Week of May 17th

1 minute read

What I did

Monday: Energy mesh interpolation codes
Tuesday - Thursday: Felt dead from the seconds shot of the Moderna covid vaccine. Didn’t do much. I’ve thought about how to fix the energy mesh interpolation, but I’m having some issues because of how I’ve previously concatenated the dataframe. I think I’ll be able to fix it quickly when I feel better, but I’m not being very productive now.
Friday: Yeah, it took me like 10 minutes to re-write the dataframe generator function. Both datasets are now complete, using the same energy mesh, so they are ready for neural network training.

Monday: The problem right now is that when I went to averaged all 13,000 files into 1,000 structures, I found that the energy range for each varied enough that I couldn’t just provide an energy mesh used previously for the interpolation. Even using cubic-spline interpolation, which solves this problem, doesn’t work, because some simulations have 81 energy points and some have 82. This means I need to create an energy mesh that will work for all 1,000 structures and then redo my skew-norm averaged spectra as well. I might as well write a script to generalize this better, i.e. loop through all 1,0000 structures and then create a map that will work for all structures. I’ve already written the hard parts of this code. This time, however, I want to save the mesh as a csv so that I can manually edit it more easily if needed and also use it for other structures.
Friday: I had a problem trying to create a dataframe with two arrays:
```
arr_1 = [1, 2, 3, 4]
arr_2 = [5, 4, 3, 2]
df = pd.DataFrame({'A': arr_1, 'B': arr_2)
```
this returns an error, because it wants indexes. Instead, use numpy arrays for arr_1 and arr_2 and you won’t have a problem.