Build graph from `recipe.yaml`

[Hofer-Julian]

Tests:

• [ ] CLI tests for parse-feedstock
• [x] recipe.yaml is rendered and parsed correctly
• [x] Returned graph contains expected data

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 82.65306% with 17 lines in your changes missing coverage. Please review.

Project coverage is 76.96%. Comparing base (0642d39) to head (82639a6).

Files	Patch %	Lines
conda_forge_tick/utils.py	77.77%	12 Missing :warning:
conda_forge_tick/feedstock_parser.py	82.60%	4 Missing :warning:
conda_forge_tick/update_sources.py	75.00%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #2851 +/- ## ========================================== + Coverage 76.92% 76.96% +0.04% ========================================== Files 111 112 +1 Lines 11870 11943 +73 ========================================== + Hits 9131 9192 +61 - Misses 2739 2751 +12 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[jezdez]

This PR title seems to bury the led that this is adding rattler-build to conda-forge?

[beckermr]

Well "adding to conda-forge" is a bit misleading since full support in all of our tools will take a long, long time.

It is true that with this PR, we will be using rattler-build as the current parser for the new format when building the bot's graph.

However, that is a design choice we can change later should other parsers come along.

[Hofer-Julian]

Thanks for the review comments during the process @beckermr, they were very helpful!

This is the current status of the PR:

• rattler-build is used to render the recipe, if it exists with a failure an exception is thrown, please let me know if you prefer a different behavior
• then, it is parsed to reflect the conda build's schema
• the biggest difference is that I don't fill the test field, according to @wolfv the two recipe formats differ too much here for that to make sense
• .ci_support files should be accounted for, I haven't added tests for that one though
• as you mentioned for_pinning is not supported yet, and will be implemented in a follow-up
• I've moved to pathlib for functions that I've touched, please let me know if you want me to refrain from these drive-by improvements in the future
• I haven't added specific tests for the added CLI command yet, as far as I can tell the meta-yaml also doesn't have them. If you have specific suggestions, I am happy to implement them