Add a devcontainer #637
Add a devcontainer #637
Conversation
Add a devcontainer for easy codespace startup. It will start the container with a preconfigured python environment and all needed dev tooling to start working on the project.
CONTRIBUTING.md
Outdated
|
|
||
| ## Updating Javascript packages. | ||
|
|
||
| This project uses lerna to manage the multiple packages inside the packages folder. If you run npm command at the root it will run the command for each subpackage. This is good for working across packages, however more commonly we want to work with an individual package. To do this change directories into the package you wish to alter, then npm will run only for that package. |
There was a problem hiding this comment.
Should probably say "npm commands" here, or "an npm command"
Co-authored-by: Adam B <el-mapache@users.noreply.github.com>
.devcontainer/scripts/setup.sh
Outdated
| jupyter nbextension enable --py nbdime --sys-prefix | ||
|
|
||
| jupyter labextension link ./packages/nbdime --no-build | ||
| jupyter labextension install ./packages/labextension |
There was a problem hiding this comment.
@vidartf This codespace now includes Jupyter Lab, this will allow people starting up the codespace to have a pre-install Jupyter Lab to develop against.
However I am not sure the best way to link the local version of nbdime to the JupyterLab server.
This works, but I am not sure how to get changes to show up, I need to rebuild the docker container every time which is not great.
Is there some sort of script I could put together to automatically update Jupyter Lab? Thanks..
There was a problem hiding this comment.
I was trying to use this command
jupyter labextension develop --overwrite .
But it was not working as I would expect. I wonder if there is some prereq I am not aware of that the cookie cutter does?
There was a problem hiding this comment.
jupyter labextension develop . --overwrite is correct, but not sufficient on its own. If you change some code in the nbdime package, you would then need to
- build the nbdime code (
npm run buildin packages/nbdime) - build typescript code of tha labextension, and then bundle it into the prebuilt extension (both covered by command
npm run build:devin packages/labextension)
To have it build automatically (and quicker), you can set up watchers:
npm run watchin packages/nbdimenpm run watchin packages/labextensionjupyter labextension watch? This one currently doesn't work for me, so might be worth asking lab devs about if there is a good watch command..
There was a problem hiding this comment.
Cool thanks. I will dig into this
.vscode/scripts/watch-webapp.sh
Outdated
| @@ -0,0 +1,14 @@ | |||
| #!/usr/bin/env bash | |||
There was a problem hiding this comment.
I can't run these shell script tasks on Windows :( If you could write those shell scripts as npm commands, that would be much more platform independent.
There was a problem hiding this comment.
You could run them in linux on a codespace 😅, not a windows fan my self. I will move them.
.vscode/tasks.json
Outdated
| @@ -0,0 +1,86 @@ | |||
| { | |||
| "tasks": [ | |||
There was a problem hiding this comment.
In general I would prefer not to include IDE specific files in a repo. Especially since that means that whoever maintains that repo will then have to maintain all the config files for those IDEs, even if they don't use it themselves. Making these tasks into "convenience" npm commands, and documenting them would be a better alternative.
|
@vidartf I have made some updates to this to make it more codespaces centric. Wanna kick the tires and see what you think? |
|
I am just going to simplify this and go for docs for now. |
Add a devcontainer for easy codespace startup. It will start the container with a preconfigured python environment and all needed dev tooling to start workin on the project.
cc @craiglpeters