def populate_mime_data_for_drag(self, mime_data, size: Geometry.IntSize):
if self.__display_item:
mime_data.set_data_as_string(MimeTypes.DISPLAY_ITEM_MIME_TYPE, str(self.__display_item.uuid))
rgba_image_data = self.__thumbnail_source.thumbnail_data
thumbnail = Image.get_rgba_data_from_rgba(Image.scaled(Image.get_rgba_view_from_rgba_data(rgba_image_data), (size.width, size.height))) if rgba_image_data is not None else None
return True, thumbnail
return False, None
def function_rgb_channel(data_and_metadata_in: _DataAndMetadataLike,
channel: int) -> DataAndMetadata.DataAndMetadata:
data_and_metadata = DataAndMetadata.promote_ndarray(data_and_metadata_in)
if channel < 0 or channel > 3:
raise ValueError("RGB channel: invalid channel.")
data = data_and_metadata.data
if not Image.is_data_valid(data):
raise ValueError("RGB channel: invalid data.")
if not data_and_metadata.is_data_rgb_type:
raise ValueError("RGB channel: data is not RGB type.")
assert data is not None
channel_data: _ImageDataType
if Image.is_shape_and_dtype_rgb(data.shape, data.dtype):
if channel == 3:
channel_data = numpy.ones(data.shape, int)
else:
channel_data = data[..., channel].astype(int)
elif Image.is_shape_and_dtype_rgba(data.shape, data.dtype):
channel_data = data[..., channel].astype(int)
else:
raise ValueError("RGB channel: unable to extract channel.")
return DataAndMetadata.new_data_and_metadata(
channel_data,
intensity_calibration=data_and_metadata.intensity_calibration,
dimensional_calibrations=data_and_metadata.dimensional_calibrations)
def function_rgb_linear_combine(
data_and_metadata_in: _DataAndMetadataLike, red_weight: float,
green_weight: float,
blue_weight: float) -> DataAndMetadata.DataAndMetadata:
data_and_metadata = DataAndMetadata.promote_ndarray(data_and_metadata_in)
data = data_and_metadata.data
if not Image.is_data_valid(data):
raise ValueError("RGB linear combine: invalid data.")
if not data_and_metadata.is_data_rgb_type:
raise ValueError("RGB linear combine: data is not RGB type.")
assert data is not None
combined_data: _ImageDataType
if Image.is_shape_and_dtype_rgb(data.shape, data.dtype):
combined_data = numpy.sum(
data[..., :] * (blue_weight, green_weight, red_weight), 2)
elif Image.is_shape_and_dtype_rgba(data.shape, data.dtype):
combined_data = numpy.sum(
data[..., :] * (blue_weight, green_weight, red_weight, 0.0), 2)
else:
raise ValueError("RGB channel: unable to extract channel.")
return DataAndMetadata.new_data_and_metadata(
combined_data,
intensity_calibration=data_and_metadata.intensity_calibration,
dimensional_calibrations=data_and_metadata.dimensional_calibrations)
def test_create_rgba_image_from_array(self) -> None:
image_1d_16 = numpy.zeros((16, ), dtype=numpy.double)
image_1d_16x1 = numpy.zeros((16, 1), dtype=numpy.double)
self.assertIsNotNone(Image.create_rgba_image_from_array(image_1d_16))
self.assertIsNotNone(Image.create_rgba_image_from_array(image_1d_16x1))
image_1d_rgb = numpy.zeros((16, 3), dtype=numpy.uint8)
self.assertIsNotNone(Image.create_rgba_image_from_array(image_1d_rgb))
def _repaint(self, drawing_context: DrawingContext.DrawingContext) -> None:
"""Repaint the canvas item. This will occur on a thread."""
