a subset of image tiles.
If the ``refresh`` argument is set to ``True``, the statistics will be
computed from the data directly, and the cache will be updated with the
result.
If a persistent cache value is found, that value is returned. For
raster formats using Persistent Auxiliary Metadata (PAM) services, the
statistics might be cached in an auxiliary file. In some cases this
metadata might be out of sync with the pixel values or cause values
from a previous call to be returned which don't reflect the value of
the ``approximate`` argument. In such cases, use the ``refresh``
argument to get updated values and store them in the cache.
For empty bands (where all pixel values are "no data"), all statistics
are returned as ``None``.
The statistics can also be retrieved directly by accessing the
:attr:`min`, :attr:`max`, :attr:`mean`, and :attr:`std` properties.
.. attribute:: min
The minimum pixel value of the band (excluding the "no data" value).
.. attribute:: max
The maximum pixel value of the band (excluding the "no data" value).
.. attribute:: mean
The mean of all pixel values of the band (excluding the "no data"
value).
.. attribute:: std
The standard deviation of all pixel values of the band (excluding the
"no data" value).
.. attribute:: nodata_value
The "no data" value for a band is generally a special marker value used
to mark pixels that are not valid data. Such pixels should generally not
be displayed, nor contribute to analysis operations.
To delete an existing "no data" value, set this property to ``None``.
.. method:: datatype(as_string=False)
The data type contained in the band, as an integer constant between 0
(Unknown) and 11. If ``as_string`` is ``True``, the data type is
returned as a string with the following possible values:
``GDT_Unknown``, ``GDT_Byte``, ``GDT_UInt16``, ``GDT_Int16``,
``GDT_UInt32``, ``GDT_Int32``, ``GDT_Float32``, ``GDT_Float64``,
``GDT_CInt16``, ``GDT_CInt32``, ``GDT_CFloat32``, and ``GDT_CFloat64``.
.. method:: color_interp(as_string=False)
The color interpretation for the band, as an integer between 0and 16.
If ``as_string`` is ``True``, the data type is returned as a string
with the following possible values:
``GCI_Undefined``, ``GCI_GrayIndex``, ``GCI_PaletteIndex``,
``GCI_RedBand``, ``GCI_GreenBand``, ``GCI_BlueBand``, ``GCI_AlphaBand``,
``GCI_HueBand``, ``GCI_SaturationBand``, ``GCI_LightnessBand``,
``GCI_CyanBand``, ``GCI_MagentaBand``, ``GCI_YellowBand``,
``GCI_BlackBand``, ``GCI_YCbCr_YBand``, ``GCI_YCbCr_CbBand``, and
``GCI_YCbCr_CrBand``. ``GCI_YCbCr_CrBand`` also represents ``GCI_Max``
because both correspond to the integer 16, but only ``GCI_YCbCr_CrBand``
is returned as a string.
.. method:: data(data=None, offset=None, size=None, shape=None)
The accessor to the pixel values of the ``GDALBand``. Returns the complete
data array if no parameters are provided. A subset of the pixel array can
be requested by specifying an offset and block size as tuples.
If NumPy is available, the data is returned as NumPy array. For performance
reasons, it is highly recommended to use NumPy.
Data is written to the ``GDALBand`` if the ``data`` parameter is provided.
The input can be of one of the following types - packed string, buffer, list,
array, and NumPy array. The number of items in the input should normally
correspond to the total number of pixels in the band, or to the number
of pixels for a specific block of pixel values if the ``offset`` and
``size`` parameters are provided.
If the number of items in the input is different from the target pixel
block, the ``shape`` parameter must be specified. The shape is a tuple
that specifies the width and height of the input data in pixels. The
data is then replicated to update the pixel values of the selected
block. This is useful to fill an entire band with a single value, for
instance.
