Code

Bivaraite with testing#41

Code

Merged

danOSU merged 62 commits intodevelopbandframework/Taweret:developfrom bivaraite_with_testingbandframework/Taweret:bivaraite_with_testingCopy head branch name to clipboardAug 14, 2023

+4,082-1,866

Commits

Commits on Jan 20, 2023

• mixing jetscape models

  danOSU

  committed
  b068578

  Copy full SHA for b068578

• adding jetscape files as they are taken from original repo

  danOSU

  committed
  27e247d

  Copy full SHA for 27e247d

• bash script to set workdir

  danOSU

  committed
  6dbcb6f

  Copy full SHA for 6dbcb6f

• change where exp data is

  danOSU

  committed
  228e5f3

  Copy full SHA for 228e5f3

• get rid of hetroscadastic stuff

  danOSU

  committed
  d7abcdf

  Copy full SHA for d7abcdf

Commits on Jan 22, 2023

• changes made to original jetscape sims repo, mostly path variables

  danOSU

  committed
  b5fb8ca

  Copy full SHA for b5fb8ca

Commits on Jan 25, 2023

• new mixing method, improved version of addstep which allow non symetrical combinations of step functions

  danOSU

  committed
  3a0be57

  Copy full SHA for 3a0be57

• added cache functionality to evaluate emulators for MAP values. Hopefully this would increase the speed

  danOSU

  committed
  7ed893e

  Copy full SHA for 7ed893e

• cache method didn't work. Trying a new way of calculating emu output at the initialization of jetscape methods and then using it throughout. Seems to work. 3 times speed up

  danOSU

  committed
  347af45

  Copy full SHA for 347af45

Commits on Jan 31, 2023

• introduce new parameter same_params. So that when calibration + learning of weights if two models have same parameters it will only infer the common set of parameters instead of treating them as en…

  Show description for e5a8569
  danOSU

  committed
  e5a8569

  Copy full SHA for e5a8569

• add model priors, There were negative values in the variance for v22 and v42, They happen for very small values. I enforce taking the abs value hence of the varaince before taking the sqrt to find sd

  danOSU

  committed
  e76bb8b

  Copy full SHA for e76bb8b

• If same paramters flag is on this will only take as input the common set of model paramters instead of two times that

  danOSU

  committed
  36aec59

  Copy full SHA for 36aec59

• calibration + learning of weights for jetscape grad and ptb models

  danOSU

  committed
  7fb1839

  Copy full SHA for 7fb1839

Commits on Feb 10, 2023

• Update environment.yml

  danOSU

  authored
  69dfa17

  Copy full SHA for 69dfa17

Commits on Mar 21, 2023

• notebook with grad ptb mixing results

  danOSU

  committed
  d130c09

  Copy full SHA for d130c09

• calibrate either grad or the ptb model

  danOSU

  committed
  245392b

  Copy full SHA for 245392b

• python script for grad ptb local mixing

  danOSU

  committed
  dcc5b37

  Copy full SHA for dcc5b37

• optional script to find MAP value in the local mixing

  danOSU

  committed
  2e94ddf

  Copy full SHA for 2e94ddf

• final likelihood is somehow created using a product instead of a log sum. We might be able to modify this to include error correlations

  danOSU

  committed
  e8bd3b9

  Copy full SHA for e8bd3b9

• calibrate only one of the models and not the other

  danOSU

  committed
  fabeabf

  Copy full SHA for fabeabf

• plotting scripts from VAH

  danOSU

  committed
  3e0e099

  Copy full SHA for 3e0e099

• variable set for submiting into oSC

  danOSU

  committed
  8338616

  Copy full SHA for 8338616

• only calibrate one model

  danOSU

  committed
  a1660f7

  Copy full SHA for a1660f7

• other submit scripts to OWENS

  danOSU

  committed
  e75fb5d

  Copy full SHA for e75fb5d

• add product likelihood option (take product of last exponential instead of sum) and same parameter option

  danOSU

  committed
  ad387c0

  Copy full SHA for ad387c0

• add calibrate only one model weights

  danOSU

  committed
  1b0ae13

  Copy full SHA for 1b0ae13

• Merge branch 'jetscape' of https://github.com/danOSU/Taweret into jetscape

  Show description for d3ef87f
  danOSU

  committed
  d3ef87f

  Copy full SHA for d3ef87f

• updates from osc

  danOSU

  committed
  32493ac

  Copy full SHA for 32493ac

• notebooks for plotting local mixing

  danOSU

  committed
  8f510a6

  Copy full SHA for 8f510a6

Commits on Apr 6, 2023

• make sure no priors are selected for model parameters when models are fixed to MAP

  danOSU

  committed
  1dd70d2

  Copy full SHA for 1dd70d2

• python scripts for BMM

  danOSU

  committed
  4a7eb0f

  Copy full SHA for 4a7eb0f

• mixing of calibrated models, submision script

  danOSU

  committed
  6946462

  Copy full SHA for 6946462

• about files in the noebook directory

  danOSU

  committed
  eb9f372

  Copy full SHA for eb9f372

Commits on Apr 22, 2023

• more comments

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  71536a0

  Copy full SHA for 71536a0

Commits on Apr 23, 2023

• adding the ability to do linear mixing while taking into account the full error covariance matrix into account