canvas_size = self.canvas_size
if canvas_size:
with drawing_context.saver():
if self.__color_map_data is not None:
rgba_image: numpy.typing.NDArray[numpy.uint32] = numpy.empty((4,) + self.__color_map_data.shape[:-1], dtype=numpy.uint32)
Image.get_rgb_view(rgba_image)[:] = self.__color_map_data[numpy.newaxis, :, :] # scalar data assigned to each component of rgb view
Image.get_alpha_view(rgba_image)[:] = 255
drawing_context.draw_image(rgba_image, 0, 0, canvas_size.width, canvas_size.height)
def drag_started(self, ui: UserInterface.UserInterface, x: int, y: int, modifiers: UserInterface.KeyboardModifiers) -> typing.Tuple[typing.Optional[UserInterface.MimeData], typing.Optional[numpy.ndarray]]:
if self.__display_item:
mime_data = self.ui.create_mime_data()
if self.__display_item:
MimeTypes.mime_data_put_display_item(mime_data, self.__display_item)
thumbnail_data = self.calculate_thumbnail_data()
if thumbnail_data is not None:
# scaling is very slow
thumbnail_data = Image.get_rgba_data_from_rgba(Image.scaled(Image.get_rgba_view_from_rgba_data(thumbnail_data), Geometry.IntSize(w=80, h=80)))
return mime_data, thumbnail_data
return None, None
def calculate_data():
data = data_and_metadata.data
if not Image.is_data_valid(data):
return None
if Image.is_shape_and_dtype_rgb(data.shape, data.dtype):
return numpy.sum(
data[..., :] * (blue_weight, green_weight, red_weight), 2)
elif Image.is_shape_and_dtype_rgba(data.shape, data.dtype):
return numpy.sum(
data[..., :] * (blue_weight, green_weight, red_weight, 0.0), 2)
else:
return None
def populate_mime_data_for_drag(self, mime_data: UserInterface.MimeData,
size: Geometry.IntSize):
if self.__display_item:
MimeTypes.mime_data_put_display_item(mime_data,
self.__display_item)
rgba_image_data = self.__thumbnail_source.thumbnail_data
thumbnail = Image.get_rgba_data_from_rgba(
Image.scaled(
Image.get_rgba_view_from_rgba_data(rgba_image_data),
(size.width,
size.height))) if rgba_image_data is not None else None
return True, thumbnail
return False, None
def populate_mime_data_for_drag(
self, mime_data: UserInterface.MimeData, size: Geometry.IntSize
) -> typing.Tuple[bool, typing.Optional[_NDArray]]:
if self.__display_item:
MimeTypes.mime_data_put_display_item(mime_data,
self.__display_item)
rgba_image_data = self.__thumbnail_source.thumbnail_data if self.__thumbnail_source else None
thumbnail = Image.get_rgba_data_from_rgba(
Image.scaled(
Image.get_rgba_view_from_rgba_data(rgba_image_data),
(80, 80))) if rgba_image_data is not None else None
return True, thumbnail
return False, None
def test_rgba_can_be_created_from_h5py_array(self) -> None:
current_working_directory = os.getcwd()
workspace_dir = os.path.join(current_working_directory, "__Test")
if os.path.exists(workspace_dir):
shutil.rmtree(workspace_dir)
os.makedirs(workspace_dir)
try:
with h5py.File(os.path.join(workspace_dir, "file.h5"), "w") as f:
dataset = f.create_dataset("data", data=numpy.ones((4, 4, 4), dtype=numpy.uint8))
Image.create_rgba_image_from_array(dataset)
finally:
# print(f"rmtree {workspace_dir}")
shutil.rmtree(workspace_dir)
def calculate_data():
data = data_and_metadata.data
if channel < 0 or channel > 3:
return None
if not Image.is_data_valid(data):
return None
if Image.is_shape_and_dtype_rgb(data.shape, data.dtype):
if channel == 3:
return numpy.ones(data.shape, int)
return data[..., channel].astype(int)
elif Image.is_shape_and_dtype_rgba(data.shape, data.dtype):
return data[..., channel].astype(int)
else:
return None
def test_rebin_expand_has_even_expansion(self) -> None:
# NOTE: statistical tests are only valid if expanded length is multiple of src length
src = numpy.arange(0, 10)
expanded = Image.rebin_1d(src, 50)
self.assertAlmostEqual(numpy.mean(src), numpy.mean(expanded))
self.assertAlmostEqual(numpy.var(src), numpy.var(expanded))
src = numpy.arange(0, 10)
expanded = Image.rebin_1d(src, 500)
self.assertAlmostEqual(numpy.mean(src), numpy.mean(expanded))
self.assertAlmostEqual(numpy.var(src), numpy.var(expanded))
# test larger values to make sure linear mapping works (failed once)
src = numpy.arange(0, 200)
expanded = Image.rebin_1d(src, 600)
self.assertAlmostEqual(numpy.mean(src), numpy.mean(expanded))
self.assertAlmostEqual(numpy.var(src), numpy.var(expanded))
def _repaint(self, drawing_context):
"""Repaint the canvas item. This will occur on a thread."""