For example:
>>> rst = GDALRaster({'width': 4, 'height': 4, 'srid': 4326, 'datatype': 1, 'nr_of_bands': 1})
>>> bnd = rst.bands[0]
>>> bnd.data(range(16))
>>> bnd.data()
array([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11],
[12, 13, 14, 15]], dtype=int8)
>>> bnd.data(offset=(1, 1), size=(2, 2))
array([[ 5, 6],
[ 9, 10]], dtype=int8)
>>> bnd.data(data=[-1, -2, -3, -4], offset=(1, 1), size=(2, 2))
>>> bnd.data()
array([[ 0, 1, 2, 3],
[ 4, -1, -2, 7],
[ 8, -3, -4, 11],
[12, 13, 14, 15]], dtype=int8)
>>> bnd.data(data='\x9d\xa8\xb3\xbe', offset=(1, 1), size=(2, 2))
>>> bnd.data()
array([[ 0, 1, 2, 3],
[ 4, -99, -88, 7],
[ 8, -77, -66, 11],
[ 12, 13, 14, 15]], dtype=int8)
>>> bnd.data([1], shape=(1, 1))
>>> bnd.data()
array([[1, 1, 1, 1],
[1, 1, 1, 1],
[1, 1, 1, 1],
[1, 1, 1, 1]], dtype=uint8)
>>> bnd.data(range(4), shape=(1, 4))
array([[0, 0, 0, 0],
[1, 1, 1, 1],
[2, 2, 2, 2],
[3, 3, 3, 3]], dtype=uint8)
.. attribute:: metadata
The metadata of this band. The functionality is identical to
:attr:`GDALRaster.metadata`.
.. _gdal-raster-ds-input:
Creating rasters from data
--------------------------
This section describes how to create rasters from scratch using the
``ds_input`` parameter.
A new raster is created when a ``dict`` is passed to the :class:`GDALRaster`
constructor. The dictionary contains defining parameters of the new raster,
such as the origin, size, or spatial reference system. The dictionary can also
contain pixel data and information about the format of the new raster. The
resulting raster can therefore be file-based or memory-based, depending on the
driver specified.
There's no standard for describing raster data in a dictionary or JSON flavor.
The definition of the dictionary input to the :class:`GDALRaster` class is
therefore specific to Django. It's inspired by the `geojson`__ format, but the
``geojson`` standard is currently limited to vector formats.
Examples of using the different keys when creating rasters can be found in the
documentation of the corresponding attributes and methods of the
:class:`GDALRaster` and :class:`GDALBand` classes.
__ https://geojson.org/
The ``ds_input`` dictionary
~~~~~~~~~~~~~~~~~~~~~~~~~~~
Only a few keys are required in the ``ds_input`` dictionary to create a raster:
``width``, ``height``, and ``srid``. All other parameters have default values
(see the table below). The list of keys that can be passed in the ``ds_input``
dictionary is closely related but not identical to the :class:`GDALRaster`
properties. Many of the parameters are mapped directly to those properties;
the others are described below.
The following table describes all keys that can be set in the ``ds_input``
dictionary.
================= ======== ==================================================
Key Default Usage
================= ======== ==================================================
``srid`` required Mapped to the :attr:`~GDALRaster.srid` attribute
``width`` required Mapped to the :attr:`~GDALRaster.width` attribute
``height`` required Mapped to the :attr:`~GDALRaster.height` attribute
``driver`` ``MEM`` Mapped to the :attr:`~GDALRaster.driver` attribute
``name`` ``''`` See below
``origin`` ``0`` Mapped to the :attr:`~GDALRaster.origin` attribute
``scale`` ``0`` Mapped to the :attr:`~GDALRaster.scale` attribute
``skew`` ``0`` Mapped to the :attr:`~GDALRaster.width` attribute
``bands`` ``[]`` See below
``nr_of_bands`` ``0`` See below
``datatype`` ``6`` See below
``papsz_options`` ``{}`` See below
================= ======== ==================================================
.. object:: name
String representing the name of the raster. When creating a file-based
raster, this parameter must be the file path for the new raster. If the
name starts with ``/vsimem/``, the raster is created in GDAL's virtual
filesystem.
