Fix readme badges and expected number of tests

[f0uriest]

Some stuff that was missing from #1213

• Updates readme badges to point to the renamed gh actions
• Tells codecov to expect 14 batches of coverage data, not 10 (we used to have 6 regression + 4 unit, now 6 regression + 8 unit)

[YigitElma]

Oh sorry, I didn't know these exist.
Can we try to resolve #1221 here?

[YigitElma]

I assume it will not affect a lot of PR's but some small ones (like this one or just docs update) can be faster. The conditions can be

• if no change in desc or requirements don't run notebook tests
• if no change in desc, tests (without benchmarks) or requirements don't run unit and regression tests (also disable code-cov)
• if no change in desc, tests/benchmarks or requirements don't run benchmarks

Also, I asked ChatGPT if there is a way to check changes that are made to comments (not to actual code). Here is a basic example,

name: Check Code Changes

on: [push, pull_request]

jobs:
  check_code_changes:
    runs-on: ubuntu-latest
    steps:
    - name: Check out the repository
      uses: actions/checkout@v2

    - name: Set up Python
      uses: actions/setup-python@v2
      with:
        python-version: '3.x'

    - name: Install git diff tool (if needed)
      run: sudo apt-get install diffutils

    - name: Check if there are code changes
      run: |
        git diff HEAD^1 -- '*.py' > changes.diff
        if grep -q -v -e '^\s*#' -e '^\s*$' changes.diff; then
          echo "There are code changes!"
          exit 1
        else
          echo "No code changes detected."
          exit 0
        fi

[f0uriest]

I'd rather fix this for now since it makes the main landing page look bad right now, we can do the testing improvements in another one.

[github-actions]

|             benchmark_name             |         dt(%)          |         dt(s)          |        t_new(s)        |        t_old(s)        | 
| -------------------------------------- | ---------------------- | ---------------------- | ---------------------- | ---------------------- |
 test_build_transform_fft_lowres         |     +0.13 +/- 3.90     | +6.61e-04 +/- 2.04e-02 |  5.25e-01 +/- 7.4e-03  |  5.24e-01 +/- 1.9e-02  |
 test_equilibrium_init_medres            |     +0.51 +/- 0.72     | +2.10e-02 +/- 2.99e-02 |  4.16e+00 +/- 2.2e-02  |  4.14e+00 +/- 2.0e-02  |
 test_equilibrium_init_highres           |     +0.45 +/- 1.42     | +2.44e-02 +/- 7.80e-02 |  5.50e+00 +/- 7.2e-02  |  5.48e+00 +/- 3.0e-02  |
 test_objective_compile_dshape_current   |     +0.92 +/- 1.25     | +3.51e-02 +/- 4.80e-02 |  3.87e+00 +/- 3.8e-02  |  3.83e+00 +/- 2.9e-02  |
 test_objective_compute_dshape_current   |     -1.24 +/- 1.36     | -4.35e-05 +/- 4.77e-05 |  3.46e-03 +/- 3.5e-05  |  3.51e-03 +/- 3.2e-05  |
 test_objective_jac_dshape_current       |     +4.06 +/- 5.15     | +1.58e-03 +/- 2.00e-03 |  4.04e-02 +/- 1.2e-03  |  3.88e-02 +/- 1.6e-03  |
 test_perturb_2                          |     +0.94 +/- 3.77     | +1.64e-01 +/- 6.56e-01 |  1.75e+01 +/- 6.3e-01  |  1.74e+01 +/- 1.8e-01  |
 test_proximal_freeb_jac                 |     -1.80 +/- 2.56     | -1.37e-01 +/- 1.95e-01 |  7.46e+00 +/- 8.4e-02  |  7.60e+00 +/- 1.8e-01  |
 test_solve_fixed_iter                   |     +0.19 +/- 60.62    | +9.54e-03 +/- 3.06e+00 |  5.06e+00 +/- 2.1e+00  |  5.05e+00 +/- 2.2e+00  |
 test_build_transform_fft_midres         |     +0.80 +/- 7.01     | +5.04e-03 +/- 4.44e-02 |  6.38e-01 +/- 3.8e-02  |  6.33e-01 +/- 2.3e-02  |
 test_build_transform_fft_highres        |     -0.50 +/- 4.81     | -5.11e-03 +/- 4.94e-02 |  1.02e+00 +/- 3.4e-02  |  1.03e+00 +/- 3.6e-02  |
 test_equilibrium_init_lowres            |     -1.26 +/- 4.69     | -5.22e-02 +/- 1.94e-01 |  4.09e+00 +/- 1.7e-01  |  4.14e+00 +/- 9.3e-02  |
 test_objective_compile_atf              |     +1.84 +/- 5.66     | +1.45e-01 +/- 4.48e-01 |  8.06e+00 +/- 3.5e-01  |  7.92e+00 +/- 2.7e-01  |
 test_objective_compute_atf              |    +10.33 +/- 5.65     | +1.06e-03 +/- 5.82e-04 |  1.14e-02 +/- 5.4e-04  |  1.03e-02 +/- 2.2e-04  |
 test_objective_jac_atf                  |     -0.07 +/- 2.13     | -1.30e-03 +/- 4.22e-02 |  1.98e+00 +/- 2.2e-02  |  1.98e+00 +/- 3.6e-02  |
 test_perturb_1                          |     +3.32 +/- 5.73     | +4.25e-01 +/- 7.33e-01 |  1.32e+01 +/- 6.7e-01  |  1.28e+01 +/- 3.0e-01  |
 test_proximal_jac_atf                   |     +0.07 +/- 1.11     | +5.56e-03 +/- 9.21e-02 |  8.29e+00 +/- 7.1e-02  |  8.29e+00 +/- 5.8e-02  |
 test_proximal_freeb_compute             |     -0.70 +/- 1.99     | -1.31e-03 +/- 3.71e-03 |  1.85e-01 +/- 1.5e-03  |  1.86e-01 +/- 3.4e-03  |

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.34%. Comparing base (416b5fb) to head (788a197).
Report is 1887 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1235   +/-   ##
=======================================
  Coverage   95.33%   95.34%           
=======================================
  Files          90       90           
  Lines       22702    22702           
=======================================
+ Hits        21643    21645    +2     
+ Misses       1059     1057    -2     

see 1 file with indirect coverage changes

[YigitElma]

Btw, it is not related to #1213 because even before that this was the case, but why do we have extremely high standard deviation in test_fixed_iter?

[f0uriest]

Btw, it is not related to #1213 because even before that this was the case, but why do we have extremely high standard deviation in test_fixed_iter?

Could be due to the first step being compiled? We used to properly account for that but since #1043 it recompiles a lot less so the later ones are faster?

[YigitElma]

Btw, it is not related to #1213 because even before that this was the case, but why do we have extremely high standard deviation in test_fixed_iter?

Could be due to the first step being compiled? We used to properly account for that but since #1043 it recompiles a lot less so the later ones are faster?

Ah yes, it looks like the std dev increased after that. Nice