General class for N body grid integrations #191
Conversation
There was a problem hiding this comment.
It looks good to me. I'd let @Ali-Tehrani or @FarnazH review it before merging, however.
Can you add a tutorial notebook showing how to integrate the k-electron distribution function from Hartree-Fock (a Slater determinant) with the QC-Devs packages? That notebook could be added in the Horton3 repository, which is where most notebooks that use multiple packages go.
|
Thanks, @PaulWAyers for your feedback. I will certainly make the tutorial. Before that, I plan to implement the porcelain needed for the class so that the obtained grid (used for representing the k-electron distribution function) can be used for calculating the coulomb type integrals seamlessly (without having undefined values for any of the points). |
|
At first I'd just compute normalization of the distribution functions. Then we can work on coulomb integrals in a second step. |
|
Perfect, I will do so then. Thank you very much. |
|
@sanchezw17 The Basegrid should not be changed to accommodate the Ngrid class. I will work on an example and see if I can reproduce your problem. |
The function to integrate now has few requirements. It has to be a function of n variables where n is the number of grids, all input parameters must have the same dimensions as the corresponding grid points. Added two tests too.
|
@sanchezw17 I made the corrections to the ngrid and added some tests. Now It is possible to use grids of different dimensions for the same function. At some point an example notebook will be needed but for now, you can see the tests. Please let me know if you have any problem after the corrections. |
Added module and class for generic N grid integrations. This class is the base framework and on top of it, some porcelain will be added for the coulomb-exchange integrals.