GitLab

Add the sfhdelayedbq module implementing a delayed SFH with a burst/quench. It is described in detail in Ciesla et al. (2017).

Add the possibility of computing emission lines through special filters when underlying absorption lines are not subtracted.

Completely change the SED cache. Now we rely on the fact that for a given set of modules we only need to store only one model in cache. If there is a cache miss, it means we can discard this one model. In turn the cache key simply becomes the number of modules used, making it easier and faster to access. To check whether the stored module corresponds to the requested one, we store both the model parameters and the SED in a tuple.

Avoid making a copy when storing the redshifted SED. This SED will never be modified. This is done by copying what is returned by get_sed rather than having get_sed return directly a copy.

It turns out that elevating 10 to some power is very slow. Because the attenuation curve will not change between two calls, we save its shape for the young and old ages rather than recomputing it every single time. This leads to a massive speedup.

The counter can become a bottleneck when many parallel processes are running concurrently because of the lock to update it. To address this we implement a Counter class that only updates the counter periodically rather than at each iteration. The number of models actually computed at any moment is kept in a local counter.

Make the cache faster by avoiding a serialisation with marshel. Rather let's use tuples of of the values of the parameters. This works because they are stored in ordered dictionaries.

Start the parallel processes with fork rather than spawn on Linux platforms, resulting in a faster startup.

Add the possibility to have an exponential burst with the sfhdelayed module in order to model the latest episode of star formation.

When using different line widths during a single run, ensure that the fluxes and other quantities are always computed correctly

Ensure that the computation of the continuum level is correct when determining the equivalent width, in particular when the line width is very narrow.

The formula from Sawicki et al. (2012) used to compute the χ² in the presence of upper limits was not correct. This led the χ² to depend directly on the absolute value of the upper limit. The formula has been rederived and corrected.

Fν was computed by calculating Fλ and then converting to Fν, which led to typical differences in fluxes of typically 1-2% and a bit more for a handful of pathological filters. Now Fν is computed directly and a bit faster. At the same time rather than computing an effective wavelength (which was more a mean wavelength), we compute the pivot wavelength, which is more appropriate when converting between Fν and Fλ.

When computing the attenuation curve, the bump is now added so that its relative strength does not depend on δ.