# canvas size
canvas_width = self.canvas_size[1]
canvas_height = self.canvas_size[0]
# draw background
if self.background_color:
with drawing_context.saver():
drawing_context.begin_path()
drawing_context.move_to(0, 0)
drawing_context.line_to(canvas_width, 0)
drawing_context.line_to(canvas_width, canvas_height)
drawing_context.line_to(0, canvas_height)
drawing_context.close_path()
drawing_context.fill_style = self.background_color
drawing_context.fill()
# draw the data, if any
if (self.data is not None and len(self.data) > 0):
# draw the histogram itself
with drawing_context.saver():
drawing_context.begin_path()
binned_data = Image.rebin_1d(
self.data, int(canvas_width), self.__retained_rebin_1d
) if int(canvas_width) != self.data.shape[0] else self.data
for i in range(canvas_width):
drawing_context.move_to(i, canvas_height)
drawing_context.line_to(
i, canvas_height * (1 - binned_data[i]))
drawing_context.line_width = 1
drawing_context.stroke_style = "#444"
drawing_context.stroke()
def read_image_from_file(filename: pathlib.Path) -> _DataArrayType:
if str(filename).startswith(":"):
return numpy.zeros((20, 20, 4), numpy.uint8)
image = imageio.imread(filename)
if image is not None:
image_u8 = imageio.core.image_as_uint(image)
if len(image_u8.shape) == 3:
rgba_image: numpy.typing.NDArray[numpy.uint8]
if image_u8.shape[-1] == 3:
rgba_image = numpy.empty(image_u8.shape[:-1] + (4, ),
numpy.uint8)
rgba_image[..., 0] = image_u8[..., 2]
rgba_image[..., 1] = image_u8[..., 1]
rgba_image[..., 2] = image_u8[..., 0]
rgba_image[..., 3] = 255
else:
assert image_u8.shape[-1] == 4
rgba_image = numpy.empty(image_u8.shape[:-1] + (4, ),
numpy.uint8)
rgba_image[..., 0] = image_u8[..., 3]
rgba_image[..., 1] = image_u8[..., 2]
rgba_image[..., 2] = image_u8[..., 1]
rgba_image[..., 3] = image_u8[..., 0]
if is_grayscale(rgba_image):
rgba_image = Image.convert_to_grayscale(rgba_image)
else:
assert len(image_u8.shape) == 2
rgba_image = image_u8
assert rgba_image is not None
return rgba_image
raise IOError()
def function_rgba(
red_data_and_metadata_in: _DataAndMetadataIndeterminateSizeLike,
green_data_and_metadata_in: _DataAndMetadataIndeterminateSizeLike,
blue_data_and_metadata_in: _DataAndMetadataIndeterminateSizeLike,
alpha_data_and_metadata_in: _DataAndMetadataIndeterminateSizeLike
) -> DataAndMetadata.DataAndMetadata:
red_data_and_metadata_c = DataAndMetadata.promote_indeterminate_array(
red_data_and_metadata_in)
green_data_and_metadata_c = DataAndMetadata.promote_indeterminate_array(
green_data_and_metadata_in)
blue_data_and_metadata_c = DataAndMetadata.promote_indeterminate_array(
blue_data_and_metadata_in)
alpha_data_and_metadata_c = DataAndMetadata.promote_indeterminate_array(
alpha_data_and_metadata_in)
shape = DataAndMetadata.determine_shape(red_data_and_metadata_c,
green_data_and_metadata_c,
blue_data_and_metadata_c)
if shape is None:
raise ValueError("RGBA: data shapes do not match or are indeterminate")
red_data_and_metadata = DataAndMetadata.promote_constant(
red_data_and_metadata_c, shape)
green_data_and_metadata = DataAndMetadata.promote_constant(
green_data_and_metadata_c, shape)
