Out of bounds fix #85
Conversation
Added a simple, and wholly incorrect, try-execute catch in `field.py`to return zero values for attempts at out-of-bounds field.data sampling. This is not anticipated as a final solution, but rather is a temporary fix to allow some simulation work to continue on HPC clusters.
|
I totally understand that you need a quick fix to the out-of-bounds error, so thanks for creating this (slightly dirty) hack, @Jacketless While I'd be happy to temporarily add this code to master, we need a better solution in the long term of course. The other option, if we don't want to 'spoil' master with these quick fixes, is that you temporarily fork parcels and then add it in your own forked version. @mlange05, what do you think? |
|
Ok, IMHO I would prefer a well-documented hack over maintaining a separate fork, since a hacky solution is better than no solution at all. I would like a clear in-place developer comment though that points to the relevant github issue, in case interested people stumble across this. |
|
I agree with @mlange05. So if you can add some more comments, pointing to the relevant issues on the github page, @Jacketless, then I'll pull this into master |
Added some documentation to reference the github issues this hot fix refers to, as well as some more useful print output when out-of-bounds errors are caught
Not suggesting that this be merged with the master branch, but wanted to have a short-term fix branch that can be checked out to the HPC cluster so I can keep running simulations there.
However, if we want, this branch could form the baseline for our eventual solutions discussed in #47 #61 and #76