* Bump `base-notebook`
* OS version
* `jupyterlab`
* Bump `scipy-notebook`
* `beautifulsoup4`
* `cloudpickle`
* `dask`
* `numba`
* `xlrd`: version specified (it was not the case)
* Bump `datascience-notebook`
* `r-devtools`
* `r-forecast`
* Julia version
* Test added to check if Julia is correctly installed (`julia --version`)
* Note: A new version of `r-base` is out `4.0.0` however it cannot be installed yet since dependencies with other packages cannot be resolved.
* Bump `r-notebook`
* Same as `datascience-notebook` except Julia.
* Bump `all-spark-notebook`
* `r-sparklyr`
- Bump root container version
- Bump packages of
- scipy-notebook
- r-notebook
- datascience-notebook
- add -y flag to jupyter lab commands in scipy-notebook
- Fix a bug in check outdated packages when packages installed through pip
Add another layer of abstraction to the base container, namely the root container `ROOT_CONTAINER`, so that the user can have the flexibility to build the whole series of images with `make` based on a different base image.
Add optional environment variable RESTARTABLE to enable Jupyter to
be restarted without causing the container to exit. When the variable
is set, run-one-constantly (from the Ubuntu run-one package) is
inserted into the Jupyter startup command.
Pin the conda version so that a call to `conda update --all` will not
cause conda to update itself to a higher version with possible breaking
changes.
Motivating example: Upgrade from __4.6.*__ to __4.7.5__ causes the build to fail
due to a conda error:
[Errno 13] Permission denied: '/.cph_tmpilpdzq7x'
Resolves#891
This commit adds the additional `-f` force command to all uses of `conda
clean --all` through the repo. Size should be smaller, but still testing
if anything breaks. See issue #861.
