# Copyright Crown and Cartopy Contributors
#
# This file is part of Cartopy and is released under the BSD 3-clause license.
# See LICENSE in the root of the repository for full licensing details.
"""
Provides an interface for representing images.
"""
import collections
from pathlib import Path
import numpy as np
from PIL import Image
import shapely.geometry as sgeom
_img_class_attrs = ['filename', 'extent', 'origin', 'pixel_size']
[docs]
class Img(collections.namedtuple('Img', _img_class_attrs)):
def __new__(cls, *args, **kwargs):
# ensure any lists given as args or kwargs are turned into tuples.
new_args = []
for item in args:
if isinstance(item, list):
item = tuple(item)
new_args.append(item)
new_kwargs = {}
for k, item in kwargs.items():
if isinstance(item, list):
item = tuple(item)
new_kwargs[k] = item
return super().__new__(cls, *new_args, **new_kwargs)
def __init__(self, *args, **kwargs):
"""
Represents a simple geo-located image.
Parameters
----------
filename
Filename of the image tile.
extent
The (x_lower, x_upper, y_lower, y_upper) extent of the
image in units of the native projection.
origin
Name of the origin.
pixel_size
The (x_scale, y_scale) pixel width, in units of the
native projection per pixel.
Note
----
API is likely to change in the future to include a CRS.
"""
self._bbox = None
def __getstate__(self):
"""
Override the default to ensure when pickling that any new attributes
introduced are included in the pickled object.
"""
return self.__dict__
[docs]
def bbox(self):
"""
Return a :class:`~shapely.geometry.polygon.Polygon` instance for
this image's extents.
"""
if self._bbox is None:
x0, x1, y0, y1 = self.extent
self._bbox = sgeom.box(x0, y0, x1, y1)
return self._bbox
[docs]
@staticmethod
def world_files(fname):
"""
Determine potential world filename combinations, without checking
their existence.
For example, a ``'*.tif'`` file may have one of the following
popular conventions for world file extensions ``'*.tifw'``,
``'*.tfw'``, ``'*.TIFW'`` or ``'*.TFW'``.
Given the possible world file extensions, the upper case basename
combinations are also generated. For example, the file 'map.tif'
will generate the following world file variations, 'map.tifw',
'map.tfw', 'map.TIFW', 'map.TFW', 'MAP.tifw', 'MAP.tfw', 'MAP.TIFW'
and 'MAP.TFW'.
Parameters
----------
fname
Name of the file for which to get all the possible world
filename combinations.
Returns
-------
A list of possible world filename combinations.
Examples
--------
>>> from cartopy.io.img_nest import Img
>>> Img.world_files('img.png')[:6]
['img.pngw', 'img.pgw', 'img.PNGW', 'img.PGW', 'IMG.pngw', 'IMG.pgw']
>>> Img.world_files('/path/to/img.TIF')[:2]
['/path/to/img.tifw', '/path/to/img.tfw']
>>> Img.world_files('/path/to/img/with_no_extension')[0]
'/path/to/img/with_no_extension.w'
"""
path = Path(fname)
fext = path.suffix[1::].lower()
# If there was no extension to the filename.
if len(fext) < 3:
result = [path.with_suffix('.w'), path.with_suffix('.W')]
else:
fext_types = [f'.{fext}w', f'.{fext[0]}{fext[-1]}w']
fext_types.extend([ext.upper() for ext in fext_types])
result = [path.with_suffix(ext) for ext in fext_types]
result += [p.with_name(p.stem.swapcase() + p.suffix) for p in result]
# return string paths rather than Path objects
return [str(r) for r in result]
def __array__(self):
return np.array(Image.open(self.filename))
[docs]
@classmethod
def from_world_file(cls, img_fname, world_fname):
"""
Return an Img instance from the given image filename and
worldfile filename.
"""
im = Image.open(img_fname)
with open(world_fname) as world_fh:
extent, pix_size = cls.world_file_extent(world_fh, im.size)
if hasattr(im, 'close'):
im.close()
return cls(img_fname, extent, 'lower', pix_size)
[docs]
@staticmethod
def world_file_extent(worldfile_handle, im_shape):
"""
Return the extent ``(x0, x1, y0, y1)`` and pixel size
``(x_width, y_width)`` as defined in the given worldfile file handle
and associated image shape ``(x, y)``.
