Fix NLTE Normalization - issue #784 #831
Conversation
| @@ -190,7 +190,7 @@ def _main_nlte_calculation( | |||
| else: | |||
| raise e | |||
| general_level_boltzmann_factor[i].ix[species] = \ | |||
| level_boltzmann_factor | |||
| level_boltzmann_factor * g[species][0] / level_boltzmann_factor[0] | |||
There was a problem hiding this comment.
@livnehra can you make this g.loc[species, 0]
| @@ -190,7 +190,7 @@ def _main_nlte_calculation( | |||
| else: | |||
| raise e | |||
| general_level_boltzmann_factor[i].ix[species] = \ | |||
| level_boltzmann_factor * g[species][0] / level_boltzmann_factor[0] | |||
| level_boltzmann_factor * g.loc[species][0] / level_boltzmann_factor[0] | |||
There was a problem hiding this comment.
That doesn't work...
There was a problem hiding this comment.
hmm - that's interesting. I'll check it tomorrow.
There was a problem hiding this comment.
yes you're right - I screwed things up in my mind. I think the way you do it now is good.
There was a problem hiding this comment.
Remember, for example species=(14,1) for Si II
|
@livnehra the tests fail - your change somehow breaks the code - have a look at the continuous integration to see what's wrong. |
|
Hopefully should pass the tests now. |
|
So the fails now are due to NLTE results changing, which makes sense if the previous results were obtained with the erroneous g_ratio_matrices. |
|
@livnehra - okay I'm ready to update the reference data. I still think it would be good to get a spectrum beforehand. |
|
@wkerzendorf, @livnehra - what's the idea about proceeding with this PR? |
|
@livnehra - we haven't forgotten about you. We need regenerate the reference data and then we'll merge. |
| @@ -442,7 +442,7 @@ def _create_collision_coefficient_matrix(self): | |||
| C_ul_matrix[level_number_lower, level_number_upper, :] = line.values[2:] | |||
| delta_E_matrix[level_number_lower, level_number_upper] = line['delta_e'] | |||
| #TODO TARDISATOMIC fix change the g_ratio to be the otherway round - I flip them now here. | |||
There was a problem hiding this comment.
It would be possible to remove the TODO.
Fix NLTE normalization according to @ssim suggestion in the discussion in #784