[Bug]: PsychPy is unable to load images from a public online source. Specifically, it is unable to load an image from a public AWS s3 bucket, as in: https://neurotype-images.s3.us-east-2.amazonaws.com/affective-images/004504.jpg about psychopy HOT 4 OPEN

I'm going to relabel this as an enhancement as currently loading images from a URL isn't something ImageStim is supposed to be able to do - that would be a new feature rather than a fix. Would be nice though, and very possible as both requests and PIL are already in PsychoPy's core requirements, so I'll add the "Good first issue" tag too.

One thing to be aware of with this is that accessing a resource from a server means your IP is exposed. This is fine if it's e.g. the Pavlovia server, but when connecting to something like amazonaws we would definitely need to have some kind of warning. There was a case in 2022 where a German court ruled that accessing the Google Fonts server without warning violated GDPR, which is why PsychoPy warns you when you try to use a Google Font in TextStim/TextBox.

from psychopy.

Just to add... PsychoPy is designed for high-performance, very low-latency operation, so adding in the delay of fetching resources from the web when they are needed should be treated very carefully considering those goals. I would strongly recommend fetching local copies of images for your studies in advance when running in the lab

from psychopy.

If you definitely need to get it from online and are okay with the loss of speed from having to query a server each time, the following code should work (using url for the image url and img for the ImageStim):

# get raw data from server
raw = requests.get(url, stream=True).raw
# create a PIL image from raw data
data = pil.open(raw)
# create an image stim
img.setImage(data)

(you could also use requests.get, pil.open and pil.save to programmatically get and save all the images just once, so you can use the local files)

from psychopy.

from psychopy.