Closed sansona closed 5 years ago
coveralls
commented
5 years ago | Files with Coverage Reduction | New Missed Lines | % | ||
|---|---|---|---|---|
| pylj/forcefields.py | 1 | 66.0% | ||
| <!-- | Total: | 1 | --> |
| Totals | |
|---|---|
| Change from base Build 358: | 0.0% |
| Covered Lines: | 238 |
| Relevant Lines: | 338 |
coveralls
commented
5 years ago | Changes Missing Coverage | Covered Lines | Changed/Added Lines | % | ||
|---|---|---|---|---|---|
| pylj/forcefields.py | 0 | 5 | 0.0% | ||
| <!-- | Total: | 0 | 5 | 0.0% | --> |
| Totals | |
|---|---|
| Change from base Build 358: | -1.9% |
| Covered Lines: | 238 |
| Relevant Lines: | 343 |
arm61
commented
5 years ago Sorry for the delay on this (I am really busy between work and viva prep) but would it not be more straight forward to say that the forcefield functions only accept np.ndarray type objects?
sansona
commented
5 years ago No worries at all. It would definitely be. I'm not familiar enough to comment on whether users would attempt arrays as inputs - if that's not the case then I'll add that to the documentation and remove the handling for lists. The other forcefield functions can handle lists so for consistency, I see value in having all the forcefield functions be able to handle the same inputs
arm61
commented
5 years ago I have moved this over to #48 cause there was a problem with the travis continuous integration and I didn't wanna bog you down with trying to fix it. All of your commits will still be registered and I will merge that pull request once the tests pass!
Thanks again for an awesome contribution :100: :+1:
For #46
Buckingham forcefield was raising a TypeError when dr was an array. Error was in the np.exp() term which was only used in the Buckingham function but not the other forcefields:
When type(dr) == float, the term inside was fine, but multiplication of a non-int and a float i.e
2.0 * [2.0]isn't allowed in python. I fixed this by converting dr to a numpy array by default. Numpy allows this multiplication in their arrays i.e2.0 * np.array([2.0])returns 4.0. The function should now be able to handle both floats and arrays (lists and numpy arrays) fine.