.. object:: datatype
Integer representing the data type for all the bands. Defaults to ``6``
(Float32). All bands of a new raster are required to have the same datatype.
The value mapping is:
===== =============== ===============================
Value GDAL Pixel Type Description
===== =============== ===============================
1 GDT_Byte Eight bit unsigned integer
2 GDT_UInt16 Sixteen bit unsigned integer
3 GDT_Int16 Sixteen bit signed integer
4 GDT_UInt32 Thirty-two bit unsigned integer
5 GDT_Int32 Thirty-two bit signed integer
6 GDT_Float32 Thirty-two bit floating point
7 GDT_Float64 Sixty-four bit floating point
===== =============== ===============================
.. object:: nr_of_bands
Integer representing the number of bands of the raster. A raster can be
created without passing band data upon creation. If the number of bands
isn't specified, it's automatically calculated from the length of the
``bands`` input. The number of bands can't be changed after creation.
.. object:: bands
A list of ``band_input`` dictionaries with band input data. The resulting
band indices are the same as in the list provided. The definition of the
band input dictionary is given below. If band data isn't provided, the
raster bands values are instantiated as an array of zeros and the "no
data" value is set to ``None``.
.. object:: papsz_options
A dictionary with raster creation options. The key-value pairs of the
input dictionary are passed to the driver on creation of the raster.
The available options are driver-specific and are described in the
documentation of each driver.
The values in the dictionary are not case-sensitive and are automatically
converted to the correct string format upon creation.
The following example uses some of the options available for the
`GTiff driver`__. The result is a compressed signed byte raster with an
internal tiling scheme. The internal tiles have a block size of 23 by 23:
.. code-block:: pycon
>>> GDALRaster(
... {
... "driver": "GTiff",
... "name": "/path/to/new/file.tif",
... "srid": 4326,
... "width": 255,
... "height": 255,
... "nr_of_bands": 1,
... "papsz_options": {
... "compress": "packbits",
... "pixeltype": "signedbyte",
... "tiled": "yes",
... "blockxsize": 23,
... "blockysize": 23,
... },
... }
... )
__ https://gdal.org/drivers/raster/gtiff.html
The ``band_input`` dictionary
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The ``bands`` key in the ``ds_input`` dictionary is a list of ``band_input``
dictionaries. Each ``band_input`` dictionary can contain pixel values and the
"no data" value to be set on the bands of the new raster. The data array can
have the full size of the new raster or be smaller. For arrays that are smaller
than the full raster, the ``size``, ``shape``, and ``offset`` keys control the
pixel values. The corresponding keys are passed to the :meth:`~GDALBand.data`
method. Their functionality is the same as setting the band data with that
method. The following table describes the keys that can be used.
================ ================================= ======================================================
Key Default Usage
================ ================================= ======================================================
``nodata_value`` ``None`` Mapped to the :attr:`~GDALBand.nodata_value` attribute
``data`` Same as ``nodata_value`` or ``0`` Passed to the :meth:`~GDALBand.data` method
``size`` ``(with, height)`` of raster Passed to the :meth:`~GDALBand.data` method
``shape`` Same as size Passed to the :meth:`~GDALBand.data` method
``offset`` ``(0, 0)`` Passed to the :meth:`~GDALBand.data` method
================ ================================= ======================================================
.. _gdal-raster-vsimem:
Using GDAL's Virtual Filesystem
-------------------------------
GDAL can access files stored in the filesystem, but also supports virtual
filesystems to abstract accessing other kind of files, such as compressed,
encrypted, or remote files.
Using memory-based Virtual Filesystem
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
GDAL has an internal memory-based filesystem, which allows treating blocks of
memory as files. It can be used to read and write :class:`GDALRaster` objects
to and from binary file buffers.
This is useful in web contexts where rasters might be obtained as a buffer
from a remote storage or returned from a view without being written to disk.
