search

SunnySuite / Sunny.jl

Spin dynamics and generalization to SU(N) coherent states
Other
86 stars 19 forks source link

Wavevectors back to RLU #123

Closed kbarros closed 1 year ago

kbarros commented 1 year ago

This is a work in progress. Tests are passing, but I think @Lazersmoke needs to update binning code.

This won't be ready until the generated docs for fei2_classical.jl and binning_tutorial.jl look good.

Good news: The examples are now serving as nice tests!

kbarros commented 1 year ago

Note that LSWT now uses q_reshaped internally, which can be our naming convention for RLU for the reshaped supercell. This is what @Hao-Phys requested.

kbarros commented 1 year ago

Other updates I think we should include with this PR (help requested from @Lazersmoke and @ddahlbom):

  1. Adopt a consistent naming convention throughout Sunny to indicate the units of a wavevector. Previously @ddahlbom was distinguishing q and k, but I think that's too subtle. One possibility (started in this PR) is q for a wavevector in original RLU, q_reshaped for a wavevector in RLU for the reshaped crystal, and q_absolute for a wavevector in absolute units. We need all three because: (a) LSWT and SampledCorrelations want to work with q_reshaped internally, (b) Sometimes we need q_absolute, e.g., for calculating the form factor, and (c) The user interface is entirely in q, i.e. the original RLU.

  2. I have renamed connected_path to reciprocal_space_path, which reflects the more specific nature of this function. We will need some analogous renaming of connected_path_bins.