PSLmodels / OG-UK

An overlapping generations model to simulate fiscal policy the United Kingdom.

https://pslmodels.github.io/OG-UK

Other

6 stars 7 forks source link

Update with microsimulation changes #27

Closed nikhilwoodruff closed 3 years ago

nikhilwoodruff commented 3 years ago

@jdebacker I think this should have the changes to get it running on your machine

jdebacker commented 3 years ago

@nikhilwoodruff Thanks for this PR. Can you update environment.yml to install OG-USA 0.6.4 from https://github.com/PSLmodels/OG-USA/releases/tag/v0.6.4?

Thanks!

nikhilwoodruff commented 3 years ago

@jdebacker done!

jdebacker commented 3 years ago

@nikhilwoodruff Why the hard pin to 3.7.7? Is something incompatible with Python 3.8?

jdebacker commented 3 years ago

@nikhilwoodruff I'd like to see if we can at least get the test_get_micro_data.py::test_frs and test_run_example.py::test_run_og_uk tests to pass. Do you understand the test failures and what needs to be done given changes in openfisca-uk and data?

nikhilwoodruff commented 3 years ago

@jdebacker I agree - the tricky thing here is that the GitHub action isn't going to be able to use the actual microdata until we can confirm with the UKDA that it's allowed. Right now I'm going to try and get it working with the synthetic dataset

jdebacker commented 3 years ago

@jdebacker I agree - the tricky thing here is that the GitHub action isn't going to be able to use the actual microdata until we can confirm with the UKDA that it's allowed. Right now I'm going to try and get it working with the synthetic dataset

Testing on the synthetic is fine here. But it'd be good to have the capability use both inget_micro_data.py, even if only the synthetic is tested on GH.

nikhilwoodruff commented 3 years ago

@jdebacker Just got the test working here for get_micro_data, it uses the synthetic FRS when it can't find the actual FRS.

codecov[bot] commented 3 years ago

Codecov Report

Merging #27 (46e2cd8) into main (bc35185) will decrease coverage by 8.12%. The diff coverage is 38.77%.

@@            Coverage Diff             @@
##             main      #27      +/-   ##
==========================================
- Coverage   53.98%   45.85%   -8.13%     
==========================================
  Files           4        6       +2     
  Lines         113      205      +92     
==========================================
+ Hits           61       94      +33     
- Misses         52      111      +59

Flag	Coverage Δ
unittests	`45.85% <38.77%> (-8.13%)`	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
plot.py	`0.00% <0.00%> (ø)`
small_ubi_reform.py	`0.00% <0.00%> (ø)`
oguk/get_micro_data.py	`76.31% <97.43%> (+18.17%)`	:arrow_up:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update bc35185...46e2cd8. Read the comment docs.

nikhilwoodruff commented 3 years ago

@jdebacker the tests are now passing. Though I'm still not able to find the steady-state locally.

jdebacker commented 3 years ago

@nikhilwoodruff Thanks for this PR. We'll address calibration and other issues that are likely the root of the SS failure in future PRs. This PR is important to get the link to the microsimulation model working.