microsoft / AIforEarthDataSets

Notebooks and documentation for AI-for-Earth-managed datasets on Azure

Home Page:https://microsoft.github.io/AIforEarthDataSets/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Adding requirements in term of dependencies

yvanlebras opened this issue · comments

Hi AIforEarthDataSets amazing team!

THANK YOU so much for this amazing work! As we (through usegalaxy.eu / Galaxy for Ecology / Galaxy for Climate / BiodiFAIRse GO FAIR Implementation network / PNDB French Biodiversity e-infrastructure / French Challenge IA-Biodiv ) want to capitalize onthis work and facilitate the reuse of these notebooks, it appears to me that adding informations about requirements in terms of dependencies like R packages or Python libraries for each notebook can be of great interest to rise the FAIRness of each. A preliminary idea can be to add a requirements.txt file like this example https://github.com/galaxyecology/webhook_SPIPOLL_Flash/blob/master/requirements.txt and for sure something better an environment.yml listing conda packages like this example https://github.com/IFB-ElixirFr/ReproHackathon/blob/master/reprohackathon1/galaxy/environment.yml to each notebook, for example gbif_environment.yml for the gbif notebook.

If you think this can be relevant, please don't hesitate to comment / propose way to do so.

Whishing you a very good end of week,

Best,

Yvan

Thanks for your suggestion! This is a good idea and we will add it to the backlog, though I don't think we're likely to get to this that quickly. We definitely welcome contributions, so if you have built requirements files for one or more notebooks, feel free to submit a PR.

In general, we also try to list package dependencies in the first cell of every notebook, and we try to specify in comments when one would need to do anything other than "pip install [x]". If you find notebooks where reality is more complicated than that, let us know.

Thanks!

Thank you for your answer!

I understand, and I think we will go starting this making PR for some notebooks we are interested to use / test.

Good to see that you already try to facilitate dependencies management. Unfortunately, if I am not wrong, the notebook I am looking for is not presenting dependencies in any of the mentionned manner https://github.com/microsoft/AIforEarthDataSets/blob/main/data/gbif.ipynb that's why I was opening this PR. I think I will start with this one so.

Wishing you a very good week,

Thanks !