import numpy as np
from yt._maintenance.ipython_compat import IS_IPYTHON
from yt.config import ytcfg
from yt.funcs import mylog
from yt.units.yt_array import YTQuantity
from yt.utilities import png_writer as pw
from yt.utilities.exceptions import YTNotInsideNotebook
from yt.utilities.lib import image_utilities as au
from yt.visualization.color_maps import get_colormap_lut
from ._commons import get_canvas, validate_image_name
[docs]
def scale_image(image, mi=None, ma=None):
r"""Scale an image ([NxNxM] where M = 1-4) to be uint8 and values scaled
from [0,255].
Parameters
----------
image : array_like or tuple of image info
Examples
--------
>>> image = scale_image(image)
>>> image = scale_image(image, min=0, max=1000)
"""
if isinstance(image, np.ndarray) and image.dtype == np.uint8:
return image
if isinstance(image, (tuple, list)):
image, mi, ma = image
if mi is None:
mi = image.min()
if ma is None:
ma = image.max()
image = np.clip((image - mi) / (ma - mi) * 255, 0, 255).astype("uint8")
return image
[docs]
def multi_image_composite(
fn, red_channel, blue_channel, green_channel=None, alpha_channel=None
):
r"""Write an image with different color channels corresponding to different
quantities.
Accepts at least a red and a blue array, of shape (N,N) each, that are
optionally scaled and composited into a final image, written into `fn`.
Can also accept green and alpha.
Parameters
----------
fn : string
Filename to save
red_channel : array_like or tuple of image info
Array, of shape (N,N), to be written into the red channel of the output
image. If not already uint8, will be converted (and scaled) into
uint8. Optionally, you can also specify a tuple that includes scaling
information, in the form of (array_to_plot, min_value_to_scale,
max_value_to_scale).
blue_channel : array_like or tuple of image info
Array, of shape (N,N), to be written into the blue channel of the output
image. If not already uint8, will be converted (and scaled) into
uint8. Optionally, you can also specify a tuple that includes scaling
information, in the form of (array_to_plot, min_value_to_scale,
max_value_to_scale).
green_channel : array_like or tuple of image info, optional
Array, of shape (N,N), to be written into the green channel of the
output image. If not already uint8, will be converted (and scaled)
into uint8. If not supplied, will be left empty. Optionally, you can
also specify a tuple that includes scaling information, in the form of
(array_to_plot, min_value_to_scale, max_value_to_scale).
alpha_channel : array_like or tuple of image info, optional
Array, of shape (N,N), to be written into the alpha channel of the output
image. If not already uint8, will be converted (and scaled) into uint8.
If not supplied, will be made fully opaque. Optionally, you can also
specify a tuple that includes scaling information, in the form of
(array_to_plot, min_value_to_scale, max_value_to_scale).
Examples
--------
>>> red_channel = np.log10(frb["gas", "temperature"])
>>> blue_channel = np.log10(frb["gas", "density"])
>>> multi_image_composite("multi_channel1.png", red_channel, blue_channel)
"""
red_channel = scale_image(red_channel)
blue_channel = scale_image(blue_channel)
if green_channel is None:
green_channel = np.zeros(red_channel.shape, dtype="uint8")
else:
green_channel = scale_image(green_channel)
if alpha_channel is None:
alpha_channel = np.zeros(red_channel.shape, dtype="uint8") + 255
else:
alpha_channel = scale_image(alpha_channel)
image = np.array([red_channel, green_channel, blue_channel, alpha_channel])
image = image.transpose().copy() # Have to make sure it's contiguous
pw.write_png(image, fn)
[docs]
def write_bitmap(bitmap_array, filename, max_val=None, transpose=False):
r"""Write out a bitmapped image directly to a PNG file.
This accepts a three- or four-channel `bitmap_array`. If the image is not
already uint8, it will be scaled and converted. If it is four channel,
only the first three channels will be scaled, while the fourth channel is
assumed to be in the range of [0,1]. If it is not four channel, a fourth
alpha channel will be added and set to fully opaque. The resultant image
will be directly written to `filename` as a PNG with no colormap applied.
`max_val` is a value used if the array is passed in as anything other than
uint8; it will be the value used for scaling and clipping in the first
three channels when the array is converted. Additionally, the minimum is
assumed to be zero; this makes it primarily suited for the results of
volume rendered images, rather than misaligned projections.
Parameters
----------
bitmap_array : array_like
Array of shape (N,M,3) or (N,M,4), to be written. If it is not already
a uint8 array, it will be scaled and converted to uint8.
filename : string
Filename to save to. If None, PNG contents will be returned as a
string.
max_val : float, optional
The upper limit to clip values to in the output, if converting to uint8.
