Convert the TempFile to opencv in a Jupyter notebook

Anvil Q&A
1

What I’m trying to do: Create an AI app with anvil.
To do this, I upload a picture and send it to a Jupyter notebook.
There the image should be converted to opencv, classified and returned to Anvil to be viewed on the web.

What I’ve tried and what doesn’t work:
I looked at the tutorial but it didn’t work.

How could I convert the anvil.media.TempFile to opencv in an Jupyter notebook process it and send it back to anvil?

Code Sample:

import anvil.media

@anvil.server.callable
def classify_image(file):
    with anvil.media.TempFile(file) as filename:
        img = load_img(filename)
??????? to opencv

return processed_image
2

This seems more about opencv than Anvil. Opencv is a library that will accept an image file. Anvil simply has a file storage wrapper it uses, and you can convert that image or write it to disk, depending n what you need.

Typically if I’m doing some ML on an image I use PIL. Torch for example uses pillow easily. Does opencv expect a file path or is an image in memory acceptable?

4

Here’s how I handled a video file with OpenCV. You may find you can adapt it suit your needs:

1 Like
6

Hi Owen,

Thanks, but I think it’s way too complicated.
I just want a way to transfer a image from anvil to the Jupyter notebook, save it there, open it with opencv, send it back to anvil and display it in the wep app.
Do you have a solution for that?

Kind regards
Erwin

7

Hi Robert,

It didn’t work with an image in memory. Do you have some way to save the TempFile image in the Jupyter notebook using PIL? I could open it then with opencv and do my transformation and ML stuff.
Do you have a solution for that?

Kind regards
Erwin

1 Like
8

You could write the file to the hard drive with some generic Python. Or you can use Anvil’s built-in method to do that:

Media Object to temporary file

1 Like