Gaussian energies #1815
Gaussian energies #1815
Conversation
Previously the gaussian log file would read the SCF energy if coupled cluster or similar methods were being used. This would increase the error in calculation. This commit allows for the correct energy to be obtained from MP2, CCSD, and CCSD(T) methods. It also adds a debug message which specifies what type of energy is being taken from the log file.
|
@dranasinghe, I would merge the first two commits, since they relate to the same effect. The tests are showing an error when reading I don't know the CBS always appears in |
Codecov Report
@@ Coverage Diff @@
## master #1815 +/- ##
=======================================
Coverage 42.99% 42.99%
=======================================
Files 80 80
Lines 21090 21090
Branches 5513 5513
=======================================
Hits 9067 9067
+ Misses 11008 10995 -13
- Partials 1015 1028 +13
Continue to review full report at Codecov.
|
| @@ -274,7 +274,7 @@ def load_energy(self, zpe_scale_factor=1.): | |||
| if 'SCF Done:' in line: | |||
| e_elect = float(line.split()[4]) * constants.E_h * constants.Na | |||
| elect_energy_source = 'SCF' | |||
| elif ' E2(' in line: | |||
| elif ' E2(' in line and ' E(' in line: | |||
There was a problem hiding this comment.
I am probably missing something. What line in ethyl_cbsqb3.log should this be parsing?
When I type grep ' E2(' ethyl_cbsqb3.log | grep ' E(', I don't find any matching lines.
The unit tests are passing, though I don't understand why.
There was a problem hiding this comment.
Yes, that how it should be. In my previous commit, I only put " E2(" and it found a match at the end of the summary gaussian print. That's why the unit test failed. " E2(" and "E(" does not appear together in the cbs-qb3 calculation and it will correctly parse the " cbs-qb3 energy" , but it will appear in the double-hybrid calculations.
|
Dear @goldmanm Mark, i made the changes you suggested. Further, i added the file previously gave the error as a test. |
Motivation or Problem
Arkane parser for Gaussian does not parse MP2, double hybrid DFT, CCSD, CCSD(T) energies. Previous to this change, HF energy was appended as e0 or energies should be given manually in the Arkane input file.
Description of Changes
CBS-QB3 calculation contains CCSD(T), MP2, and CCSD energies, therefore, new if statements were added before CBS-QB3. The elif statements were arranged in the order of appearance in the output file
Testing
unit test added
Reviewer Tips
run a Gaussian calculation and check e0