If `bitmap_array` is already uint8, this will be ignore.
transpose : boolean, optional
If transpose is False, we assume that the incoming bitmap_array is such
that the first element resides in the upper-left corner. If True, the
first element will be placed in the lower-left corner.
"""
if len(bitmap_array.shape) != 3 or bitmap_array.shape[-1] not in (3, 4):
raise RuntimeError(
"Expecting image array of shape (N,M,3) or "
f"(N,M,4), received {str(bitmap_array.shape)}"
)
if bitmap_array.dtype != np.uint8:
s1, s2 = bitmap_array.shape[:2]
if bitmap_array.shape[-1] == 3:
alpha_channel = 255 * np.ones((s1, s2, 1), dtype="uint8")
else:
alpha_channel = (255 * bitmap_array[:, :, 3]).astype("uint8")
alpha_channel.shape = s1, s2, 1
if max_val is None:
max_val = bitmap_array[:, :, :3].max()
if max_val == 0.0:
# avoid dividing by zero for blank images
max_val = 1.0
bitmap_array = np.clip(bitmap_array[:, :, :3] / max_val, 0.0, 1.0) * 255
bitmap_array = np.concatenate(
[bitmap_array.astype("uint8"), alpha_channel], axis=-1
)
if transpose:
bitmap_array = bitmap_array.swapaxes(0, 1).copy(order="C")
if filename is not None:
pw.write_png(bitmap_array, filename)
else:
return pw.write_png_to_string(bitmap_array.copy())
return bitmap_array
[docs]
def write_image(image, filename, color_bounds=None, cmap_name=None, func=lambda x: x):
r"""Write out a floating point array directly to a PNG file, scaling it and
applying a colormap.
This function will scale an image and directly call libpng to write out a
colormapped version of that image. It is designed for rapid-fire saving of
image buffers generated using `yt.visualization.api.FixedResolutionBuffers`
and the likes.
Parameters
----------
image : array_like
This is an (unscaled) array of floating point values, shape (N,N,) to
save in a PNG file.
filename : string
Filename to save as.
color_bounds : tuple of floats, optional
The min and max to scale between. Outlying values will be clipped.
cmap_name : string, optional
An acceptable colormap. See either yt.visualization.color_maps or
https://scipy-cookbook.readthedocs.io/items/Matplotlib_Show_colormaps.html .
func : function, optional
A function to transform the buffer before applying a colormap.
Returns
-------
scaled_image : uint8 image that has been saved
Examples
--------
>>> sl = ds.slice(0, 0.5, "Density")
>>> frb1 = FixedResolutionBuffer(sl, (0.2, 0.3, 0.4, 0.5), (1024, 1024))
>>> write_image(frb1["gas", "density"], "saved.png")
"""
if cmap_name is None:
cmap_name = ytcfg.get("yt", "default_colormap")
if len(image.shape) == 3:
mylog.info("Using only channel 1 of supplied image")
image = image[:, :, 0]
to_plot = apply_colormap(image, color_bounds=color_bounds, cmap_name=cmap_name)
pw.write_png(to_plot, filename)
return to_plot
[docs]
def apply_colormap(image, color_bounds=None, cmap_name=None, func=lambda x: x):
r"""Apply a colormap to a floating point image, scaling to uint8.
This function will scale an image and directly call libpng to write out a
colormapped version of that image. It is designed for rapid-fire saving of
image buffers generated using `yt.visualization.api.FixedResolutionBuffers`
and the likes.
Parameters
----------
image : array_like
This is an (unscaled) array of floating point values, shape (N,N,) to
save in a PNG file.
color_bounds : tuple of floats, optional
The min and max to scale between. Outlying values will be clipped.
cmap_name : string, optional
An acceptable colormap. See either yt.visualization.color_maps or
https://scipy-cookbook.readthedocs.io/items/Matplotlib_Show_colormaps.html .
func : function, optional
A function to transform the buffer before applying a colormap.
Returns
-------
to_plot : uint8 image with colorbar applied.
