spacepy / dbprocessing

Automated processing controller for heliophysics data
5 stars 4 forks source link

Make ProcessQueue S3-aware #133

Open jtniehof opened 9 months ago

jtniehof commented 9 months ago

Because ProcessQueue knows the input and output filenames, and builds the command line, it should be able to retrieve the inputs for the processing code and put the

Relation to an issue

131 and #132, although there are no dependencies.

Proposed enhancement

dbprocessing could be updated to be aware of S3-style URIs as valid file paths. ProcessQueue could then pull all needed input files from S3, place them in a mounted filesystem space, call the code with the mounted filesystem paths instead of S3, and then retrieve the created file and put it in S3.

Handling of the error directory would be potentially interesting.

Closure condition

Closed when can use S3 URIs as file paths without the code being S3-aware. This will need to be unit tested, which is again potentially interesting.

This may break into multiple subissues.