Populate the molecular cloud library! about openfermion HOT 5 OPEN

Hi Ryan,
Sorry for the late reply. Tarini ran calculations for many molecules (and I think she also included couple of molecules from alkane, alkene, and alkyne series) and has been running calculations for hydrogen lattices. She has been trying few other things. So, I can talk to her and James and we can contribute the data that makes the most sense for OpenFermion.
Also, the implementation for the cloud library looks good. It is not completely straightforward, but we understand your constraints.

from openfermion.

Over here, the test set we've been playing around with the AE6 and BH6 library (Lynch BJ, Truhlar DG (2003) J Phys Chem A 107:8996). The subset is designed to be representative of the larger Database/3 collection of atomization energies and barrier heights. As Kanav says, we've had a look at the implementation and we're discussing some of the design decision before cleaning and uploading our test cases.

from openfermion.

Sounds great James. We'd love to hear thoughts on the implementation. Keep in mind that any data bucket which allows users to upload a large data file and also download that data file will surely invite security exploits in the form of people using the system for storage or sharing of content that is surely inappropriate for OpenFermion. So any system needs to involve a secure upload where we are able to audit the data prior to transferring things over to the public facing database. We were (unfortunately) also not able to easily configure a system where such secure uploads could be received in a safe way without us manually generating signed URLs. One could setup a server to facilitate this, which is something we might consider in the future if this feature is popular. If you have ideas on better ways to download the data, or to organize it, let us know!

from openfermion.

have you decided on a quality assurance (NIST repo standard ) and a propper way to store these for benchmarking and future usage ?

where do one upload to the molecular cloud library (I assume most here would prefer a https://cloud.google.com/ EssentialmolecularQloudlibrary)

from openfermion.

I am not sure exactly what sort of "quality assurance" you have in mind but we have determined a reasonable way to store these. This is explained (along with instructions for uploading), here: https://github.com/quantumlib/OpenFermion/tree/master/cloud_library

from openfermion.