mapteksdk.data.images module

Image data objects.

Image objects are used to apply complicated textures to other objects.

class Raster(object_id=None, lock_type=LockType.READWRITE, width=None, height=None, *, image=None)

Bases: DataObject

Class representing raster images which can be draped onto other objects.

Topology objects which support raster association possess an associate_raster function which accepts a Raster and a RasterRegistration object which allows the raster to be draped onto that object.

Parameters

width (int) – The width of the raster.
height (int) – The height of the raster.
image (string, pathlib.Path or PIL.Image.Image) – The path to an image file. The image will be opened using pillow and used to populate the pixels, width and height of the raster. Alternatively, this can be a PIL.Image.Image object which will be used to populate the pixels, width and height of the raster. If this argument is specified also specifying the width and height arguments will raise an error.

Raises

TypeError – If image is not a string, pathlib.Path or Pil.Image.Image.
ValueError – If image is specified and width or height is specified.
FileNotFoundError – If image is the path to a non-existent file.
PIL.UnidentifiedImageError – If image is the path to a file which pillow does not recognise as an image.
PIL.DecompressionBombError – If image is the path to an image file which seems to be a “decompression bomb” - a maliciously crafted file intended to crash or otherwise disrupt the application when it is imported.

Notes

This object provides a consistent interface to the pixels of the raster image regardless of the underlying format. If the underlying format is JPEG or another format which does not support alpha, the alpha will always be read as 255 and any changes to the alpha components will be ignored.

See also

mapteksdk.data.facets.Surface.associate_raster: associate a raster with a surface.

Examples

Create a Raster sourcing the dimensions and pixels of the raster from a file called image.png.

>>> from mapteksdk.project import Project
>>> from mapteksdk.data import Raster
>>> project = Project()
>>> # Raster typically have a path of None because they are associated
>>> # with other objects.
>>> with project.new(None, Raster(image="path/to/image.jpg")) as raster:
...     # Perform operations on the raster here and associate it with
...     # an object.
...     pass

classmethod static_type()

Return the type of raster as stored in a Project.

This can be used for determining if the type of an object is a raster.

property id: ObjectID[Raster]

Object ID that uniquely references this object in the project.

Returns: The unique id of this object.
Return type: ObjectID

property width: The width of the raster. This is the number of pixels in each row.

property height: The height of the raster. This is the number of pixels in each column.

resize(new_width, new_height, resize_image=True)

Resizes the raster to the new width and height.

Parameters

new_width (int) – The new width for the raster. Pass None to keep the width unchanged.
new_height (int) – The new height for the raster. Pass None to keep the height unchanged.
resize_image (bool) – If True (default) The raster will be resized to fill the new size using a simple nearest neighbour search if the size is reduced, or simple bilinear interpolation. This will also change the format to JPEG (and hence the alpha component of the pixels will be discarded). If False, the current pixels are discarded and replaced with transparent white (or white if the format does not support transparency). This will not change the underlying format.

Raises

TypeError – If width or height cannot be converted to an integer.
ValueError – If width or height is less than one.
RuntimeError – If called when creating a new raster.

Warning

After calling resize with resize_image=True it is an error to access pixels or pixels_2d until the object is closed and reopened.

Examples

Halve the size of all rasters on an object. Note that because resize_image is true, the existing pixels will be changed to make a smaller version of the image.

>>> from mapteksdk.project import Project
>>> project = Project()
>>> with project.read("surfaces/target") as read_object:
>>>     for raster in read_object.rasters.values():
...         with project.edit(raster) as edit_raster:
...             edit_raster.resize(edit_raster.width // 2,
...                                edit_raster.height // 2,
...                                resize_image=True)

property pixel_count: The total number of pixels in the raster.

property pixels

The pixels of the raster object represented as a numpy array of shape (pixel_count, 4) where each row is the colour of a pixel in the form: [Red, Green, Blue, Alpha].

See pixels_2d for the pixels reshaped to match the width and height of the raster.

Raises: RuntimeError – If accessed after calling resize with resize_image = True.

Notes

The default colour is transparent white. This may appear as white if the raster is stored in a format which does not support alpha.

Examples

Accessing the pixels via this function is best when the two dimensional nature of the raster is not relevant or useful. The below example shows removing the green component from all of the pixels in an raster. Has no effect if the object at surfaces/target does not have an associated raster.

>>> from mapteksdk.project import Project
>>> project = Project()
>>> with project.read("surfaces/target") as read_object:
>>>     for raster in read_object.rasters.values():
...         with project.edit(raster) as edit_raster:
...             edit_raster.pixels[:, 1] = 0

property pixels_2d

The pixels reshaped to match the width and height of the raster. pixels_2d[0][0] is the colour of the pixel in the bottom left hand corner of the raster. pixels_2d[i][j] is the colour of the pixel i pixels to the right of the bottom left hand corner and j pixels above the bottom left hand corner.

The returned array will have shape (height, width, 4).

Raises

ValueError – If set using a string which cannot be converted to an integer.
ValueError – If set to a value which cannot be broadcast to the right shape.
TypeError – If set to a value which cannot be converted to an integer.