blue_data_and_metadata = DataAndMetadata.promote_constant(
blue_data_and_metadata_c, shape)
alpha_data_and_metadata = DataAndMetadata.promote_constant(
alpha_data_and_metadata_c, shape)
channels = (blue_data_and_metadata, green_data_and_metadata,
red_data_and_metadata, alpha_data_and_metadata)
if any([
not Image.is_data_valid(data_and_metadata.data)
for data_and_metadata in channels
]):
raise ValueError("RGB: invalid data")
rgba_image = numpy.empty(shape + (4, ), numpy.uint8)
for channel_index, channel in enumerate(channels):
data = channel._data_ex
if data.dtype.kind in 'iu':
rgba_image[..., channel_index] = numpy.clip(data, 0, 255)
elif data.dtype.kind in 'f':
rgba_image[...,
channel_index] = numpy.clip(numpy.multiply(data, 255),
0, 255)
return DataAndMetadata.new_data_and_metadata(
rgba_image,
intensity_calibration=red_data_and_metadata.intensity_calibration,
dimensional_calibrations=red_data_and_metadata.dimensional_calibrations
)
def test_scale_linear_is_symmetry(self) -> None:
src1 = numpy.zeros((8, 8))
src2 = numpy.zeros((9, 9))
src1[3:5, 3:5] = 1
src2[3:6, 3:6] = 1
src1s: _ImageDataType = typing.cast(_ImageDataType, Image.scaled(src1, (12, 12), 'linear')*1000).astype(numpy.int32)
src2s: _ImageDataType = typing.cast(_ImageDataType, Image.scaled(src1, (12, 12), 'linear')*1000).astype(numpy.int32)
src1t: _ImageDataType = typing.cast(_ImageDataType, Image.scaled(src1, (13, 13), 'linear')*1000).astype(numpy.int32)
src2t: _ImageDataType = typing.cast(_ImageDataType, Image.scaled(src1, (13, 13), 'linear')*1000).astype(numpy.int32)
self.assertTrue(numpy.array_equal(src1s[0:6, 0:6], src1s[0:6, 12:5:-1]))
self.assertTrue(numpy.array_equal(src1s[0:6, 0:6], src1s[12:5:-1, 12:5:-1]))
self.assertTrue(numpy.array_equal(src1s[0:6, 0:6], src1s[12:5:-1, 0:6]))
self.assertTrue(numpy.array_equal(src2s[0:6, 0:6], src2s[0:6, 12:5:-1]))
self.assertTrue(numpy.array_equal(src2s[0:6, 0:6], src2s[12:5:-1, 12:5:-1]))
self.assertTrue(numpy.array_equal(src2s[0:6, 0:6], src2s[12:5:-1, 0:6]))
self.assertTrue(numpy.array_equal(src1t[0:6, 0:6], src1t[0:6, 13:6:-1]))
self.assertTrue(numpy.array_equal(src1t[0:6, 0:6], src1t[13:6:-1, 13:6:-1]))
self.assertTrue(numpy.array_equal(src1t[0:6, 0:6], src1t[13:6:-1, 0:6]))
self.assertTrue(numpy.array_equal(src2t[0:6, 0:6], src2t[0:6, 13:6:-1]))
self.assertTrue(numpy.array_equal(src2t[0:6, 0:6], src2t[13:6:-1, 13:6:-1]))
self.assertTrue(numpy.array_equal(src2t[0:6, 0:6], src2t[13:6:-1, 0:6]))
def plot_features(self, data: _NDArray, offset_m: Geometry.FloatPoint,
fov_size_nm: Geometry.FloatSize,
extra_nm: Geometry.FloatPoint,
center_nm: Geometry.FloatPoint,
used_size: Geometry.IntSize) -> None:
range_nm = 80
# print(f"{offset_m * 1E9}, {fov_size_nm} {extra_nm}")
# calculate destination bounds in nm
left_nm = -offset_m.x * 1E9 - (fov_size_nm.width + extra_nm.x) / 2
top_nm = -offset_m.y * 1E9 - (fov_size_nm.height + extra_nm.y) / 2
right_nm = left_nm + fov_size_nm.width + extra_nm.x
bottom_nm = top_nm + fov_size_nm.height + extra_nm.y
