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.
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 simple data sets that you can use for testing. You can download them here:
$ wget https://raw.githubusercontent.com/django/django/master/tests/gis_tests/data/cities/cities.{shp,prj,shx,dbf}
$ wget https://raw.githubusercontent.com/django/django/master/tests/gis_tests/data/rasters/raster.tif
DataSource¶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 simple, consistent interface. Each
data source is represented by a DataSource object which contains
one or more layers of data. Each layer, represented by a Layer
object, contains some number of geographic features (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 (Field) of data that may be associated with
each feature in that layer.
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 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 layer_count property, or
(equivalently) by using the len() function. For information on
accessing the layers of data themselves, see the next section:
>>> 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
layer_count¶Returns the number of layers in the data source.
name¶Returns the name of the data source.
Layer¶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 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 geom_type property of a layer is an OGRGeomType that
identifies the feature type. We can use it to print out some basic
information about each layer in a DataSource:
>>> 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:
>>> layer = ds[0]
name¶Returns the name of this layer in the data source.
>>> layer.name
'cities'
num_feat¶Returns the number of features in the layer. Same as len(layer):
>>> layer.num_feat
3
geom_type¶Returns the geometry type of the layer, as an OGRGeomType object:
>>> layer.geom_type.name
'Point'
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:
>>> layer.num_fields
4
fields¶Returns a list of the names of each of the fields in this layer:
>>> layer.fields
['Name', 'Population', 'Density', 'Created']
Returns a list of the data types of each of the fields in this layer. These
are subclasses of Field, discussed below:
>>> [ft.__name__ for ft in layer.field_types]
['OFTString', 'OFTReal', 'OFTReal', 'OFTDate']
field_widths¶Returns a list of the maximum field widths for each of the fields in this layer:
>>> layer.field_widths
[80, 11, 24, 10]
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:
>>> layer.field_precisions
[0, 0, 15, 0]
extent¶Returns the spatial extent of this layer, as an Envelope object:
>>> layer.extent.tuple
(-104.609252, 29.763374, -95.23506, 38.971823)
srs¶Property that returns the SpatialReference associated with this
layer:
>>> 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 Layer has no spatial reference information associated
with it, None is returned.
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 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:
>>> 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
get_fields()¶A method that returns a list of the values of a given field for each feature in the layer:
>>> layer.get_fields('Name')
['Pueblo', 'Lawrence', 'Houston']
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 GEOSGeometry
objects. Otherwise, they are returned as OGRGeometry objects:
>>> [pt.tuple for pt in layer.get_geoms()]
[(-104.609252, 38.255001), (-95.23506, 38.971823), (-95.363151, 29.763374)]
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¶Feature¶Feature wraps an OGR feature. You never create a Feature object
directly. Instead, you retrieve them from a 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 OGRGeometry object. A Feature behaves like a
standard Python container for its fields, which it returns as 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.
geom¶Returns the geometry for this feature, as an OGRGeometry object:
>>> city.geom.tuple
(-104.609252, 38.255001)
get¶A method that returns the value of the given field (specified by name)
for this feature, not a Field wrapper object:
>>> city.get('Population')
102121
geom_type¶Returns the type of geometry for this feature, as an OGRGeomType
object. This will be the same for all features in a given layer and is
equivalent to the Layer.geom_type property of the Layer
object the feature came from.
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
Layer.num_fields property of the Layer object the feature
came from.
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 Layer.fields property of the Layer
object the feature came from.