"""
lines = worldfile_handle.readlines()
if len(lines) != 6:
raise ValueError('Only world files with 6 lines are supported.')
pix_size = (float(lines[0]), float(lines[3]))
pix_rotation = (float(lines[1]), float(lines[2]))
if pix_rotation != (0., 0.):
raise ValueError('Rotated pixels in world files is not currently '
'supported.')
ul_corner = (float(lines[4]), float(lines[5]))
min_x, max_x = (ul_corner[0] - pix_size[0] / 2.,
ul_corner[0] + pix_size[0] * im_shape[0] -
pix_size[0] / 2.)
min_y, max_y = (ul_corner[1] - pix_size[1] / 2.,
ul_corner[1] + pix_size[1] * im_shape[1] -
pix_size[1] / 2.)
return (min_x, max_x, min_y, max_y), pix_size
[docs]
class ImageCollection:
def __init__(self, name, crs, images=None):
"""
Represents a collection of images at the same logical level.
Typically these are images at the same zoom level or resolution.
Parameters
----------
name
The name of the image collection.
crs
The :class:`~cartopy.crs.Projection` instance.
images: optional
A list of one or more :class:`~cartopy.io.img_nest.Img`
instances. Defaults to None.
"""
self.name = name
self.crs = crs
self.images = images or []
[docs]
def scan_dir_for_imgs(self, directory, glob_pattern='*.tif',
img_class=Img):
"""
Search the given directory for the associated world files
of the image files.
Parameters
----------
directory
The directory path to search for image files.
glob_pattern: optional
The image filename glob pattern to search with.
Defaults to ``'*.tif'``.
img_class: optional
The class used to construct each image in the Collection.
Note
----
Does not recursively search sub-directories.
"""
imgs = Path(directory).glob(glob_pattern)
for img in imgs:
dirname, fname = img.parent, img.name
worlds = img_class.world_files(fname)
for fworld in worlds:
fworld = dirname / fworld
if fworld.exists():
break
else:
raise ValueError(
f'Image file {img!r} has no associated world file')
self.images.append(img_class.from_world_file(img, fworld))
[docs]
class NestedImageCollection:
def __init__(self, name, crs, collections, _ancestry=None):
"""
Represents a complex nest of ImageCollections.
On construction, the image collections are scanned for ancestry,
leading to fast image finding capabilities.
A complex (and time consuming to create) NestedImageCollection instance
can be saved as a pickle file and subsequently be (quickly) restored.
There is a simplified creation interface for NestedImageCollection
``from_configuration`` for more detail.
Parameters
----------
name
The name of the nested image collection.
crs
The native :class:`~cartopy.crs.Projection` of all the image
collections.
collections
A list of one or more :class:`~cartopy.io.img_nest.ImageCollection`
instances.
"""
# NOTE: all collections must have the same crs.
_names = {collection.name for collection in collections}
assert len(_names) == len(collections), \
'The collections must have unique names.'
self.name = name
self.crs = crs
self._collections_by_name = {collection.name: collection
for collection in collections}
def sort_func(c):
return np.max([image.bbox().area for image in c.images])
self._collections = sorted(collections, key=sort_func, reverse=True)
self._ancestry = {}
"""
maps (collection name, image) to a list of children
(collection name, image).
"""
if _ancestry is not None:
self._ancestry = _ancestry
else:
parent_wth_children = zip(self._collections,
self._collections[1:])
for parent_collection, collection in parent_wth_children:
for parent_image in parent_collection.images:
for image in collection.images:
if self._is_parent(parent_image, image):
# append the child image to the parent's ancestry
key = (parent_collection.name, parent_image)
self._ancestry.setdefault(key, []).append(
(collection.name, image))
# TODO check that the ancestry is in a good state (i.e. that each
# collection has child images)
@staticmethod
def _is_parent(parent, child):
"""
Return whether the given Image is the parent of image.
Used by __init__.
"""
result = False
pbox = parent.bbox()
cbox = child.bbox()
if pbox.area > cbox.area:
result = pbox.intersects(cbox) and not pbox.touches(cbox)
return result
[docs]
def image_for_domain(self, target_domain, target_z):
"""
Determine the image that provides complete coverage of target
location.
The composed image is merged from one or more image tiles that overlay
the target location and provide complete image coverage of the target
location.
Parameters
----------
target_domain
A :class:`~shapely.geometry.linestring.LineString`
instance that specifies the target location requiring image
coverage.
target_z
The name of the target :class`~cartopy.io.img_nest.ImageCollection`
which specifies the target zoom level (resolution) of the required
images.
Returns
-------
img, extent, origin
A tuple containing three items, consisting of the target
location :class:`numpy.ndarray` image data, the
(x-lower, x-upper, y-lower, y-upper) extent of the image, and the
origin for the target location.
"""
# XXX Copied from cartopy.io.img_tiles
if target_z not in self._collections_by_name:
# TODO: Handle integer depths also?
raise ValueError(
f'{target_z!r} is not one of the possible collections.')
tiles = []
for tile in self.find_images(target_domain, target_z):
try:
img, extent, origin = self.get_image(tile)
except OSError:
continue
img = np.array(img)
x = np.linspace(extent[0], extent[1], img.shape[1],
endpoint=False)
y = np.linspace(extent[2], extent[3], img.shape[0],
endpoint=False)
tiles.append([np.array(img), x, y, origin])
from cartopy.io.img_tiles import _merge_tiles
img, extent, origin = _merge_tiles(tiles)
return img, extent, origin
[docs]
def find_images(self, target_domain, target_z, start_tiles=None):
"""
A generator that finds all images that overlap the bounded
target location.
