Aurélien Jaquier

I think the formatting issues have been solved in the `master` branch of `BlueBrain/BluePyOpt`. So a `merge branch 'master' into 'lfpy'` should solve them.

The V2::AP_end_indices just takes the 1st value to be above threshold. When we make it similar to V5::AP_begin_indices, it takes the 1st value only when all next values in a...

Could this cause any unexpected behavior in other features that depend on ISI_values? I know that it would at least in `burst_mean_freq`. Although you can ignore this particular feature because...

@DrTaDa Yes, but the user might want to set `ignore_first_spike` to `False`, and compute other features. I think that I will make `burst_mean_freq` dependent on the value of `ignore_first_spike` when...

Wait... if you want not to skip the 1st ISI, can't you just use `LibV5::all_ISI_values`?

@wvangeit I just checked, and only burst_mean_freq would need to be modified after this merge, the other features would behave as expected.

For the new features I added in the last PR to the master branch, you just have to adapt the `burst_indices` feature. It should be fairly easy, you just have...

I would like to rename ignore_first_spike into ignore_first_ISI for clarity, since it only affects the 1st ISI, and features depending on it, and not features depending on the 1st spike...

Changes made in the last commit : - renamed ignore_first_spike into ignore_first_ISI - made burst features consistent with burst_indices when ignore_first_ISI is 0 - allow the new burst suite (strict_burst_*...

It follows the algorithm given in the documentation, but that algorithm is flawed. It does not use the right burst duration, nor the right number of peaks in the burst....