HydraConfig support

[otherman16]

Hydra support

1. MNIST classification test pipeline was added
2. catalyst/experiments/hydra_config.py: HydraConfigExperiment was added
3. catalyst/runners/hydra_supervised.py: HydraSupervisedExperiment was added
4. catalyst/dl/script/hydra_run.py: @hydra.main() was added as entry point for hydra configuration experiment
5. catalyst/dl/script/run.py: optional argument '--hydra' was added to change entrypoint to hydra config experiment one
6. catalyst/settings.py: optional 'hydra' package import was added
7. catalyst/utils/hydra_config.py: hydra config utils was added
8. catalyst/utils/hydra_scripts.py: hydra scripts utils was added
9. catalyst/utils/hydra_sys.py: hydra sys utils was added
10. requirements/requirements-hydra.txt: optional 'hydra-core' package was added

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• Did you read the contribution guide?
• Did you check the code style? catalyst-make-codestyle && catalyst-check-codestyle (pip install -U catalyst-codestyle).
• Did you make sure to update the docs? We use Google format for all the methods and classes.
• Did you check the docs with make check-docs?
• Did you write any new necessary tests?
• Did you check that your code passes the unit tests pytest . ?
• Did you add your new functionality to the docs?
• Did you update the CHANGELOG?

Description

Hydra support for catalyst-dl run

Previous PR: OmegaConf support for "catalyst-dl run" #943

Related Issue

Hydra/OmegaConf configuration structure #912

Type of Change

• Examples / docs / tutorials / contributors update
• Bug fix (non-breaking change which fixes an issue)
• Improvement (non-breaking change which improves an existing feature)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

PS

• I know, that I could join slack for pull request discussion.

[otherman16]

I don't know what's wrong with this Error

Traceback (most recent call last):
  File "catalyst/dl/scripts/run.py", line 9, in <module>
    from catalyst.settings import IS_HYDRA_AVAILABLE
  File "/home/runner/work/catalyst/catalyst/catalyst/settings.py", line 1, in <module>
    from typing import Any, Dict, List, Optional, Tuple
  File "/home/runner/work/catalyst/catalyst/catalyst/typing.py", line 4, in <module>
    from typing import Dict, Union
ImportError: cannot import name 'Dict' from partially initialized module 'typing' (most likely due to a circular import) (/home/runner/work/catalyst/catalyst/catalyst/typing.py)

Think I need some review help guys.

[Scitator]

Dear @otherman16 ,
should we close #943 then?

[Scitator]

thrilled by this PR, amazing work!

[otherman16]

Yes we can close #943

[omry]

Thanks for working on this!
While I understand you have backward compatibility requirements that might be in conflict with the suggestions here, I wanted to provide them anyway (at least as food for thought) :
I'd recommend utilizing config composition to get this instead of manually editing the config.
for example, the defaults list of the primary config can look like:

defaults:
  - catalyst/config  # mandatory catalyst config (provided by catalyst)
  - foo: bar # user config groups

catalyst/config can contain everything needed by catalyst (what you are adding above).

A few extra points:

1. I am guessing this is designed to be backward compatible with the standard command line.
   If possible, break away from that and utilize hierarchy to give catalyst a space of it's own where it wouldn't conflict with other frameworks/user config.
   for example:
   catalst/config.yaml:

catalyst:
  resume: null
  seed: 42
  ...

2. Consider using Structured Configs to get config type safety, the above yaml file can be replaced with something like:

@dataclass
class CatalystConf:
  resume: Optional[bool] = None
  seed: int = 42

ConfigStore.instance().store(group="catalst", name="config", node=CatalystConf)

(See the Structured Configs tutorial for more info).

[otherman16]

Thank you. Thinking about it.

[otherman16]

I do not know what is wrong with import catalyst.experiments.functional as F in catalyst/experiments/config.py in VMs with python3.6 inside.
Can you help me please?

[Scitator]

I do not know what is wrong with import catalyst.experiments.functional as F in catalyst/experiments/config.py in VMs with python3.6 inside.
Can you help me please?

could you try something like import catalyst.experiments.functional import my_required_function?