Notes

This returns the pixels in an ideal format to be passed to the raster.fromarray function in the 3rd party pillow library.

Examples

As pixels_2d allows access to the two dimensional nature of the raster, it can allow different transformations to be applied to different parts of the raster. The example below performs a different transformation to each quarter of the raster. Has no effect if the object at surfaces/target has no associated rasters.

>>> from mapteksdk.project import Project
>>> project = Project()
>>> with project.read("surfaces/target") as read_object:
...     for raster in read_object.rasters.values():
...         with project.edit(raster) as edit_raster:
...             width = edit_raster.width
...             height = edit_raster.height
...             half_width = edit_raster.width // 2
...             half_height = edit_raster.height // 2
...             # Remove red from the bottom left hand corner.
...             edit_raster.pixels_2d[0:half_height, 0:half_width, 0] = 0
...             # Remove green from the top left hand corner.
...             edit_raster.pixels_2d[half_height:height,
...                                   0:half_width, 1] = 0
...             # Remove blue from the bottom right hand corner.
...             edit_raster.pixels_2d[0:half_height,
...                                   half_width:width, 2] = 0
...             # Maximizes the red component in the top right hand corner.
...             edit_raster.pixels_2d[half_height:height,
...                                   half_width:width, 0] = 255

property title

The title of the raster.

This is shown in the manage images panel. Generally this is the name of the file the raster was imported from.

to_pillow()

Returns a Pillow.Image.Image object of the Raster.

Returns: A pillow image object with the same pixels as the raster.
Return type: PIL.Image.Image

Notes

Changing the returned object will not affect the Raster.

Examples

WARNING: These examples are incomplete and will not run by themselves. They assume a Project instance called project has been created and that raster_id has been set to the ObjectID of a raster object.

Save the pixels of the raster to a PNG file.

>>> with project.read(raster_id) as raster:
...     pillow_image = raster.to_pillow()
...     pillow_image.save(f"{raster.title}.png", "png")

Show the raster to the user in the standard image viewer.

>>> with project.read(raster_id) as raster:
...     pillow_image = raster.to_pillow()
...     pillow_image.show()

property registration

Returns the registration object which defines how the raster is draped onto Topology Objects.

If no raster registration is set, this will return a RasterRegistrationNone object. If raster registration is set, but the SDK does not support the type of registration, it will return a RasterRegistrationUnsupported object. Otherwise it will return a RasterRegistration subclass representing the existing registration.

Raises

TypeError – If set to a value which is not a RasterRegistration.
ValueError – If set to an invalid RasterRegistration.

Notes

You should not assign to this property directly. Instead pass the registration to the associate_raster() function of the object you would like to associate the raster with.

save(): Saves changes to the raster to the Project.

attribute_names()

Returns a list containing the names of all object-level attributes. Use this to iterate over the object attributes.

Returns: List containing the attribute names.
Return type: list

Examples

Iterate over all object attributes of the object stared at “target” and print their values.

>>> from mapteksdk.project import Project
>>> project = Project()
>>> with project.read("target") as read_object:
...     for name in read_object.attribute_names():
...         print(name, ":", read_object.get_attribute(name))

close()

Closes the object.

This should be called as soon as you are finished working with an object. To avoid needing to remember to call this function, open the object using a with block and project.read(), project.new() or project.edit(). Those functions automatically call this function at the end of the with block.

A closed object cannot be used for further reading or writing. The ID of a closed object may be queried and this can then be used to re-open the object.

property created_date: datetime

The date and time (in UTC) of when this object was created.

Returns: The date and time the object was created. 0:0:0 1/1/1970 if the operation failed.
Return type: datetime.datetime

delete_all_attributes()

Delete all object attributes attached to an object.

This only deletes object attributes and has no effect on PrimitiveAttributes.

Raises: RuntimeError – If all attributes cannot be deleted.

delete_attribute(attribute)

Deletes a single object-level attribute.

Deleting a non-existent object attribute will not raise an error.

Parameters: attribute (str) – Name of attribute to delete.
Returns: True if the object attribute existed and was deleted; False if the object attribute did not exist.
Return type: bool
Raises: RuntimeError – If the attribute cannot be deleted.

get_attribute(name)

Returns the value for the attribute with the specified name.

Parameters: name (str) – The name of the object attribute to get the value for.
Returns: The value of the object attribute name. For dtype = datetime.datetime this is an integer representing the number of milliseconds since 1st Jan 1970. For dtype = datetime.date this is a tuple of the form: (year, month, day).
Return type: ObjectAttributeTypes
Raises: KeyError – If there is no object attribute called name.

Warning

In the future this function may be changed to return datetime.datetime and datetime.date objects instead of the current representation for object attributes of type datetime.datetime or datetime.date.

get_attribute_type(name)

Returns the type of the attribute with the specified name.

Parameters: name (str) – Name of the attribute whose type should be returned.
Returns: The type of the object attribute name.
Return type: ObjectAttributeDataTypes
Raises: KeyError – If there is no object attribute called name.

property lock_type: LockType

Indicates whether operating in read-only or read-write mode.

Returns: The type of lock on this object. This will be LockType.ReadWrite if the object is open for editing and LockType.Read if the object is open for reading.
Return type: LockType

property modified_date: datetime

The date and time (in UTC) of when this object was last modified.

Returns: The date and time this object was last modified. 0:0:0 1/1/1970 if the operation failed.
Return type: datetime.datetime

set_attribute(name, dtype, data)

Sets the value for the object attribute with the specified name.

This will overwrite any existing attribute with the specified name.

Parameters

name – The name of the object attribute for which the value should be set.
dtype – The type of data to assign to the attribute. This should be a type from the ctypes module or datetime.datetime or datetime.date. Passing bool is equivalent to passing ctypes.c_bool. Passing str is equivalent to passing ctypes.c_char_p. Passing int is equivalent to passing ctypes.c_int16. Passing float is equivalent to passing ctypes.c_double.
data – The value to assign to object attribute name. For dtype = datetime.datetime this can either be a datetime object or timestamp which will be passed directly to datetime.utcfromtimestamp(). For dtype = datetime.date this can either be a date object or a tuple of the form: (year, month, day).

Raises

ValueError – If dtype is an unsupported type.
TypeError – If value is an inappropriate type for object attribute name.
RuntimeError – If a different error occurs.

Warning

Object attributes are saved separately from the object itself - any changes made by this function (assuming it does not raise an error) will be saved even if save() is not called (for example, due to an error being raised by another function).

Examples

Create an object attribute on an object at “target” and then read its value.

>>> import ctypes
>>> from mapteksdk.project import Project
>>> project = Project()
>>> with project.edit("target") as edit_object:
...     edit_object.set_attribute("count", ctypes.c_int16, 0)
... with project.read("target") as read_object:
...     print(read_object.get_attribute("count"))
0

class RasterRegistration

Bases: object

Base class for all types of raster registration. This is useful for type checking.

property is_valid: True if the object is valid.

raise_if_invalid()

Checks if the raster is invalid and raises a ValueError if any invalid information is detected.

If is_valid is False then calling this function will raise a ValueError containing information on why the registration is considered invalid.

Raises: ValueError – If the raster is invalid.

class RasterRegistrationNone

Bases: RasterRegistration

Class representing no raster registration is present.

Notes

This is always considered valid, so raise_if_valid will never raise an error for this object.

property is_valid: True if the object is valid.

raise_if_invalid()

Checks if the raster is invalid and raises a ValueError if any invalid information is detected.

If is_valid is False then calling this function will raise a ValueError containing information on why the registration is considered invalid.

Raises: ValueError – If the raster is invalid.

class RasterRegistrationUnsupported

Bases: RasterRegistration

Class representing a raster registration which is not supported by the SDK.

If you would like an unsupported registration method to be supported then use request support.

Notes

This is always considered invalid so raise_if_valid will always raise an error.

property is_valid: True if the object is valid.

raise_if_invalid()

Checks if the raster is invalid and raises a ValueError if any invalid information is detected.

If is_valid is False then calling this function will raise a ValueError containing information on why the registration is considered invalid.

Raises: ValueError – If the raster is invalid.

class PointPairRegistrationBase

Bases: RasterRegistration

Base class for registration objects which use image/world point pairs to register the image to a Surface.

classmethod minimum_point_pairs()

The minimum number of world / image point pairs required for this type of registration.

Returns: The minimum number of world / image point pairs required.
Return type: int

property is_valid: True if the object is valid.

raise_if_invalid()

Checks if the raster is invalid and raises a ValueError if any invalid information is detected.

If is_valid is False then calling this function will raise a ValueError containing information on why the registration is considered invalid.

Raises: ValueError – If the raster is invalid.

property image_points

The points on the image used to map the raster onto an object. This is a numpy array of points in image coordinates where [0, 0] is the bottom left hand corner of the image and [width, height] is the top right hand corner of the image.

Each of these points should match one of the world points. If the raster is mapped onto an object, the pixel at image_points[i] will be placed at world_points[i] on the surface.

Raises

ValueError – If set to a value which cannot be converted to a two dimensional array containing two dimensional points or if any value in the array cannot be converted to a floating point number.
TypeError – If set to a value which cannot be converted to a numpy array.

property world_points

The world points used to map the raster onto an object. This is a numpy array of points in world coordinates.

Each of these points should match one of the image points. If the raster is mapped onto an object, the pixel at image_points[i] will be placed at world_points[i] on the surface.

Raises

ValueError – If set to a value which cannot be converted to a two dimensional array containing three dimensional points or if any value in the array cannot be converted to a floating point number.
TypeError – If set to a value which cannot be converted to a numpy array.

class RasterRegistrationTwoPoint(image_points=None, world_points=None, orientation=None)

Bases: PointPairRegistrationBase

Represents a simple raster registration which uses two points and an orientation to project a raster onto an object (typically a Surface).

Parameters

image_points (numpy.ndarray) – The image points to assign to the object. See the property for more details.
world_points (numpy.ndarray) – The world points to assign to the object. See the property for more details.
orientation (numpy.ndarray) – The orientation to assign to the object. See the property for more details.