nomad-coe / electronic-parsers

Apache License 2.0
18 stars 7 forks source link

problematic frequency parsing from Gaussian files #229

Closed schumannj closed 1 week ago

schumannj commented 3 weeks ago

I have 2 Gaussian files with frequency calcualtions, but the frequencies are not parsed correctly. in x_gaussian_section_frequncies the x_gaussian_frequency vector lists all frequencies twice and has some empty rows. Can you fix this? See here for the link to the data. It is shared with @ndaelman-hu and @ladinesa . Let me know if you need something else.

JosePizarro3 commented 3 weeks ago

Are these Raman calculations? Sorry, I cannot access the information on the link.

If so, I would suggest the following, depending on your deadline @schumannj: we will transfer the Gaussian to its own plugin, and then take care of defining a RamanSpectra property with the new Simulations schema. This will have the advantage that you will not need to use these quantities x_gaussian..., but they will rather be quantities properly found under data.outputs.raman_spectra.

If you need it to be fixed asap, we can take a look on the parser and do a quick patch.

@ladinesa what do you think of this? Do you prefer to fix this asap and later transfer Gaussian anyways to a plugin?

ladinesa commented 3 weeks ago

I would also put up a fix in the old schema unless @ndaelman-hu would llike to take over?

ndaelman-hu commented 2 weeks ago

Update: found the reason.

Gaussian prints the vibrational section 2x (not sure yet why). However, some quantities, e.g. reduced mass and normal modes are denoted slightly differently. The main difference is the normal mode notation.

Our parser cannot handle the first output format, though the Gaussian schema did distinguish between these 2 headers and sections. I've removed this now

ndaelman-hu commented 2 weeks ago

@schumannj We aren't extracting the (harmonic) force constants nor the IR intensities. Do you need these?

schumannj commented 2 weeks ago

That would be great, thank you!

ndaelman-hu commented 2 weeks ago

@schumannj have added IR intensities and Harmonic force constants.

ndaelman-hu commented 2 weeks ago

closed with #232 @schumannj you can update parser version in the main

ndaelman-hu commented 1 week ago

Found a small error with the unit conversion that had some Gaussian tests fail (hard to tell from the verbose list with warnings). Will patch this quickly.

ndaelman-hu commented 1 week ago

Patched