Closed JensWehner closed 3 years ago
@votca-bot format
@votca-bot changelog: overhaul complete option handling
I have started with doing a simple dftgwbse
. When I generate the options with
xtp_tools -p dftgwbse -o dftgwbse.xml
I get the following output
Writing options for calculator dftgwbse to dftgwbse.xml
nothing to be done - stopping here
Which reads like an error message (I know it isn't) which is confusing, because it had something to do and it even did it.
I don't know how much control there is over the ordering of the options in the -p
generated xml files, but it is confusing that some options are so far from one another. Two examples the package name and package executable are separated by 5 lines and the functional and basisset options are not close either.
It also seems that there are still some options that could be hidden from users completely like the
temp
, that could be hard coded I think, the name of the file shouldn't matter it is temporary)options
When you use orca to do the dft calculation it crashes now. It can't parse the functional.
To reproduce use the defaults and run a dftgwbse calculation on any molecule.
Thanks. I will fix the orca problem.
Concerning hiding the options, they have sensible defaults now, so the user does not have to modify them and they are effectively hidden. We could hardcode them, but I would do that in a new PR later down the line.
concerning the nothing to be done - stopping here
you are right, but it is in tools::application and I do not want change that at the moment, because I do not want to touch csg as well. I agree with you in principle.
It also seems that there are still some options that could be hidden from users completely like the
- scratch
- cleanup and
- temporary files (this is just an option that names a temporary file
temp
, that could be hard coded I think, the name of the file shouldn't matter it is temporary)options
Regarding the scratch
and temp
options: if I am not mistaken, they were mainly used on cluster machines to move some IO intensive steps (like orca's scratch files) from network drives to fast local disks, no? Why do we have two?
What do you mean with
Why do we have two?
? Scratch is the directory and temporary file is the filename used for the files. We could isolate all of this if we remove the input,dft,parse
and turn it into dft
and encapsulate the qmpackages
much more.
I wasn't sure anymore that scratch
-> directory and temp
-> filename.
@JensWehner can you fix the bad merge in your latest commit and then merge this in master?
@votca-bot format
I cannot merge it into master yet, because for that we have to fix the documentation.
@votca-bot format
@votca-bot format
Merging #704 (fd5288c) into master (b677c32) will increase coverage by
0.2%
. The diff coverage is28.7%
.
@@ Coverage Diff @@
## master #704 +/- ##
========================================
+ Coverage 51.4% 51.6% +0.2%
========================================
Files 293 289 -4
Lines 27229 27060 -169
========================================
- Hits 14006 13988 -18
+ Misses 13223 13072 -151
Impacted Files | Coverage Δ | |
---|---|---|
include/votca/xtp/bsecoupling.h | 100.0% <ø> (ø) |
|
include/votca/xtp/dftengine.h | 100.0% <ø> (ø) |
|
include/votca/xtp/esp2multipole.h | 0.0% <ø> (ø) |
|
include/votca/xtp/gwbseengine.h | 0.0% <ø> (ø) |
|
include/votca/xtp/jobcalculator.h | 0.0% <0.0%> (ø) |
|
include/votca/xtp/jobtopology.h | 22.2% <ø> (ø) |
|
include/votca/xtp/parallelxjobcalc.h | 0.0% <ø> (ø) |
|
include/votca/xtp/polarregion.h | 0.0% <0.0%> (ø) |
|
include/votca/xtp/qmcalculator.h | 0.0% <0.0%> (ø) |
|
include/votca/xtp/qmpackage.h | 31.5% <0.0%> (+1.5%) |
:arrow_up: |
... and 83 more |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update b677c32...fd5288c. Read the comment docs.
The beginning of the options refactor.
closes #704