# print(f"{left_nm}, {top_nm} x {right_nm - left_nm}, {bottom_nm - top_nm}")
# map into fractional coordinates (0, 1) of source area where (-range_nm, range_nm) is the range in both axes
intersection_left_nm = max(left_nm, -range_nm)
intersection_top_nm = max(top_nm, -range_nm)
intersection_right_nm = min(right_nm, range_nm)
intersection_bottom_nm = min(bottom_nm, range_nm)
# print(f"{intersection_left_nm}, {intersection_top_nm} x {intersection_right_nm - intersection_left_nm}, {intersection_bottom_nm - intersection_top_nm}")
if intersection_left_nm < intersection_right_nm and intersection_top_nm < intersection_bottom_nm:
src_left = int(self.__amorphous.shape[1] * max(
(intersection_left_nm + range_nm) / (range_nm * 2), 0))
src_top = int(self.__amorphous.shape[0] * max(
(intersection_top_nm + range_nm) / (range_nm * 2), 0))
src_right = int(self.__amorphous.shape[1] * min(
(intersection_right_nm + range_nm) / (range_nm * 2), 1))
src_bottom = int(self.__amorphous.shape[0] * min(
(intersection_bottom_nm + range_nm) / (range_nm * 2), 1))
dst_left = int(data.shape[1] * max(
(intersection_left_nm - left_nm) / (right_nm - left_nm), 0))
dst_top = int(data.shape[0] * max(
(intersection_top_nm - top_nm) / (bottom_nm - top_nm), 0))
dst_right = int(data.shape[1] * min(
(intersection_right_nm - left_nm) / (right_nm - left_nm), 1))
dst_bottom = int(data.shape[0] * min(
(intersection_bottom_nm - top_nm) / (bottom_nm - top_nm), 1))
# print(f"{src_left}, {src_top} x {src_right - src_left}, {src_bottom - src_top} => {dst_left}, {dst_top} x {dst_right - dst_left}, {dst_bottom - dst_top}")
src = self.__amorphous[src_top:src_bottom, src_left:src_right]
src = scipy.ndimage.gaussian_filter(src, 3)
# may be faster, but doesn't work for non-square src
# src = numpy.fft.ifft2(scipy.ndimage.fourier_gaussian(src * 1j, 3)).real
src = 4 * (src - numpy.amin(src)) / numpy.ptp(src)
data[dst_top:dst_bottom, dst_left:dst_right] = Image.scaled(
src, (dst_bottom - dst_top, dst_right - dst_left))
def test_scale_cubic_is_symmetry(self):
src1 = numpy.zeros((8, 8))
src2 = numpy.zeros((9, 9))
src1[3:5, 3:5] = 1
src2[3:6, 3:6] = 1
src1s = (Image.scaled(src1,
(12, 12), 'cubic') * 1000).astype(numpy.int32)
src2s = (Image.scaled(src1,
(12, 12), 'cubic') * 1000).astype(numpy.int32)
src1t = (Image.scaled(src1,
(13, 13), 'cubic') * 1000).astype(numpy.int32)
src2t = (Image.scaled(src1,
(13, 13), 'cubic') * 1000).astype(numpy.int32)
self.assertTrue(numpy.array_equal(src1s[0:6, 0:6], src1s[0:6,
12:5:-1]))
self.assertTrue(
numpy.array_equal(src1s[0:6, 0:6], src1s[12:5:-1, 12:5:-1]))
self.assertTrue(numpy.array_equal(src1s[0:6, 0:6], src1s[12:5:-1,
0:6]))
self.assertTrue(numpy.array_equal(src2s[0:6, 0:6], src2s[0:6,
12:5:-1]))
self.assertTrue(
numpy.array_equal(src2s[0:6, 0:6], src2s[12:5:-1, 12:5:-1]))
self.assertTrue(numpy.array_equal(src2s[0:6, 0:6], src2s[12:5:-1,
0:6]))
self.assertTrue(numpy.array_equal(src1t[0:6, 0:6], src1t[0:6,
13:6:-1]))
self.assertTrue(
numpy.array_equal(src1t[0:6, 0:6], src1t[13:6:-1, 13:6:-1]))
self.assertTrue(numpy.array_equal(src1t[0:6, 0:6], src1t[13:6:-1,
0:6]))
self.assertTrue(numpy.array_equal(src2t[0:6, 0:6], src2t[0:6,
13:6:-1]))
self.assertTrue(