fid¶Returns the feature identifier within the layer:
>>> city.fid
0
layer_name¶Returns the name of the Layer that the feature came from. This
will be the same for all features in a given layer:
>>> city.layer_name
'cities'
index¶A method that returns the index of the given field name. This will be the same for all features in a given layer:
>>> city.index('Population')
1
Field¶Field¶name¶Returns the name of this field:
>>> city['Name'].name
'Name'
type¶Returns the OGR type of this field, as an integer. The FIELD_CLASSES
dictionary maps these values onto subclasses of Field:
>>> city['Density'].type
2
type_name¶Returns a string with the name of the data type of this field:
>>> city['Name'].type_name
'String'
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:
>>> city['Population'].value
102121
width¶Returns the width of this field:
>>> city['Name'].width
80
precision¶Returns the numeric precision of this field. This is meaningless (and set to zero) for non-numeric fields:
>>> city['Density'].precision
15
as_double()¶Returns the value of the field as a double (float):
>>> city['Density'].as_double()
874.7
as_int()¶Returns the value of the field as an integer:
>>> city['Population'].as_int()
102121
as_string()¶Returns the value of the field as a string:
>>> city['Name'].as_string()
'Pueblo'
as_datetime()¶Returns the value of the field as a tuple of date and time components:
>>> 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¶Driver(dr_input)¶The Driver class is used internally to wrap an OGR DataSource
driver.
driver_count¶Returns the number of OGR vector drivers currently registered.
OGRGeometry¶OGRGeometry objects share similar functionality with
GEOSGeometry objects and are thin wrappers
around OGR’s internal geometry representation. Thus, they allow for more
efficient access to data when using DataSource. Unlike its GEOS
counterpart, OGRGeometry supports spatial reference systems and
coordinate transformation:
>>> from django.contrib.gis.gdal import OGRGeometry
>>> polygon = OGRGeometry('POLYGON((0 0, 5 0, 5 5, 0 5))')
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 OGRGeomType object. These objects are also returned from the
Feature.geom attribute, when reading vector data from
Layer (which is in turn a part of a DataSource).
from_gml(gml_string)¶Constructs an OGRGeometry from the given GML string.
from_bbox(bbox)¶Constructs a Polygon from the given bounding-box (a 4-tuple).
__len__()¶Returns the number of points in a LineString, the number of rings
in a Polygon, or the number of geometries in a
GeometryCollection. Not applicable to other geometry types.
__iter__()¶Iterates over the points in a LineString, the rings in a
Polygon, or the geometries in a GeometryCollection.
Not applicable to other geometry types.
__getitem__()¶Returns the point at the specified index for a LineString, the
interior ring at the specified index for a Polygon, or the geometry
at the specified index in a GeometryCollection. Not applicable to
other geometry types.
dimension¶Returns the number of coordinated dimensions of the geometry, i.e. 0 for points, 1 for lines, and so forth:
>> polygon.dimension
2
coord_dim¶Returns or sets the coordinate dimension of this geometry. For example, the value would be 2 for two-dimensional geometries.
geom_count¶Returns the number of elements in this geometry:
>>> polygon.geom_count
1
point_count¶Returns the number of points used to describe this geometry:
>>> polygon.point_count
4
num_points¶Alias for point_count.
num_coords¶Alias for point_count.
geom_type¶Returns the type of this geometry, as an OGRGeomType object.
geom_name¶Returns the name of the type of this geometry:
>>> polygon.geom_name
'POLYGON'
area¶Returns the area of this geometry, or 0 for geometries that do not contain an area:
>>> polygon.area
25.0
envelope¶Returns the envelope of this geometry, as an Envelope object.
extent¶Returns the envelope of this geometry as a 4-tuple, instead of as an
Envelope object:
>>> point.extent
(0.0, 0.0, 5.0, 5.0)
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 SpatialReference
object. It may be set with another SpatialReference object,
or any input that SpatialReference accepts. Example:
>>> city.geom.srs.name
'GCS_WGS_1984'
srid¶Returns or sets the spatial reference identifier corresponding to
SpatialReference of this geometry. Returns None if
there is no spatial reference information associated with this
geometry, or if an SRID cannot be determined.
geos¶Returns a GEOSGeometry object
corresponding to this geometry.
gml¶Returns a string representation of this geometry in GML format:
>>> OGRGeometry('POINT(1 2)').gml
'<gml:Point><gml:coordinates>1,2</gml:coordinates></gml:Point>'
hex¶Returns a string representation of this geometry in HEX WKB format:
>>> OGRGeometry('POINT(1 2)').hex
'0101000000000000000000F03F0000000000000040'
json¶Returns a string representation of this geometry in JSON format:
>>> OGRGeometry('POINT(1 2)').json
'{ "type": "Point", "coordinates": [ 1.000000, 2.000000 ] }'
kml¶Returns a string representation of this geometry in KML format.
