Amy Wooding
Amy Wooding
Right now the script invokes `conda-forge` even when `conda-forge` isn't invoked in the `environment.yml`. We need to be able to pull from other self-hosted channels as well.
The current way we handle data hashing doesn't survive package upgrades. For example, with pandas, we have been dumping dataframes and the hashes change (even if the data itself doesn't)...
Autogenerated MIT license includes both the MIT license and a Copyright line (for proprietary). Is this how the MIT license works?
`run_notebook` in src/utils.py needs to optionally create output paths if they don't exist.
We love pathlib. It is great for using paths in a platform independent way (windows/mac/linux). We use it in easydata when we give everything in an easydata repo a standard...
take out extra instructions from the readme
Add unit tests for notebooks that are part of the documentation (via mkdocs).
If I change my data_path in `config.ini` to another location, it doesn't survive restarting my kernel in my notebook. Reloading the kernel overwrites `config.ini`.
Add a warning message (or something like it) if there is no license.
interim/cache/datasets interim/cache/models