numpy.array_equal(src2t[0:6, 0:6], src2t[13:6:-1, 13:6:-1]))
self.assertTrue(numpy.array_equal(src2t[0:6, 0:6], src2t[13:6:-1,
0:6]))
def get_processed_data(self, data_sources, values):
data = data_sources[0].data
if data is None:
return None
img = Image.create_rgba_image_from_array(data) # inefficient since we're just converting back to gray
if id(img) == id(data):
img = img.copy()
if id(img.base) == id(data):
img = img.copy()
img = img.view(numpy.uint8).reshape(img.shape + (4,)) # expand the color into uint8s
img_gray = cv2.cvtColor(img, cv.CV_RGB2GRAY)
img_gray = cv2.equalizeHist(img_gray)
rects = detect(img_gray, relative_file(__file__, "haarcascade_frontalface_alt.xml"))
draw_rects(img, rects, (0, 255, 0))
return img
def get_processed_data(self, data_sources, values):
data = data_sources[0].data
if data is None:
return None
img = Image.create_rgba_image_from_array(
data) # inefficient since we're just converting back to gray
if id(img) == id(data):
img = img.copy()
if id(img.base) == id(data):
img = img.copy()
img = img.view(numpy.uint8).reshape(
img.shape + (4, )) # expand the color into uint8s
img_gray = cv2.cvtColor(img, cv.CV_RGB2GRAY)
img_gray = cv2.equalizeHist(img_gray)
rects = detect(
img_gray, relative_file(__file__,
"haarcascade_frontalface_alt.xml"))
draw_rects(img, rects, (0, 255, 0))
return img
def read_data_elements(self, ui, extension, path):
data = None
try:
data = Image.read_image_from_file(ui, path)
except Exception as e:
pass
if data is not None:
data_element = dict()
data_element["version"] = 1
data_element["data"] = data
if os.path.exists(path) or path.startswith(":"): # check for colon is for testing
try:
file_datetime = datetime.datetime.fromtimestamp(os.path.getmtime(path))
except:
file_datetime = None
if file_datetime is not None:
data_element["datetime_modified"] = Utility.get_datetime_item_from_datetime(file_datetime)
return [data_element]
return list()
def calculate_statistics(display_data_and_metadata_func,
display_data_range, region,
displayed_intensity_calibration):
display_data_and_metadata = display_data_and_metadata_func()
data = display_data_and_metadata.data if display_data_and_metadata else None
data_range = display_data_range
if data is not None and data.size > 0 and displayed_intensity_calibration:
mean = numpy.mean(data)
std = numpy.std(data)
rms = numpy.sqrt(numpy.mean(numpy.square(
numpy.absolute(data))))
sum_data = mean * functools.reduce(
operator.mul,
Image.dimensional_shape_from_shape_and_dtype(
data.shape, data.dtype))
if region is None:
data_min, data_max = data_range if data_range is not None else (
None, None)
else:
data_min, data_max = numpy.amin(data), numpy.amax(data)
mean_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
mean)
std_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
std)
data_min_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
data_min)
data_max_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
data_max)
rms_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
rms)
sum_data_str = displayed_intensity_calibration.convert_to_calibrated_value_str(
sum_data)
return {
"mean": mean_str,
"std": std_str,