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  56259bc

  Copy full SHA for 56259bc

• correcting a bug in finding simulation uncertainity in evaluate (gtting the diagonal fron np.diagonal() does not work as expected for 1,n,n arrays). And then updating the models to include a log_li…

  Show description for 91eb5cd
  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  91eb5cd

  Copy full SHA for 91eb5cd

• Merge branch 'jetscape_cross_correlation_mix' of https://github.com/danOSU/Taweret into jetscape_cross_correlation_mix

  Show description for 5a5ed84
  danOSU

  committed
  5a5ed84

  Copy full SHA for 5a5ed84

Commits on Apr 27, 2023

• Make a weight array so that we can directly multiply all exprimental data with it

  danOSU

  committed
  97166cc

  Copy full SHA for 97166cc

Commits on May 2, 2023

• including mean mixing capability in bivariate mixing

  danOSU

  committed
  4166509

  Copy full SHA for 4166509

• temporary add the normed likelihood calculation method from jetscape to utile. We can copy the source code and put it in util later on

  danOSU

  committed
  a3745fe

  Copy full SHA for a3745fe

• plot only non viscous parameters (do not plot mixing parameters in that plot)

  danOSU

  committed
  80fd816

  Copy full SHA for 80fd816

Commits on May 29, 2023

• docs update

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  e727dfb

  Copy full SHA for e727dfb

• we will keep covaraiance untoched in the mean mixing. There is no weighing there

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  ee4e3c3

  Copy full SHA for ee4e3c3

Commits on May 30, 2023

• bug fix for mean mixing--define log likelihood for BMMcor within bivariate rather than in the model -- some argument name changes

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  ed0a730

  Copy full SHA for ed0a730

Commits on Jul 31, 2023

• add the ability to include or remove full error covariance matrix in the initialization of the bivaraite mixing method

  danOSU

  committed
  a7d2fea

  Copy full SHA for a7d2fea

• the log likelihood does not consider full covariance. Can be changed if one wants to

  danOSU

  committed
  bbeee22

  Copy full SHA for bbeee22

• put the log likelihood calculation method from jetscape sims to here so it can be put to general use without sims package dependence

  danOSU

  committed
  0e9a232

  Copy full SHA for 0e9a232

• no interpolation for centrality in jetscape models and return None if the data is unavialable

  danOSU

  committed
  8eb58e7

  Copy full SHA for 8eb58e7

• bivariate mixing method with missing data. This can handle None values given by the model when it is evaluvated at input parameters with no experimental data

  danOSU

  committed
  2787243

  Copy full SHA for 2787243

Commits on Aug 3, 2023

• replacing bivariate linear with bivariate linear with missing data.Can be merged. No need to keeep as two seperate mixing methods

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  d5f106b

  Copy full SHA for d5f106b

Commits on Aug 7, 2023

• update documentation to fix typos and include type settings

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  a439896

  Copy full SHA for a439896

• merge with the normal bivaraite method since there are minimal differences. Taking into accoun Nan is done by numpy

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  53aed69

  Copy full SHA for 53aed69

Commits on Aug 8, 2023

• updated coleman models to run with updated bivariate linear mixing methods

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  edf6af4

  Copy full SHA for edf6af4

• update utils elementwise log likelihood to use evaluvate instead of predict

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  4afc026

  Copy full SHA for 4afc026

• Coleman toy models using all three different mixing methods available under bivariate linear mixing

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  ea34234

  Copy full SHA for ea34234

Commits on Aug 14, 2023

• testing for the three mixing methods in bivariatelinear.py

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  7b19d09

  Copy full SHA for 7b19d09

• example notebook of how to use all three mixing methods in bivaraite linear mixing

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  7f01419

  Copy full SHA for 7f01419

• remove jetscape sub package from this branch

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  5962722

  Copy full SHA for 5962722

• remove jetscape models

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  67fb55a

  Copy full SHA for 67fb55a

• clean repo from Jetscape mixing files and add some lots for coleman methods

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  ceeff77

  Copy full SHA for ceeff77

• no need to het jetscape data as submodule

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  134b4dc

  Copy full SHA for 134b4dc

• remove more jetscape related files

  

  Dananjaya Liyanage

  authored and

  Dananjaya Liyanage

  committed
  5a83e44

  Copy full SHA for 5a83e44