stackSentinel background task fix
There is a problem with https://github.com/isce-framework/isce2/issues/223 in which any tasks running in the background will continue in the background when the next step of the process begins (eg: run_03 can bleed into run_04). This causes issues when some steps try and use incomplete data.
The proposed solution writes a final wait in any shell scripts which leverage num_process.
I welcome any feedback.
@yunjunz I would be happy to change
line_cnt = 1
configObj.write_wrapper_config2run_file(configName, line_cnt)
to
configObj.write_wrapper_config2run_file(configName)
in the three occurrences. I left as above as it might be simpler to follow.
Presently the stackSentinel.py num_process is buggy so I think good to get this merged. Apologies for the delay.
Thank you for getting back to this @RussellGrew. Could you do the following?
- Rebase this PR against the latest
mainbranch, as the stack processor has been updated and GitHub flagged conflicts. - Apply the same
enumeratestyle call to the new functions added within theclass run(object).
@yunjunz I would be happy to change
line_cnt = 1 configObj.write_wrapper_config2run_file(configName, line_cnt)to
configObj.write_wrapper_config2run_file(configName)in the three occurrences. I left as above as it might be simpler to follow.
I would prefer the latter one. We don't need to show the details if we don't have to.
@yunjunz I think that is everything.
These two still stand I believe:
- Rebase this PR against the latest
mainbranch, as the stack processor has been updated and GitHub flagged conflicts.- Apply the same
enumeratestyle call to the new functions added within theclass run(object). E.g.,filtIon(). They will be visible in this PR after rebasing.
@yunjunz see https://github.com/isce-framework/isce2/pull/271/commits/4ffbffff0b269d13e53775394cb65da4278a1b6f and https://github.com/isce-framework/isce2/pull/271/commits/8971f5292f5803b036ac3b14e5a3893d842a2bac
Though this was the first time I have attempted to update a branch for an in progress PR. Hopefully it went ok.
Edit - I see that it did not. Leave it with me.
That should be much closer.
You did this to yourself with your notes of encouragement on https://mintpy.readthedocs.io/en/latest/#5_contributing
I see the CI has now passed.
I am not familiar with these new ionosphere steps. I checked that the config is written out, and that step 17 and 18 which consider the numProcess are written correctly.
Cheers.