"min": data_min_str,
"max": data_max_str,
"rms": rms_str,
"sum": sum_data_str
}
return dict()
def calculate_data():
rgb_image = numpy.empty(shape + (3, ), numpy.uint8)
channels = (blue_data_and_metadata, green_data_and_metadata,
red_data_and_metadata)
for channel_index, channel in enumerate(channels):
data = channel.data
if not Image.is_data_valid(data):
return None
if tuple(data.shape) != shape:
return None
if data.dtype.kind in 'iu':
rgb_image[..., channel_index] = numpy.clip(data, 0, 255)
elif data.dtype.kind in 'f':
rgb_image[..., channel_index] = numpy.clip(
numpy.multiply(data, 255), 0, 255)
else:
return None
return rgb_image
def calculate_statistics(display_data_and_metadata_func: typing.Callable[[], typing.Optional[DataAndMetadata.DataAndMetadata]], display_data_range: typing.Optional[typing.Tuple[float, float]], region: typing.Optional[Graphics.Graphic], displayed_intensity_calibration: typing.Optional[Calibration.Calibration]) -> typing.Dict[str, str]:
display_data_and_metadata = display_data_and_metadata_func()
data = display_data_and_metadata.data if display_data_and_metadata else None
data_range = display_data_range
if data is not None and data.size > 0 and displayed_intensity_calibration:
mean = numpy.mean(data)
std = numpy.std(data)
rms = numpy.sqrt(numpy.mean(numpy.square(numpy.absolute(data))))
dimensional_shape = Image.dimensional_shape_from_shape_and_dtype(data.shape, data.dtype) or (1, 1)
sum_data = mean * functools.reduce(operator.mul, dimensional_shape)
if region is None:
data_min, data_max = data_range if data_range is not None else (None, None)
else:
data_min, data_max = numpy.amin(data), numpy.amax(data)
mean_str = displayed_intensity_calibration.convert_to_calibrated_value_str(mean)
std_str = displayed_intensity_calibration.convert_to_calibrated_value_str(std)
data_min_str = displayed_intensity_calibration.convert_to_calibrated_value_str(data_min) if data_min is not None else str()
data_max_str = displayed_intensity_calibration.convert_to_calibrated_value_str(data_max) if data_max is not None else str()
rms_str = displayed_intensity_calibration.convert_to_calibrated_value_str(rms)
sum_data_str = displayed_intensity_calibration.convert_to_calibrated_value_str(sum_data)
return { "mean": mean_str, "std": std_str, "min": data_min_str, "max": data_max_str, "rms": rms_str, "sum": sum_data_str }
return dict()
def convert_data_element_to_data_and_metadata_1(data_element) -> DataAndMetadata.DataAndMetadata:
"""Convert a data element to xdata. No data copying occurs.
The data element can have the following keys:
data (required)
is_sequence, collection_dimension_count, datum_dimension_count (optional description of the data)
spatial_calibrations (optional list of spatial calibration dicts, scale, offset, units)
intensity_calibration (optional intensity calibration dict, scale, offset, units)
metadata (optional)
properties (get stored into metadata.hardware_source)
one of either timestamp or datetime_modified
if datetime_modified (dst, tz) it is converted and used as timestamp
then timezone gets stored into metadata.description.timezone.
