Clarity between `hosted_image` and `is_hosted` in `Project.upload`

I find it confusing that in Project.upload there's both a hosted_image: boolargument set by the user, and an internal is_hosted, that determines which errors may be thrown.

By setting hosted_image it's possible to bypass checks and send the image to Roboflow, which promptly returns a 500 code.

Still, maybe there's a reason behind it?

Test code:

import roboflow

img_path = "<path_to_img>"
rf = roboflow.Roboflow(api_key="API_KEY")

workspace_id = "<ws ID>"
project_id = "<proj ID>"
workspace = rf.workspace(workspace_id)
project = workspace.project(project_id)

project.upload(img_path, hosted_image=True)

This raises:

File  ... /roboflow/adapters/rfapi.py:90, in upload_image(api_key, project_url, image_path, hosted_image, split, batch_name, tag_names, sequence_number, sequence_size, **kwargs)
UploadError: Bad response: 500: {'error': 'Unknown error'}

Start at Project.upload and roboflow/adapters/rfapi.py:90, where upload_image is reported to raise UploadError after a 500 response. Trace how hosted_image and internal is_hosted are interpreted and compare the validation and error paths. Done means their distinction and expected behavior are explicit, and the shown call no longer bypasses checks into an unexplained server error.