Adds conda support to envd plugin

[thomasjpfan]

Why are the changes needed?

This PR adds conda support to envd. This makes it simple to use Nvidia's runtime environment for PyTorch:

image_spec = ImageSpec(
    base_image="ubuntu:20.04",
    python_version="3.11",
    packages=["flytekit"],
    conda_packages=["pytorch", "pytorch-cuda=12.1"],
    conda_channels=["pytorch", "nvidia"]
)

which is inline with PyTorch's docs.

Furthermore, I do not think it is possible to install cuda=12.1 with envd. Using conda allows for cuda=12.1 and makes it simple to switch between Nvidia runtimes.

What changes were proposed in this pull request?

This PR adds conda_packages and conda_channels to ImageSpec to specific the conda packages and channels.

How was this patch tested?

I added a unit test to check the expected behavior.

Docs link

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (9bf7afa) 85.76% compared to head (456e6ac) 85.78%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2020      +/-   ##
==========================================
+ Coverage   85.76%   85.78%   +0.01%     
==========================================
  Files         313      313              
  Lines       23426    23447      +21     
  Branches     3510     3510              
==========================================
+ Hits        20092    20114      +22     
  Misses       2727     2727              
+ Partials      607      606       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pingsutw]

LGTM, could you resolve the merge conflict?

[pingsutw]

@thomasjpfan unit tests in envd-plugin are failing