search

comses / miracle

Repeatable data analysis workflows for computational models
1 stars 3 forks source link

upload dataset zips #9

Closed alee closed 8 years ago

alee commented 9 years ago

zipfile with prescribed structure:

cpritcha commented 9 years ago

What benefits do you see from prescribing a structure on zipfile uploads?

I can see that file classification may be easier (every file in the data folder is of the Data type and everything in the scripts folder is of the Code type) but classification by extension should also handle most of these cases.

cpritcha commented 9 years ago

I think that having special filenames could be helpful. A README file in the base directory could be turned into the Analysis description or a requirements.txt file in the base directory could be classified as a Configuration type.

alee commented 9 years ago

I could go either way but I think I prefer the template structure because it makes the user think about curating their own data and not throw a pile of files at us and have us try to sort things out. So the main benefits are simplicity and explicitness, and hopefully reduced false positives, especially for overloaded text data, e.g., .txt files can hold CSV/TSV data just fine, but if they live in a /data directory we will interpret them as datasets.

alee commented 8 years ago

https://github.com/comses/miracle/wiki/Project-Archive-Preparation-Guidelines