UnidentifiedImageError: cannot identify image file <_io.BytesIO object at 0x7f28da86eea0>

[hdocmsu]

Hello, I installed and tried the sample code from README.md and experienced the error below. Any suggestion is appreciated.
Thanks!

---

UnidentifiedImageError Traceback (most recent call last)
/tmp/ipykernel_316921/2729893238.py in
----> 1 get_ipython().run_cell_magic('compare', '', '\nfrom skimage import data\nfrom skimage.color import rgb2gray\nimport matplotlib.pyplot as plt\n\nimg = data.chelsea()\ngrayscale_img = rgb2gray(img)\n\nplt.imshow(img)\nplt.axis("off")\nplt.show()\n\nplt.imshow(grayscale_img, cmap="gray")\nplt.axis("off")\nplt.show()\n')

~/anaconda3/lib/python3.8/site-packages/IPython/core/interactiveshell.py in run_cell_magic(self, magic_name, line, cell)
2401 with self.builtin_trap:
2402 args = (magic_arg_s, cell)
-> 2403 result = fn(*args, **kwargs)
2404 return result
2405

~/anaconda3/lib/python3.8/site-packages/decorator.py in fun(*args, **kw)
230 if not kwsyntax:
231 args, kw = fix(args, kw, sig)
--> 232 return caller(func, *(extras + args), **kw)
233 fun.name = func.name
234 fun.doc = func.doc

~/anaconda3/lib/python3.8/site-packages/IPython/core/magic.py in (f, *a, **k)
185 # but it's overkill for just that one bit of state.
186 def magic_deco(arg):
--> 187 call = lambda f, *a, **k: f(*a, **k)
188
189 if callable(arg):

~/anaconda3/lib/python3.8/site-packages/jupyter_compare_view/sw_cellmagic.py in compare(self, line, cell)
63 imgdata = b64decode(out_images_base64[0])
64 # maybe possible without the PIL dependency?
---> 65 im = Image.open(io.BytesIO(imgdata))
66 height = im.size[1]
67

~/anaconda3/lib/python3.8/site-packages/PIL/Image.py in open(fp, mode, formats)
3121 for message in accept_warnings:
3122 warnings.warn(message)
-> 3123 raise UnidentifiedImageError(
3124 "cannot identify image file %r" % (filename if filename else fp)
3125 )

UnidentifiedImageError: cannot identify image file <_io.BytesIO object at 0x7f28da86eea0>

[kolibril13]

Hi @hdocmsu !
Can you provide also the code that you were running?
And further questions:

1. What OS are you on?
2. What browser are you using?

And secondly:
Can you once run these two code cells?

import io
from base64 import b64decode

from IPython.core import magic_arguments
from IPython.core.display import HTML
from IPython.core.magic import Magics, cell_magic, magics_class
from IPython.display import display
from IPython.utils.capture import capture_output
from PIL import Image


@magics_class
class SplitViewMagic(Magics):
    @magic_arguments.magic_arguments()
    @magic_arguments.argument(
        "--position",
        "-p",
        default="50%",
        help=("The position where the slider starts"),
    )
    @magic_arguments.argument(
        "--height",
        "-h",
        default="300",
        help=(
            "The height that the widget has. The width will be adjusted automatically. \
             If height is choosen 'auto', the height will be defined by the resolution \
             in vertical direction of the first image."
        ),
    )
    @cell_magic
    def splity(self, line, cell):
        """Saves the png image and calls the splitview canvas"""

        with capture_output(stdout=False, stderr=False, display=True) as result:
            self.shell.run_cell(cell)

        # saves all jupyter output images into the out_images_base64 list
        out_images_base64 = []
        print(f"{out_images_base64 = }")
        for output in result.outputs:
            data = output.data
            print(f"{type(data) = }")
            print(f"{type(list(data.values())[0]) = }")

            if "image/png" in data:
                png_bytes_data = data["image/png"]
                out_images_base64.append(png_bytes_data)
        
        print(f"{len(out_images_base64) = }")

        # get the parameters the configure the widget
        args = magic_arguments.parse_argstring(SplitViewMagic.splity, line)

        slider_position = args.position
        widget_height = args.height

        if widget_height == "auto":
            imgdata = b64decode(out_images_base64[0])
            # maybe possible without the PIL dependency?
            im = Image.open(io.BytesIO(imgdata))
            width, height = im.size
            widget_height = height

        html_code = f"""
        <div class="outer_layer" style="position: relative; padding-top: {int(widget_height)+3}px;"> 
            <div class="juxtapose" data-startingposition="{slider_position}" style="height: {int(widget_height)}px;; width: auto; top: 1%; left: 1%; position: absolute;">
                <img src="data:image/jpeg;base64,{out_images_base64[0]}" />' <img src="data:image/jpeg;base64,{out_images_base64[1]}" />'
            </div>
        </div>
        <script src="https://cdn.knightlab.com/libs/juxtapose/latest/js/juxtapose.min.js"></script>
        <link rel="stylesheet" href="https://cdn.knightlab.com/libs/juxtapose/latest/css/juxtapose.css" />
        """
        display(HTML(html_code))
        
from IPython import get_ipython  # register cell magic

ipy = get_ipython()
ipy.register_magics(SplitViewMagic)

and this in the second cell?

%%splity --position 73% 
import matplotlib.pyplot as plt
import numpy as np

array1 = np.full((15, 30), 10)
array2 = np.random.randint(0, 10, size=(15, 30))
fig, ax1 = plt.subplots(figsize=(5, 10))
ax1.imshow(array1)
fig, ax2 = plt.subplots(figsize=(5, 10))
ax2.imshow(array2)

Output should be:

out_images_base64 = []
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
len(out_images_base64) = 2

[hdocmsu]

Dear @kolibril13, thanks a lot for your quick response!

Here is the code that I ran on Jupiter lab. The code is identical to the example in README.md.

%%compare  
from skimage import data
from skimage.color import rgb2gray
import matplotlib.pyplot as plt

img = data.chelsea()
grayscale_img = rgb2gray(img)

plt.imshow(img)
plt.axis("off")
plt.show()

plt.imshow(grayscale_img, cmap="gray")
plt.axis("off")
plt.show()

What OS are you on? --> I am using Linux Ubuntu 20x
What browser are you using? I am using Chrome
Here is the output after I ran the two cell:

out_images_base64 = []
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
type(data) = <class 'dict'>
type(list(data.values())[0]) = <class 'str'>
len(out_images_base64) = 2

[christopher-besch]

This is an issue with your (@kolibril13) way of determining the image height; I never touched that code.

@hdocmsu, could you try manually setting the height using %%compare --height 220? If my suspect is correct, the error should disappear.

@kolibril13, why is this not in the with statement above:

jupyter_compare_view/jupyter_compare_view/sw_cellmagic.py

Lines 45 to 55 in 85f7960

	# saves all jupyter output images into the out_images_base64 list
	out_images_base64 = []
	for output in result.outputs:
	data = output.data
	if "image/png" in data:
	png_bytes_data = data["image/png"]
	out_images_base64.append(png_bytes_data)
	if len(out_images_base64) < 2:
	raise ValueError(
	"There need to be at least two images for Jupyter compare_view to work."
	)

Shouldn't result be out of scope here?

[hdocmsu]

thanks for the response @christopher-besch, when i tried "%%compare --hight 220" the error does not show up but I got a blank output.

[christopher-besch]

Then this seems to be an issue with your Jupyter Lab installation. Could you try running some very basic cell that outputs an image?

[hdocmsu]

the image output is fine from Jupiter lab.