Open maedoc opened 9 years ago
Huge interest from me. We use SQL for all processing tracking and hacking my way through it has been difficult.
i would propose we create a logging service with multiple backends for this. one of the key things we want to standardize is provenance tracking. there is already some effort in provenance.py
. but instead of writing out these things at the end of execution, we literally want the information to be sent as messages.
currently we are using RDF to store such information and it's likely that we will continue to do so. you can always convert a subset of the graph to a relational table.
btw, have any of you used the provenance output from nipype?
is there still any interest in it?
Still interest for me!
Is there any interest in hooking up a SQL database to nipype interfaces, workflows, monitor what's running, etc.?