Ssc terasawa#991
Conversation
Add the function Tk3D_SSC_Terasawa22 for computing super-sample covariance. Also add some module to import.
carlosggarcia
left a comment
carlosggarcia
left a comment
There was a problem hiding this comment.
Great work. It looks good to me (modulo the lint errors that GitHub is complaining about). Just a few minor comments.
In addition, it'd be great if you can add an unit test for this function (see https://github.com/LSSTDESC/CCL/blob/master/pyccl/tests/test_tkkssc.py for reference). If you were comparing this Tk with some other code, you could also add some benchmark (see e.g. https://github.com/LSSTDESC/CCL/blob/master/benchmarks/test_covariances.py). Or maybe with the simulations that you have? What do you think @damonge?
|
|
||
| .. math:: | ||
| \\frac{\\partial P_{mm}(k)}{\\partial\\delta_L} = | ||
| \\left(1 + \\frac{26}{21}T_{h}(k) -\\frac{1}{3}\\frac{d\\log P_{mm}(k)}{d\\log k}\\right) |
There was a problem hiding this comment.
I had a quick look at Terasawa+2022 and couldn't find where the -1/3... term is coming from. If it's easy to derive or I have missed it, it'd be great to point to the Eq where it is (or that you have to use to get it) for future reference.
There was a problem hiding this comment.
The power spectrum response to the background density is consist of two components: growth response and dilation response (and change in the reference density). Terasawa+2022 focused on the growth response and the dilation term is simply the derivative of the power spectrum with respect to the wave number k. The total power spectrum response (including growth and dilation term) is found in Eq. 46 and 47 of Li, Hu, and Takada 2014a (https://arxiv.org/abs/1401.0385), for example.
| extra_parameters=extra_parameters) | ||
|
|
||
| # Growth factor | ||
| Dp = background.growth_factor_unnorm(cosmo_hp,a_arr) |
There was a problem hiding this comment.
You can remove the background dependency and use directly cosmo_hp.growth_factor_unnorm(a_arr). Same in the next line
There was a problem hiding this comment.
Thank you for pointing this out. I have removed the background dependency as you mentioned.
|
Yes, besides the unit tests Carlos mentions, it'd be important if you can add any type of validation you've done (e.g. against simulations) as a benchmark test. |
|
I am going to add a unit test and a benchmark test later. |
Update tk3d.py
Update tk3d.py
Lint errors correction
Update tk3d.py
Lint-check for SSC-Terasawa
Pull Request Test Coverage Report for Build 4290191698
💛 - Coveralls |
|
The changes in this PR is now included in |
5 participants
I have implemented a function Tk3D_SSC_Terasawa22 to compute super-sample covariance into the tk3d.py.