Parameters
----------
target_domain
A :class:`~shapely.geometry.linestring.LineString` instance that
specifies the target location requiring image coverage.
target_z
The name of the target
:class:`~cartopy.io.img_nest.ImageCollection` which specifies
the target zoom level (resolution) of the required images.
start_tiles: optional
A list of one or more tuple pairs, composed of a
:class:`~cartopy.io.img_nest.ImageCollection` name and an
:class:`~cartopy.io.img_nest.Img` instance, from which to search
for the target images.
Returns
-------
generator
A generator tuple pair composed of a
:class:`~cartopy.io.img_nest.ImageCollection` name and an
:class:`~cartopy.io.img_nest.Img` instance.
"""
# XXX Copied from cartopy.io.img_tiles
if target_z not in self._collections_by_name:
# TODO: Handle integer depths also?
raise ValueError(
f'{target_z!r} is not one of the possible collections.')
if start_tiles is None:
start_tiles = ((self._collections[0].name, img)
for img in self._collections[0].images)
for start_tile in start_tiles:
# recursively drill down to the images at the target zoom
domain = start_tile[1].bbox()
if target_domain.intersects(domain) and \
not target_domain.touches(domain):
if start_tile[0] == target_z:
yield start_tile
else:
for tile in self.subtiles(start_tile):
yield from self.find_images(target_domain,
target_z,
start_tiles=[tile])
[docs]
def subtiles(self, collection_image):
"""
Find the higher resolution image tiles that compose this parent
image tile.
Parameters
----------
collection_image
A tuple pair containing the parent
:class:`~cartopy.io.img_nest.ImageCollection` name and
:class:`~cartopy.io.img_nest.Img` instance.
Returns
-------
iterator
An iterator of tuple pairs containing the higher resolution child
:class:`~cartopy.io.img_nest.ImageCollection` name and
:class:`~cartopy.io.img_nest.Img` instance that compose the parent.
"""
return iter(self._ancestry.get(collection_image, []))
desired_tile_form = 'RGB'
[docs]
def get_image(self, collection_image):
"""
Retrieve the data of the target image from file.
Parameters
----------
collection_image:
A tuple pair containing the target
:class:`~cartopy.io.img_nest.ImageCollection` name and
:class:`~cartopy.io.img_nest.Img` instance.
Returns
-------
img_data, img.extent, img.origin
A tuple containing three items, consisting of the associated image
file data, the (x_lower, x_upper, y_lower, y_upper) extent of the
image, and the image origin.
Note
----
The format of the retrieved image file data is controlled by
:attr:`~cartopy.io.img_nest.NestedImageCollection.desired_tile_form`,
which defaults to 'RGB' format.
"""
img = collection_image[1]
img_data = Image.open(img.filename)
img_data = img_data.convert(self.desired_tile_form)
return img_data, img.extent, img.origin
[docs]
@classmethod
def from_configuration(cls, name, crs, name_dir_pairs,
glob_pattern='*.tif',
img_class=Img):
"""
Create a :class:`~cartopy.io.img_nest.NestedImageCollection` instance
given the list of image collection name and directory path pairs.
This is very convenient functionality for simple configuration level
creation of this complex object.
For example, to produce a nested collection of OS map tiles::
files = [['OS 1:1,000,000', '/directory/to/1_to_1m'],
['OS 1:250,000', '/directory/to/1_to_250k'],
['OS 1:50,000', '/directory/to/1_to_50k'],
]
r = NestedImageCollection.from_configuration('os',
ccrs.OSGB(),
files)
Parameters
----------
name
The name for the
:class:`~cartopy.io.img_nest.NestedImageCollection` instance.
crs
The :class:`~cartopy.crs.Projection` of the image collection.
name_dir_pairs
A list of image collection name and directory path pairs.
glob_pattern: optional
The image collection filename glob pattern. Defaults
to ``'*.tif'``.
img_class: optional
The class of images created in the image collection.
Returns
-------
A :class:`~cartopy.io.img_nest.NestedImageCollection` instance.
Warnings
--------
The list of image collection name and directory path pairs must be
given in increasing resolution order i.e. from low resolution to
high resolution.
"""
collections = []
for collection_name, collection_dir in name_dir_pairs:
collection = ImageCollection(collection_name, crs)
collection.scan_dir_for_imgs(collection_dir,
glob_pattern=glob_pattern,
img_class=img_class)
collections.append(collection)
return cls(name, crs, collections)
