LSSTDESC / augur

DESC forecasting and inference validation tool
MIT License
3 stars 2 forks source link

added stem method #65

Open fjaviersanchez opened 2 weeks ago

fjaviersanchez commented 2 weeks ago

This PR adds the stem derivative method as implemented by @nikosarcevic. For now, it is pretty barebones/non-optimized. Will be working on improving speed in the upcoming days.

fjaviersanchez commented 2 weeks ago

Will need to update the environment and install instructions to make this work. For reference, how I use it locally:

git clone git@github.com:nikosarcevic/Derivative-Calculator.git
ipython
import sys
sys.path.append('PATH-to-Derivative-Calculator')
from augur.analyze import Analyze
ao = Analyze('./examples/config_test.yaml')
ao.get_fisher_matrix(method='stem')

Note that this is currently very slow with the default number of parameters (12+ hours in my laptop).

arthurmloureiro commented 2 weeks ago

Will need to update the environment and install instructions to make this work. For reference, how I use it locally:

git clone git@github.com:nikosarcevic/Derivative-Calculator.git
ipython
import sys
sys.path.append('PATH-to-Derivative-Calculator')
from augur.analyze import Analyze
ao = Analyze('./examples/config_test.yaml')
ao.get_fisher_matrix(method='stem')

Note that this is currently very slow with the default number of parameters (12+ hours in my laptop).

Hi @fjaviersanchez ! I think we should avoid having to import more paths into augur.

Maybe we can ask @nikosarcevic if this derivative calculations can be packaged into pip or at least have a pip installable option?* Or maybe if @nikosarcevic can include her derivatives into augur?