"""
if cmap_name is None:
cmap_name = ytcfg.get("yt", "default_colormap")
from yt.data_objects.image_array import ImageArray
image = ImageArray(func(image))
if color_bounds is None:
mi = np.nanmin(image[~np.isinf(image)]) * image.uq
ma = np.nanmax(image[~np.isinf(image)]) * image.uq
color_bounds = mi, ma
else:
color_bounds = [YTQuantity(func(c), image.units) for c in color_bounds]
image = (image - color_bounds[0]) / (color_bounds[1] - color_bounds[0])
to_plot = map_to_colors(image, cmap_name)
to_plot = np.clip(to_plot, 0, 255)
return to_plot
[docs]
def map_to_colors(buff, cmap_name):
lut = get_colormap_lut(cmap_name)
if isinstance(cmap_name, tuple):
# If we are using the colorbrewer maps, don't interpolate
shape = buff.shape
# We add float_eps so that digitize doesn't go out of bounds
x = np.mgrid[0.0 : 1.0 + np.finfo(np.float32).eps : lut[0].shape[0] * 1j]
inds = np.digitize(buff.ravel(), x)
inds.shape = (shape[0], shape[1])
mapped = np.dstack([(v[inds] * 255).astype("uint8") for v in lut])
del inds
else:
x = np.mgrid[0.0 : 1.0 : lut[0].shape[0] * 1j]
mapped = np.dstack([(np.interp(buff, x, v) * 255).astype("uint8") for v in lut])
return mapped.copy("C")
[docs]
def splat_points(image, points_x, points_y, contribution=None, transposed=False):
if contribution is None:
contribution = 100.0
val = contribution * 1.0 / points_x.size
if transposed:
points_y = 1.0 - points_y
points_x = 1.0 - points_x
im = image.copy()
au.add_points_to_image(im, points_x, points_y, val)
return im
[docs]
def write_projection(
data,
filename,
colorbar=True,
colorbar_label=None,
title=None,
vmin=None,
vmax=None,
take_log=True,
figsize=(8, 6),
dpi=100,
cmap_name=None,
extent=None,
xlabel=None,
ylabel=None,
):
r"""Write a projection or volume rendering to disk with a variety of
pretty parameters such as limits, title, colorbar, etc. write_projection
uses the standard matplotlib interface to create the figure. N.B. This code
only works *after* you have created the projection using the standard
framework (i.e. the Camera interface or off_axis_projection).
Accepts an NxM sized array representing the projection itself as well
as the filename to which you will save this figure. Note that the final
resolution of your image will be a product of dpi/100 * figsize.
Parameters
----------
data : array_like
image array as output by off_axis_projection or camera.snapshot()
filename : string
the filename where the data will be saved
colorbar : boolean
do you want a colorbar generated to the right of the image?
colorbar_label : string
the label associated with your colorbar
title : string
the label at the top of the figure
vmin : float or None
the lower limit of the zaxis (part of matplotlib api)
vmax : float or None
the lower limit of the zaxis (part of matplotlib api)
take_log : boolean
plot the log of the data array (and take the log of the limits if set)?
figsize : array_like
width, height in inches of final image
dpi : int
final image resolution in pixels / inch
cmap_name : string
The name of the colormap.
Examples
--------
>>> image = off_axis_projection(ds, c, L, W, N, "Density", no_ghost=False)
>>> write_projection(
... image,
... "test.png",
... colorbar_label="Column Density (cm$^{-2}$)",
... title="Offaxis Projection",
... vmin=1e-5,
... vmax=1e-3,
... take_log=True,
... )
"""
if cmap_name is None:
cmap_name = ytcfg.get("yt", "default_colormap")
import matplotlib.colors
import matplotlib.figure
# If this is rendered as log, then apply now.
if take_log:
norm_cls = matplotlib.colors.LogNorm
else:
norm_cls = matplotlib.colors.Normalize
norm = norm_cls(vmin=vmin, vmax=vmax)
# Create the figure and paint the data on
fig = matplotlib.figure.Figure(figsize=figsize)
ax = fig.add_subplot(111)
cax = ax.imshow(
data.to_ndarray(),
norm=norm,
extent=extent,
cmap=cmap_name,
)
if title:
ax.set_title(title)
if xlabel:
ax.set_xlabel(xlabel)
if ylabel:
ax.set_ylabel(ylabel)
# Suppress the x and y pixel counts
if extent is None:
ax.set_xticks(())
ax.set_yticks(())
# Add a color bar and label if requested
if colorbar:
cbar = fig.colorbar(cax)
if colorbar_label:
cbar.ax.set_ylabel(colorbar_label)
filename = validate_image_name(filename)
canvas = get_canvas(fig, filename)
mylog.info("Saving plot %s", filename)
fig.tight_layout()
canvas.print_figure(filename, dpi=dpi)
return filename
[docs]
def display_in_notebook(image, max_val=None):
"""
A helper function to display images in an IPython notebook
Must be run from within an IPython notebook, or else it will raise
a YTNotInsideNotebook exception.
Parameters
----------
image : array_like
This is an (unscaled) array of floating point values, shape (N,N,3) or
(N,N,4) to display in the notebook. The first three channels will be
scaled automatically.
max_val : float, optional
The upper limit to clip values of the image. Only applies to the first
three channels.
"""
if IS_IPYTHON:
from IPython.core.displaypub import publish_display_data
data = write_bitmap(image, None, max_val=max_val)
publish_display_data(
data={"image/png": data},
source="yt.visualization.image_writer.display_in_notebook",
)
else:
raise YTNotInsideNotebook