Conversation
|
@wcarthur here's the ensemble scripts - I just created a draft PR for vidsibility |
Codecov Report
@@ Coverage Diff @@
## develop #55 +/- ##
===========================================
- Coverage 86.40% 86.29% -0.11%
===========================================
Files 26 26
Lines 1839 1839
===========================================
- Hits 1589 1587 -2
- Misses 250 252 +2
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
ensemble/run_percentiles.sh
Outdated
|
|
||
| export PYTHONPATH=$PYTHONPATH:$SOFTWARE/tcrm/$BRANCH:$SOFTWARE/tcrm/$BRANCH/Utilities | ||
|
|
||
| cd /g/data/w85/kr4383/yasi |
There was a problem hiding this comment.
Can you confirm permissions have been changed to 775 on these folders
ensemble/loss_percentiles.py
Outdated
|
|
||
| # only these filepaths need to be specified | ||
| low_res_impact_dir = os.path.join("impact", "low_res") | ||
| working_dir = "/g/data/w85/kr4383/yasi/percentile" |
There was a problem hiding this comment.
Should be a configuration option or a command line argument to the script. I usually have an
if __name__=='__main__': clause, followed by using argparse to get command line arguments for these types of variables. Means we can commit the code, run different versions and show (using the commit hash) that we're using the same code each time. If we have to change the paths in the file, might show up as multiple changes, or we end up with code like
# low_res_impact_dir = /foo/bar
low_res_impact_dir = /foo/baz
ensemble/run_percentiles.sh
Outdated
| export PYTHONPATH=$PYTHONPATH:$SOFTWARE/tcrm/$BRANCH:$SOFTWARE/tcrm/$BRANCH/Utilities | ||
|
|
||
| cd /g/data/w85/kr4383/yasi | ||
| python3 loss_percentiles.py No newline at end of file |
There was a problem hiding this comment.
Is there any STDOUT messages that could be logged to help troubleshoot problems
There was a problem hiding this comment.
yep the STDOUT from hazimp and wind multipliers is captured and sent to files
There was a problem hiding this comment.
Some comments on each file to be addressed. I prefer not to hard-code paths to input folders, rather they should be command line arguments that can be specified in the shell script that calls the python code.
See https://github.com/wcarthur/processing/blob/master/process/access/process_windgust.py and the start() function for a really comprehensive implementation. In that case, running the script defines all the functions, gets to the bottom and calls start() which handles all the command-line arguments using teh argparse module, sets up log files, then calls the functions that actually do the work. This one doesn't demand quite such a comprehensive implementation, but command line args for the paths would be helpful
Make sure that the shell script sets umask 002 so the rest of the project members can read and write to the folders used (may need to check permissions if you've not got time to run the scripts).
|
@wcarthur I think I addressed your comments if you time to look over it again |
3 participants
No description provided.