Upgrading mflike to v0.8.4 in LAT_MFLike #133

sgiardie · 2023-05-23T16:11:04Z

As the title suggests, updating the mflike version to the latest simonsobs/LAT_MFLike one (so far).
Several changes to be made also to the Foregrounds and BandPass classes, in order to substitute lists of frequencies with list of experiment_frequency, useful when using more experiments or more arrays sharing the same freq.

Also, some outdated functions in bandpass.py will be removed and the bug related to transmission in the presence of bandpass shift will be solved.

More details in this LAT_MFLike pull request.

Solving issue #130

sgiardie · 2023-06-01T11:03:29Z

I have set up the new mflike version and changed bandpass.py and foreground.py accordingly.
As already said, the main change in mflike is the fact that list of the nominal frequency of each channel is substituted by a list of the channel names experiment_freq. This is implemented also in theoryforge_MFLike.py, bandpass.py and foreground.py.
The default is now that the passbands for each channel are read from the sacc file used as data for mflike. These are then passed to bandpass.py from theoryforge_MFLike.py --> foreground.py inside a bands dictionary, to allow the computation of the bandpass transmissions then used to integrate the foreground spectra.

To allow the Theoryforge_MFLike and Foreground classes to work as standalone, I have also defined the variables exp_ch and eff_freqs in the respective yamls. This allows the construction of the bands dictionary which would normally be passed by MFLike. In this case, the passbands would be just Dirac delta at the effective frequency.
The Banpass class still requires to be called by Foreground, which passes the bands dictionary. If we would like the _bandpass_construction function to be called initializing only Banpass, we should define the exp_ch and eff_freqs variables also in the Banpass.yaml.

The tests and documentations for MFLike, Foreground and Bandpass have been fixed.

sgiardie · 2023-06-01T11:15:31Z

Regarding the passband construction, there are now three options to be selected in the BandPass.yaml:

reading the passbands stored in the bands dictionary (passed from mflike --> theoryforge --> foreground) and multiply them for the appropriate conversion factors
building a top-hat passband using the nsteps and bandwidth params. This passband can also be a Dirac delta for nsteps = 1 and bandwidth = 0. The effective frequency used comes again from the bands dictionary
reading the passbands stored in an external file, which should have exactly the name of the corresponding array in the format exp_ch (with no other extensions, e.g. LAT_93)
Also in the second and third case the appropriate dB/dT factors are applied to the passbands.

The new test_bandpass tests all of the options

soliket/BandPass.yaml

soliket/bandpass.py

soliket/foreground.py

mgerbino · 2023-06-05T16:05:07Z

soliket/mflike/MFLike.yaml

-    TE: [2, 5000]
-    ET: [2, 5000]
-    EE: [2, 5000]
+    TT: [50, 5000]


is 5000 the correct lmx for data in yaml?

not sure, I just used what was left there

ok! This might not be relevant for the current release, but maybe good to check with PS whether this is the correct lmax for data cut

soliket/mflike/TheoryForge_MFLike.yaml

mgerbino · 2023-06-05T16:06:30Z

soliket/mflike/mflike.py

@@ -37,7 +38,7 @@

 class MFLike(GaussianLikelihood, InstallableLikelihood):
    _url = "https://portal.nersc.gov/cfs/sobs/users/MFLike_data"
-    _release = "v0.6"
+    _release = "v0.7.1"


is this correct? shouldn't it be v0.8.3?

I left it like in LAT_MFLike, maybe there is no v0.8 folder on nersc

Ok! Again, it would be nice to check with PS people. I think it makes sense to update the data versioning to reflect the code versioning

codecov-commenter · 2023-06-06T10:02:32Z

Codecov Report

Merging #133 (83f673c) into master (381675b) will increase coverage by 2.29%.
The diff coverage is 83.68%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #133      +/-   ##
==========================================
+ Coverage   73.12%   75.42%   +2.29%     
==========================================
  Files          30       30              
  Lines        1935     1961      +26     
==========================================
+ Hits         1415     1479      +64     
+ Misses        520      482      -38

Impacted Files	Coverage Δ
soliket/foreground.py	`92.30% <75.00%> (-3.25%)`	⬇️
soliket/mflike/theoryforge_MFLike.py	`83.67% <75.00%> (-2.54%)`	⬇️
soliket/bandpass.py	`91.26% <87.69%> (+46.36%)`	⬆️
soliket/mflike/mflike.py	`81.00% <91.66%> (+0.11%)`	⬆️

mgerbino · 2023-06-06T10:46:08Z

soliket/mflike/MFLike.yaml

-    TE: [2, 5000]
-    ET: [2, 5000]
-    EE: [2, 5000]
+    TT: [50, 5000]


ok! This might not be relevant for the current release, but maybe good to check with PS whether this is the correct lmax for data cut

mgerbino · 2023-06-06T10:47:23Z

soliket/mflike/mflike.py

@@ -37,7 +38,7 @@

 class MFLike(GaussianLikelihood, InstallableLikelihood):
    _url = "https://portal.nersc.gov/cfs/sobs/users/MFLike_data"
-    _release = "v0.6"
+    _release = "v0.7.1"


Ok! Again, it would be nice to check with PS people. I think it makes sense to update the data versioning to reflect the code versioning

mgerbino · 2023-06-06T10:49:57Z

Closing this PR will supersede issue #90

soliket/mflike/mflike.py

itrharrison

lgtm!