wkb_size¶Returns the size of the WKB buffer needed to hold a WKB representation of this geometry:
>>> OGRGeometry('POINT(1 2)').wkb_size
21
wkb¶Returns a buffer containing a WKB representation of this geometry.
wkt¶Returns a string representation of this geometry in WKT format.
ewkt¶Returns the EWKT representation of this geometry.
clone()¶Returns a new OGRGeometry clone of this geometry object.
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:
>>> triangle = OGRGeometry('LINEARRING (0 0,0 1,1 0)')
>>> triangle.close_rings()
>>> triangle.wkt
'LINEARRING (0 0,0 1,1 0,0 0)'
transform(coord_trans, clone=False)¶Transforms this geometry to a different spatial reference system. May take
a CoordTransform object, a SpatialReference object, or
any other input accepted by SpatialReference (including spatial
reference WKT and PROJ.4 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.
intersects(other)¶Returns True if this geometry intersects the other, otherwise returns
False.
equals(other)¶Returns True if this geometry is equivalent to the other, otherwise
returns False.
disjoint(other)¶Returns True if this geometry is spatially disjoint to (i.e. does
not intersect) the other, otherwise returns False.
touches(other)¶Returns True if this geometry touches the other, otherwise returns
False.
crosses(other)¶Returns True if this geometry crosses the other, otherwise returns
False.
within(other)¶Returns True if this geometry is contained within the other, otherwise
returns False.
contains(other)¶Returns True if this geometry contains the other, otherwise returns
False.
overlaps(other)¶Returns True if this geometry overlaps the other, otherwise returns
False.
boundary()¶The boundary of this geometry, as a new OGRGeometry object.
convex_hull¶The smallest convex polygon that contains this geometry, as a new
OGRGeometry object.
difference()¶Returns the region consisting of the difference of this geometry and
the other, as a new OGRGeometry object.
intersection()¶Returns the region consisting of the intersection of this geometry and
the other, as a new OGRGeometry object.
sym_difference()¶Returns the region consisting of the symmetric difference of this
geometry and the other, as a new OGRGeometry object.
union()¶Returns the region consisting of the union of this geometry and
the other, as a new OGRGeometry object.
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:
>>> OGRGeometry('POINT (1 2)').tuple
(1.0, 2.0)
>>> OGRGeometry('LINESTRING (1 2,3 4)').tuple
((1.0, 2.0), (3.0, 4.0))
coords¶An alias for tuple.
Point¶x¶Returns the X coordinate of this point:
>>> OGRGeometry('POINT (1 2)').x
1.0
y¶Returns the Y coordinate of this point:
>>> OGRGeometry('POINT (1 2)').y
2.0
z¶Returns the Z coordinate of this point, or None if the point does not
have a Z coordinate:
>>> OGRGeometry('POINT (1 2 3)').z
3.0
LineString¶x¶Returns a list of X coordinates in this line:
>>> OGRGeometry('LINESTRING (1 2,3 4)').x
[1.0, 3.0]
y¶Returns a list of Y coordinates in this line:
>>> OGRGeometry('LINESTRING (1 2,3 4)').y
[2.0, 4.0]
z¶Returns a list of Z coordinates in this line, or None if the line does
not have Z coordinates:
>>> OGRGeometry('LINESTRING (1 2 3,4 5 6)').z
[3.0, 6.0]
OGRGeomType¶OGRGeomType(type_input)¶This class allows for the representation of an OGR geometry type in any of several ways:
>>> 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
name¶Returns a short-hand string form of the OGR Geometry type:
>>> gt1.name
'Polygon'
num¶Returns the number corresponding to the OGR geometry type:
>>> gt1.num
3
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:
>>> gt1.django
'PolygonField'
Envelope¶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.
min_x¶The value of the minimum X coordinate.
min_y¶The value of the maximum X coordinate.