"""
# data. takes ownership.
data = data_element["data"]
dimensional_shape = Image.dimensional_shape_from_data(data)
is_sequence = data_element.get("is_sequence", False)
dimension_count = len(Image.dimensional_shape_from_data(data))
adjusted_dimension_count = dimension_count - (1 if is_sequence else 0)
collection_dimension_count = data_element.get("collection_dimension_count", 2 if adjusted_dimension_count in (3, 4) else 0)
datum_dimension_count = data_element.get("datum_dimension_count", adjusted_dimension_count - collection_dimension_count)
data_descriptor = DataAndMetadata.DataDescriptor(is_sequence, collection_dimension_count, datum_dimension_count)
# dimensional calibrations
dimensional_calibrations = None
if "spatial_calibrations" in data_element:
dimensional_calibrations_list = data_element.get("spatial_calibrations")
if len(dimensional_calibrations_list) == len(dimensional_shape):
dimensional_calibrations = list()
for dimension_calibration in dimensional_calibrations_list:
offset = float(dimension_calibration.get("offset", 0.0))
scale = float(dimension_calibration.get("scale", 1.0))
units = dimension_calibration.get("units", "")
units = str(units) if units is not None else str()
if scale != 0.0:
dimensional_calibrations.append(Calibration.Calibration(offset, scale, units))
else:
dimensional_calibrations.append(Calibration.Calibration())
# intensity calibration
intensity_calibration = None
if "intensity_calibration" in data_element:
intensity_calibration_dict = data_element.get("intensity_calibration")
offset = float(intensity_calibration_dict.get("offset", 0.0))
scale = float(intensity_calibration_dict.get("scale", 1.0))
units = intensity_calibration_dict.get("units", "")
units = str(units) if units is not None else str()
if scale != 0.0:
intensity_calibration = Calibration.Calibration(offset, scale, units)
# properties (general tags)
metadata = dict()
if "metadata" in data_element:
metadata.update(Utility.clean_dict(data_element.get("metadata")))
if "properties" in data_element and data_element["properties"]:
hardware_source_metadata = metadata.setdefault("hardware_source", dict())
hardware_source_metadata.update(Utility.clean_dict(data_element.get("properties")))
# dates are _local_ time and must use this specific ISO 8601 format. 2013-11-17T08:43:21.389391
# time zones are offsets (east of UTC) in the following format "+HHMM" or "-HHMM"
# daylight savings times are time offset (east of UTC) in format "+MM" or "-MM"
# timezone is for conversion and is the Olson timezone string.
# datetime.datetime.strptime(datetime.datetime.isoformat(datetime.datetime.now()), "%Y-%m-%dT%H:%M:%S.%f" )
# datetime_modified, datetime_modified_tz, datetime_modified_dst, datetime_modified_tzname is the time at which this image was modified.
# datetime_original, datetime_original_tz, datetime_original_dst, datetime_original_tzname is the time at which this image was created.
timestamp = data_element.get("timestamp", datetime.datetime.utcnow())
datetime_item = data_element.get("datetime_modified", Utility.get_datetime_item_from_utc_datetime(timestamp))
local_datetime = Utility.get_datetime_from_datetime_item(datetime_item)
dst_value = datetime_item.get("dst", "+00")
tz_value = datetime_item.get("tz", "+0000")
timezone = datetime_item.get("timezone")
time_zone = { "dst": dst_value, "tz": tz_value}
if timezone is not None:
time_zone["timezone"] = timezone
# note: dst is informational only; tz already include dst
tz_adjust = (int(tz_value[1:3]) * 60 + int(tz_value[3:5])) * (-1 if tz_value[0] == '-' else 1)
utc_datetime = local_datetime - datetime.timedelta(minutes=tz_adjust) # tz_adjust already contains dst_adjust
timestamp = utc_datetime
return DataAndMetadata.new_data_and_metadata(data,
intensity_calibration=intensity_calibration,
dimensional_calibrations=dimensional_calibrations,
metadata=metadata,
timestamp=timestamp,
data_descriptor=data_descriptor,
timezone=timezone,
timezone_offset=tz_value)
def is_grayscale(data: _DataArrayType) -> bool:
if Image.is_data_rgb(data) or Image.is_data_rgba(data):
return numpy.array_equal(data[..., 0],
data[..., 1]) and numpy.array_equal(
data[..., 1], data[..., 2])
return True
