diff options
| -rw-r--r-- | ml_exp/representations.py | 23 |
1 files changed, 6 insertions, 17 deletions
diff --git a/ml_exp/representations.py b/ml_exp/representations.py index 86e75eb31..51ed309b9 100644 --- a/ml_exp/representations.py +++ b/ml_exp/representations.py @@ -53,7 +53,7 @@ size. Arrays are not of the right shape.') 'instead of (size).') size = n - cm = np.zeros((size, size), dtype=np.float64) + cm = np.zeros((n, n), dtype=np.float64) # Actual calculation of the coulomb matrix. for i in range(n): @@ -65,27 +65,16 @@ size. Arrays are not of the right shape.') rv = coords[i + 1:] - coords[i] r = np.linalg.norm(rv, axis=1)/cr val = nc[i]*nc[i +1:]/r - cm[i, i + 1:n] = val - cm[i + 1:n, i] = val + cm[i, i + 1:] = val + cm[i + 1:, i] = val # Now the value will be returned. if as_eig: - cm_sorted = np.sort(np.linalg.eig(cm)[0])[::-1] - # Thanks to SO for the following lines of code. - # https://stackoverflow.com/a/43011036 - - # Keep zeros at the end. - mask = cm_sorted != 0. - f_mask = mask.sum(0, keepdims=1) >\ - np.arange(cm_sorted.shape[0]-1, -1, -1) - - f_mask = f_mask[::-1] - cm_sorted[f_mask] = cm_sorted[mask] - cm_sorted[~f_mask] = 0. + cm_eigs = np.sort(np.linalg.eig(cm)[0])[::-1] - return cm_sorted + return np.pad(cm_eigs, (0, size - n), 'constant') else: - return cm + return np.pad(cm, ((0, size - n), (0, size - n)), 'constant') def lennard_jones_matrix(coords, |