max_x¶The value of the minimum Y coordinate.
max_y¶The value of the maximum Y coordinate.
ur¶The upper-right coordinate, as a tuple.
ll¶The lower-left coordinate, as a tuple.
tuple¶A tuple representing the envelope.
wkt¶A string representing this envelope as a polygon in WKT format.
expand_to_include(*args)¶SpatialReference¶SpatialReference(srs_input)¶Spatial reference objects are initialized on the given srs_input,
which may be one of the following:
'WGS84', 'WGS72',
'NAD27', 'NAD83')Example:
>>> wgs84 = SpatialReference('WGS84') # shorthand string
>>> wgs84 = SpatialReference(4326) # EPSG code
>>> wgs84 = SpatialReference('EPSG:4326') # EPSG string
>>> proj4 = '+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs '
>>> wgs84 = SpatialReference(proj4) # PROJ.4 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
__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:
>>> 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
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.
auth_name(target)¶Returns the authority name for the given string target node.
auth_code(target)¶Returns the authority code for the given string target node.
clone()¶Returns a clone of this spatial reference object.
identify_epsg()¶This method inspects the WKT of this SpatialReference and will add EPSG
authority nodes where an EPSG identifier is applicable.
from_esri()¶Morphs this SpatialReference from ESRI’s format to EPSG
to_esri()¶Morphs this SpatialReference to ESRI’s format.
validate()¶Checks to see if the given spatial reference is valid, if not an exception will be raised.
import_epsg(epsg)¶Import spatial reference from EPSG code.
import_proj(proj)¶Import spatial reference from PROJ.4 string.
import_user_input(user_input)¶import_wkt(wkt)¶Import spatial reference from WKT.
import_xml(xml)¶Import spatial reference from XML.
name¶Returns the name of this Spatial Reference.
srid¶Returns the SRID of top-level authority, or None if undefined.
linear_name¶Returns the name of the linear units.
linear_units¶Returns the value of the linear units.
angular_name¶Returns the name of the angular units.”
angular_units¶Returns the value of the angular units.
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.
ellipsoid¶Returns a tuple of the ellipsoid parameters for this spatial reference: (semimajor axis, semiminor axis, and inverse flattening).
semi_major¶Returns the semi major axis of the ellipsoid for this spatial reference.
semi_minor¶Returns the semi minor axis of the ellipsoid for this spatial reference.
inverse_flattening¶Returns the inverse flattening of the ellipsoid for this spatial reference.
geographic¶Returns True if this spatial reference is geographic (root node is
GEOGCS).
local¶Returns True if this spatial reference is local (root node is
LOCAL_CS).
projected¶Returns True if this spatial reference is a projected coordinate system
(root node is PROJCS).
wkt¶Returns the WKT representation of this spatial reference.
pretty_wkt¶Returns the ‘pretty’ representation of the WKT.
proj¶Returns the PROJ.4 representation for this spatial reference.
proj4¶Alias for SpatialReference.proj.
xml¶Returns the XML representation of this spatial reference.
CoordTransform¶CoordTransform(source, target)¶Represents a coordinate system transform. It is initialized with two
SpatialReference, representing the source and target coordinate
systems, respectively. These objects should be used when performing the same
coordinate transformation repeatedly on different geometries:
>>> ct = CoordTransform(SpatialReference('WGS84'), SpatialReference('NAD83'))
>>> for feat in layer:
... geom = feat.geom # getting clone of feature geometry
... geom.transform(ct) # transforming
GDALRaster¶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 simple, consistent interface. Each
data source is represented by a GDALRaster object which contains
one or more layers of data named bands. Each band, represented by a
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, 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.
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 representing a file path, 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 Creating rasters from data. For a detailed description of how to create rasters in the virtual filesystem, see Using GDAL’s Virtual Filesystem.
The following example shows how rasters can be created from different input sources (using the sample data from the GeoDjango tests; see also the 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'
Added the ability to read and write rasters in GDAL’s memory-based
virtual filesystem. GDALRaster objects can now be converted to and
from binary data in-memory.
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'
driver¶The name of the GDAL driver used to handle the input file. For GDALRasters created
from a file, the driver type is detected automatically. The creation of rasters from
scratch is a 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.
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'
width¶The width of the source in pixels (X-axis).
>>> GDALRaster({'width': 10, 'height': 20, 'srid': 4326}).width
10
height¶The height of the source in pixels (Y-axis).
>>> GDALRaster({'width': 10, 'height': 20, 'srid': 4326}).height
20
srs¶The spatial reference system of the raster, as a
SpatialReference instance. The SRS can be changed by
setting it to an other SpatialReference or providing any input
that is accepted by the SpatialReference constructor.
>>> rst = GDALRaster({'width': 10, 'height': 20, 'srid': 4326})
>>> rst.srs.srid
4326
>>> rst.srs = 3086
>>> rst.srs.srid
3086
srid¶The Spatial Reference System Identifier (SRID) of the raster. This
property is a shortcut to getting or setting the SRID through the
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
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 origin
(indices 0 and 3), scale (indices 1 and 5) and 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]
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]
scale¶Pixel width and height used for georeferencing the raster, as a as a
point object with x and y members. See 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]
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]
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)
bands¶List of all bands of the source, as 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)
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)
transform(srid, driver=None, name=None, resampling='NearestNeighbour', max_error=0.0)¶Returns a transformed version of this raster with the specified SRID.
This function transforms the current raster into a new spatial reference
system that can be specified with an srid. It calculates the bounds
and scale of the current raster in the new spatial reference system and
warps the raster using the 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 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 = rst.transform(4326)
>>> target.origin
[-82.98492744885776, 27.601924753080144]
info¶Returns a string with a summary of the raster. This is equivalent to the gdalinfo command line utility.
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'}}
vsi_buffer¶A bytes representation of this raster. Returns None for rasters
that are not stored in GDAL’s virtual filesystem.
is_vsi_based¶A boolean indicating if this raster is stored in GDAL’s virtual filesystem.
GDALBand¶GDALBand¶GDALBand instances are not created explicitly, but rather obtained
from a GDALRaster object, through its bands
attribute. The GDALBands contain the actual pixel values of the raster.
description¶The name or description of the band, if any.
width¶The width of the band in pixels (X-axis).
height¶The height of the band in pixels (Y-axis).
pixel_count¶The total number of pixels in this band. Is equal to width * height.
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
min, max, mean, and std properties.
min¶The minimum pixel value of the band (excluding the “no data” value).
max¶The maximum pixel value of the band (excluding the “no data” value).
mean¶The mean of all pixel values of the band (excluding the “no data” value).
std¶The standard deviation of all pixel values of the band (excluding the “no data” value).
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
(requires GDAL ≥ 2.1).
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.
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.
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)
metadata¶The metadata of this band. The functionality is identical to
GDALRaster.metadata.
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 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 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
GDALRaster and GDALBand classes.
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 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 srid attribute |
width |
required | Mapped to the width attribute |
height |
required | Mapped to the height attribute |
driver |
MEM |
Mapped to the driver attribute |
name |
'' |
See below |
origin |
0 |
Mapped to the origin attribute |
scale |
0 |
Mapped to the scale attribute |
skew |
0 |
Mapped to the width attribute |
bands |
[] |
See below |
nr_of_bands |
0 |
See below |
datatype |
6 |
See below |
papsz_options |
{} |
See below |
nameString 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.
datatypeInteger 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 |
nr_of_bandsInteger 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.
bandsA 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.
papsz_optionsA 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:
>>> 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,
... }
... })
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 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 nodata_value attribute |
data |
Same as nodata_value or 0 |
Passed to the data() method |
size |
(with, height) of raster |
Passed to the data() method |
shape |
Same as size | Passed to the data() method |
offset |
(0, 0) |
Passed to the data() method |
GDAL has an internal memory-based filesystem, which allows treating blocks of
memory as files. It can be used to read and write 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.
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:
# 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
Creating rasters from data). For virtual file-based rasters, the
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
HttpResponse:
>>> 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')
Oct 31, 2018