:class:`GDALRaster` objects are created in the virtual filesystem when a
``bytes`` object is provided as input, or when the file path starts with
``/vsimem/``.
Input provided as ``bytes`` has to be a full binary representation of a file.
For instance:
.. code-block:: pycon
# Read a raster as a file object from a remote source.
>>> from urllib.request import urlopen
>>> dat = urlopen("http://example.com/raster.tif").read()
# Instantiate a raster from the bytes object.
>>> rst = GDALRaster(dat)
# The name starts with /vsimem/, indicating that the raster lives in the
# virtual filesystem.
>>> rst.name
'/vsimem/da300bdb-129d-49a8-b336-e410a9428dad'
To create a new virtual file-based raster from scratch, use the ``ds_input``
dictionary representation and provide a ``name`` argument that starts with
``/vsimem/`` (for detail of the dictionary representation, see
:ref:`gdal-raster-ds-input`). For virtual file-based rasters, the
:attr:`~GDALRaster.vsi_buffer` attribute returns the ``bytes`` representation
of the raster.
Here's how to create a raster and return it as a file in an
:class:`~django.http.HttpResponse`:
.. code-block:: pycon
>>> from django.http import HttpResponse
>>> rst = GDALRaster(
... {
... "name": "/vsimem/temporarymemfile",
... "driver": "tif",
... "width": 6,
... "height": 6,
... "srid": 3086,
... "origin": [500000, 400000],
... "scale": [100, -100],
... "bands": [{"data": range(36), "nodata_value": 99}],
... }
... )
>>> HttpResponse(rast.vsi_buffer, "image/tiff")
Using other Virtual Filesystems
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Depending on the local build of GDAL other virtual filesystems may be
supported. You can use them by prepending the provided path with the
appropriate ``/vsi*/`` prefix. See the `GDAL Virtual Filesystems
documentation`_ for more details.
.. warning:
Rasters with names starting with `/vsi*/` will be treated as rasters from
the GDAL virtual filesystems. Django doesn't perform any extra validation.
Compressed rasters
^^^^^^^^^^^^^^^^^^
Instead decompressing the file and instantiating the resulting raster, GDAL can
directly access compressed files using the ``/vsizip/``, ``/vsigzip/``, or
``/vsitar/`` virtual filesystems:
.. code-block:: pycon
>>> from django.contrib.gis.gdal import GDALRaster
>>> rst = GDALRaster("/vsizip/path/to/your/file.zip/path/to/raster.tif")
>>> rst = GDALRaster("/vsigzip/path/to/your/file.gz")
>>> rst = GDALRaster("/vsitar/path/to/your/file.tar/path/to/raster.tif")
Network rasters
^^^^^^^^^^^^^^^
GDAL can support online resources and storage providers transparently. As long
as it's built with such capabilities.
To access a public raster file with no authentication, you can use
``/vsicurl/``:
.. code-block:: pycon
>>> from django.contrib.gis.gdal import GDALRaster
>>> rst = GDALRaster("/vsicurl/https://example.com/raster.tif")
>>> rst.name
'/vsicurl/https://example.com/raster.tif'
For commercial storage providers (e.g. ``/vsis3/``) the system should be
previously configured for authentication and possibly other settings (see the
`GDAL Virtual Filesystems documentation`_ for available options).
.. _`GDAL Virtual Filesystems documentation`: https://gdal.org/user/virtual_file_systems.html
Settings
========
.. setting:: GDAL_LIBRARY_PATH
``GDAL_LIBRARY_PATH``
---------------------
A string specifying the location of the GDAL library. Typically,
this setting is only used if the GDAL library is in a non-standard
location (e.g., ``/home/john/lib/libgdal.so``).
Exceptions
==========
.. exception:: GDALException
The base GDAL exception, indicating a GDAL-related error.
.. exception:: SRSException
An exception raised when an error occurs when constructing or using a
spatial reference system object.
Reference in New Issue View Git Blame Copy Permalink