Skip to content

Add Pk for cobaya wrapper #6

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
itrharrison wants to merge 1 commit into
base: demo-fixes
Choose a base branch
from
Draft

Add Pk for cobaya wrapper #6

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 59 additions & 3 deletions cosmopower/wrappers/cobaya/cosmopower.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,9 @@ def initialize(self) -> None:
for nw in self._networks:
self._params_required_by_networks.update({str(p) : None for p in self._networks[nw].parameters})

# z is a special case (not actually a sampled parameter)
del(self._params_required_by_networks['z'])

def must_provide(self, **requirements: dict) -> dict:
super().must_provide(**requirements)

Expand All @@ -51,6 +54,16 @@ def must_provide(self, **requirements: dict) -> dict:
if k == "Cl":
for a in v:
required_quantities.append(f"Cl/{a.lower()}")

elif "Pk_interpolator" in requirements.keys():
v = self._must_provide[('Pk_grid', True, 'delta_tot', 'delta_tot')]
self.pk_redshifts = v.pop("z")
self.pk_k_max = v.pop("k_max")
self.pk_nonlin = v.pop("nonlinear", True)
if self.pk_nonlin:
required_quantities.append("Pk/nonlin")
else:
required_quantities.append("Pk/lin")
else:
required_quantities.append(f"{k}" if v is not None else k)

Expand Down Expand Up @@ -114,6 +127,9 @@ def must_provide(self, **requirements: dict) -> dict:

self.log.debug(f"Will evaluate networks {self._networks_to_eval}")

# z is a special case (not actually a sampled parameter)
del(must_provide['z'])

return must_provide

def calculate(self, state: dict, want_derived: bool = True,
Expand Down Expand Up @@ -155,12 +171,44 @@ def calculate(self, state: dict, want_derived: bool = True,

network = self.networks[quantity]

used_params = {p: input_params[p] for p in network.parameters}
# if 'z' in network.parameters:
# input_params['z'] = np.asarray(self.pk_redshifts)

# used_params = {p: input_params[p] for p in network.parameters}

used_params = {}

for p in network.parameters:
if p == 'z':
continue # 'z' is treated as an emulated but not sampled parameter
else:
used_params[p] = input_params[p]

if self.parser.is_log(quantity):
data = network.ten_to_predictions_np(used_params)[0, :]
get_data = network.ten_to_predictions_np
else:
data = network.predictions_np(used_params)[0, :]
get_data = network.predictions_np

if 'z' in network.parameters:

data = np.empty([len(np.asarray(self.pk_redshifts)), len(network.modes)])

for iz, z_eval in enumerate(self.pk_redshifts):
used_params = {**used_params, **{'z': z_eval}}
data[iz] = get_data(used_params)[0, :]

state["z"] = self.pk_redshifts
state[('Pk_grid', True, 'delta_tot', 'delta_tot')] = data

else:
data = get_data(used_params)[0, :]

import pdb; pdb.set_trace()

# if self.parser.is_log(quantity):
# data = network.ten_to_predictions_np(used_params)[0, :]
# else:
# data = network.predictions_np(used_params)[0, :]

self.set_in_dict(state, quantity, data)
if self.parser.modes_label(quantity) == "l":
Expand Down Expand Up @@ -254,6 +302,14 @@ def get_fsigma8(self, z: Union[float, np.ndarray]) -> np.ndarray:
compute fsigma8, but neither are \
computed.")

def get_Pk_grid(self, var_pair=("delta_tot", "delta_tot"), nonlinear=True):

k = self.current_state["k"]
z = self.current_state["z"]
Pk = self.current_state[('Pk_grid', True, 'delta_tot', 'delta_tot')]

return k, z, Pk

def get_Cl(self, ell_factor: bool = False,
units: str = "FIRASmuK2") -> dict:
cls_old = self.current_state.copy()
Expand Down