It seems that when bandpass_integration is set to True in SkyGenerationParams, the output units of the generated maps are incorrect.
Comparing a map generated for the same channel, using only the center frequency vs integrating over the band, we have a factor 1e6 between them. It happens with uK_CMB and K_CMB, so it doesn't seem like a simple conversion missing from one to the other.
I'll look into the source problem as soon as I can. Meanwhile, here is a simple script showing the problem.
UPDATE: I just realized that this happens for the CMB but not for the foregrounds, which may make it easier to find the problem.
imo = lbs.Imo(flatfile_location="schema.json")
chinfo = lbs.FreqChannelInfo.from_imo(
url=f"/releases/IMo_vPostKDP2_Option2/LMHFT/L1_L/channel_info",
imo=imo,
)
params = SkyGenerationParams(
nside=256,
bandpass_integration=True,
units=Units.uK_CMB,
output_type="map",
make_cmb=True,
seed_cmb=0,
make_fg=False,
make_dipole=False,
apply_beam=True,
)
sky_gen = SkyGenerator(
parameters=params,
channels=chinfo,
)
sky_cmb = sky_gen.execute()
params = SkyGenerationParams(
nside=256,
bandpass_integration=False,
units=Units.uK_CMB,
output_type="map",
make_cmb=True,
seed_cmb=0,
make_fg=False,
make_dipole=False,
apply_beam=True,
)
sky_gen = SkyGenerator(
parameters=params,
channels=chinfo,
)
sky_cmb_noBP = sky_gen.execute()
plt.figure()
for i in range(3):
hp.mollview(sky_cmb["L1_L"].values[i], sub=(2,3,i+1))
hp.mollview(sky_cmb_noBP["L1_L"].values[i], sub=(2,3,i+4))
plt.show()
Here is the output:

It seems that when bandpass_integration is set to True in SkyGenerationParams, the output units of the generated maps are incorrect.
Comparing a map generated for the same channel, using only the center frequency vs integrating over the band, we have a factor 1e6 between them. It happens with uK_CMB and K_CMB, so it doesn't seem like a simple conversion missing from one to the other.
I'll look into the source problem as soon as I can. Meanwhile, here is a simple script showing the problem.
UPDATE: I just realized that this happens for the CMB but not for the foregrounds, which may make it easier to find the problem.
Here is the output: