htjense/pywmap

A WMAP likelihood implementation for cobaya.

pywmap's Issues

use_SZ causes crash when CLs are calculated to a value too high

Enabling the SZ template can cause an error if the Cls are calculated to an L value that is higher than 10^4, because the SZ template only goes up to that value.

This should be a simple fix where the code just checks that the highest L value does not exceed the size of the template.

Chi-square value different from the README.md when running the file `eval_wmap.yaml`

Dear people,

First, thank you for putting together this nice WMAP interface, super useful!

I just have a usage question: in the README.md you mentioned the file eval_wmap.yaml should give a chi-square value of 5923.52.. However, I run it as it is and got a different chi2: 6250.75. Should I be worried about that?

Below are the output messages for the run.

> cobaya-run cosmo-packages/code/wmaplike/eval_wmap.yaml -f

[output] Output to be read-from/written-into folder 'chains', with prefix 'eval_wmap'
[output] Found existing info files with the requested output prefix: 'chains/eval_wmap'
[output] Will delete previous products ('force' was requested).
[camb] `camb` module loaded successfully from /home/uendertandrade/mgclass_cobaya/cosmo-packages/code/CAMB/camb
[wmaplike.wmaplike] Loaded SZ.
[wmaplike.wmaplike] First evaluation: Calculating largest term...
[wmaplike.wmaplike] Done finding largest term [-759.901].
[wmaplike.wmaplike] Initialized TT gibbs.
[wmaplike.wmaplike] *WARNING* Note: you have disabled the low-ell polarisation functions, but you did not include the data. If you intend to replace the low-ell polarisation data by a tau prior, you should include `temin: 24` as a setting.
[wmaplike.wmaplike] Initialized high-l TT/TE.
[evaluate] *WARNING* No sampled parameters requested! This will fail for non-mock samplers.
[evaluate] Initialized!
[evaluate] Looking for a reference point with non-zero prior.
[evaluate] Reference point:

[evaluate] Evaluating prior and likelihoods...
/home/uendertandrade/mgclass_cobaya/cosmo-packages/code/wmaplike/wmaplike/wmap.py:540: RuntimeWarning: invalid value encountered in divide
  fisher = -self.r_off_tttt / np.sqrt(TTTT * TTTT[...,np.newaxis]) + self.epsilon / (TTTT * TTTT[...,np.newaxis])
/home/uendertandrade/mgclass_cobaya/cosmo-packages/code/wmaplike/wmaplike/wmap.py:626: RuntimeWarning: invalid value encountered in divide
  fisher = -self.r_off_tete / np.sqrt(TETE * TETE[...,np.newaxis])
lowl_TT_gibbs -3.333604636169003
MASTER_TTTT 905.902147796617
MASTER_TETE_chi2 395.03459164972844
MASTER_TETE_det 1827.7720316702637
[evaluate] log-posterior  = -3125.38
[evaluate] log-prior      = 0
[evaluate]    logprior_0 = 0
[evaluate] log-likelihood = -3125.38
[evaluate]    chi2_wmaplike.WMAPLike = 6250.75
[evaluate] Derived params:

Thank you very much,
Uendert