Open BaptisteVandecrux opened 1 year ago
We already only re-process the L0 raw
data if a change is detected in the raw/config
toml file. It could be that we have a static L3 tx
dataset that new L3 tx
data lines are appended to.
Another option is to scrap the separate processing of tx
and raw
data, and merge them under one config toml file - then append new transmissions and only re-process the entire dataset when a flag file or the config toml file changes (as stated previously). We would just need to make sure that instantaneous measurements are retained in the final L3
dataset, but perhaps not distributed with the pubilcly available dataset.
I think it is a long term goal but not a priority at all. I just wanted that Patrick's comment in the PR did not get lost after the merge.
We already only re-process the
L0 raw
data if a change is detected in theraw/config
toml file. It could be that we have a staticL3 tx
dataset that newL3 tx
data lines are appended to.
Since we are about to update a lot of flags and adjustments, and that we keep finding L0 files that were not being processed, I'd suggest that the entire processing chain is being run each time, at least for now. So we are sure that the latest raw files and a flag/adjustment files are being used each time.
Originally posted by @patrickjwright in https://github.com/GEUS-Glaciology-and-Climate/pypromice/issues/92#issuecomment-1398231020