* first modifications to switch from freqs to exp_freqs * fix parenthesys * fixing codestyle * more modifications in bandpass and foreground * more modifications and docs * updating mflike and theoryforge * fixing style * fixing bandpass test * fixing text foreground * adding test data for bandpass * fix bandpass docs * fixing mflike test and bandpass docs * fixing conflict with master on index.rst * fix path in test_bandpass * fixing problem in test_bandpass for tox tests * addressing comments * Update soliket/mflike/mflike.py --------- Co-authored-by: Ian Harrison <itrharrison@gmail.com>

sgiardie added 3 commits May 23, 2023 17:09

first modifications to switch from freqs to exp_freqs

6d6c94a

fix parenthesys

8ea303d

fixing codestyle

3d7c0ab

sgiardie requested review from itrharrison and mgerbino May 23, 2023 16:16

sgiardie added the mflike Related to mflike likelihood label May 24, 2023

sgiardie removed request for itrharrison and mgerbino May 24, 2023 10:20

sgiardie marked this pull request as draft May 24, 2023 10:20

itrharrison assigned sgiardie May 24, 2023

sgiardie added the enhancement New feature or request label May 24, 2023

sgiardie added 4 commits May 24, 2023 17:37

more modifications in bandpass and foreground

44b9c14

more modifications and docs

040031c

updating mflike and theoryforge

c9b72ae

fixing style

d014554

sgiardie changed the title ~~Upgrading mflike to v0.8.3 in LAT_MFLike~~ Upgrading mflike to v0.8.4 in LAT_MFLike May 30, 2023

sgiardie added 6 commits May 30, 2023 17:13

fixing bandpass test

91a7d02

fixing text foreground

e195b99

adding test data for bandpass

0a305ee

fix bandpass docs

fb91cb0

fixing mflike test and bandpass docs

2c5a9dd

fixing conflict with master on index.rst

0b4a0cc

sgiardie marked this pull request as ready for review June 1, 2023 10:46

Merge branch 'master' into upgrade_mflike

d8188ae

sgiardie requested review from mgerbino and itrharrison June 1, 2023 11:03

sgiardie added 2 commits June 1, 2023 13:04

fix path in test_bandpass

e5971cd

fixing problem in test_bandpass for tox tests

003d6d9

mgerbino requested changes Jun 5, 2023

View reviewed changes

mgerbino added this to the v0.1 milestone Jun 5, 2023

addressing comments

83f673c

sgiardie requested a review from mgerbino June 6, 2023 10:08

mgerbino previously approved these changes Jun 6, 2023

View reviewed changes

This was linked to issues Jun 6, 2023

Freeze SOLikeT MFLike to v0.7 from LAT_MFLike #90

Closed

Upgrade mflike version to v0.8.4 from LAT_MFLike #130

Closed

itrharrison reviewed Jun 6, 2023

View reviewed changes

soliket/mflike/mflike.py Outdated Show resolved Hide resolved

Update soliket/mflike/mflike.py

52a665a

itrharrison dismissed mgerbino’s stale review via 52a665a June 6, 2023 19:12

itrharrison approved these changes Jun 6, 2023

View reviewed changes

itrharrison merged commit cd28f57 into master Jun 6, 2023

itrharrison deleted the upgrade_mflike branch June 6, 2023 19:44

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Upgrading mflike to v0.8.4 in LAT_MFLike #133

Upgrading mflike to v0.8.4 in LAT_MFLike #133

sgiardie commented May 23, 2023 •

edited

Loading

sgiardie commented Jun 1, 2023 •

edited

Loading

sgiardie commented Jun 1, 2023 •

edited

Loading

mgerbino Jun 5, 2023

sgiardie Jun 6, 2023

mgerbino Jun 6, 2023

mgerbino Jun 5, 2023

sgiardie Jun 6, 2023

mgerbino Jun 6, 2023

codecov-commenter commented Jun 6, 2023 •

edited

Loading

mgerbino Jun 6, 2023

mgerbino Jun 6, 2023

mgerbino commented Jun 6, 2023

itrharrison left a comment

Upgrading mflike to v0.8.4 in LAT_MFLike #133

Upgrading mflike to v0.8.4 in LAT_MFLike #133

Conversation

sgiardie commented May 23, 2023 • edited Loading

sgiardie commented Jun 1, 2023 • edited Loading

sgiardie commented Jun 1, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov-commenter commented Jun 6, 2023 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mgerbino commented Jun 6, 2023

itrharrison left a comment

Choose a reason for hiding this comment

sgiardie commented May 23, 2023 •

edited

Loading

sgiardie commented Jun 1, 2023 •

edited

Loading

sgiardie commented Jun 1, 2023 •

edited

Loading

codecov-commenter commented Jun 6, 2023 •

edited

Loading