Closed iperdomo closed 2 years ago
Hey @iperdomo thanks for reporting this, and for include that helpful video to illustrate the problem. ⭐
In order to run the juptyer
command, you'll need to add the jupyterlab
dependency to your cog.yaml
file, like so:
build:
python_packages:
- "jupyterlab==3.3.4"
You should then be able to cog build
then run the notebook command above.
I will open a PR to update the documentation around this.
Hi @zeke thanks for the explanation, that was my guess. You have to install jupyter
as part of the dependencies. Adding jupyter to the dependencies make the docker image even bigger, and you don't need jupyter in the production image, just for development.
It would be interesting to define the scope of the package, e.g. production vs development. Different ecosystems solve the issue in different ways, e.g. npm
with devDependencies. I'm not sure if there is way to specify dev dependencies with pip
@all-contributors please add @iperdomo for bugs
@zeke
I've put up a pull request to add @iperdomo! :tada:
Hi,
The documentation suggests that I should be able to run a jupyter notebook by issuing a
cog run jupyter [...]
command - See #488I'm testing Cog version 0.3.0 and I'm getting the following error in a fresh sample project:
Environment
Steps to reproduce
cog init
cog run -p 8080 jupyter notebook --ip 0.0.0.0 --no-browser --port=8080 --allow-root
See attached video
https://user-images.githubusercontent.com/178474/164511281-1791c929-e498-44fc-b6ec-a5652b0aa964.mp4