Remove write_jpeg and write_png in favor of encode + write_file

[fmassa]

Follow-up from #2765

We don't need to have those functions be in C++ as they are just a combination of encode + write_file

[codecov]

Codecov Report

Merging #2766 into master will increase coverage by 0.33%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2766      +/-   ##
==========================================
+ Coverage   72.42%   72.76%   +0.33%     
==========================================
  Files          96       96              
  Lines        8313     8313              
  Branches     1293     1292       -1     
==========================================
+ Hits         6021     6049      +28     
+ Misses       1903     1870      -33     
- Partials      389      394       +5     

Impacted Files	Coverage Δ
torchvision/io/image.py	89.13% <100.00%> (+4.34%)	⬆️
torchvision/io/_video_opt.py	39.37% <0.00%> (+16.25%)	⬆️

---

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 6e639d3...1fdd677. Read the comment docs.

[andfoy]

I think we need to remove the write_* functions from the documentation, but overall it looks good

[fmassa]

The write_png and write_jpeg are still present in the codebase, I've just removed the C++ equivalents. But if we remove them in a follow-up PR we do need to change the documentation, yes