docker/api/requirements: fix api container build failure

[JenySadadia]

Fixes #303

Fix build failures of api container. Error logs can be found at https://github.com/kernelci/kernelci-api/issues/303.

Root cause:
fastapi[all] 0.68.1 package installs pyyaml 5.4.1 package internally. That again depends on CPython package. Due to the recent upgrade to CPython 3.0.0, pyyaml 5.4.1 stopped building with python 3.10. To fix this, upgrading fastapi[all] to 0.99.1 version as that uses pyyaml 6.0.1. The selected fastapi version is compatible with current fastapi-pagination and fastapi-versioning packages.

[JenySadadia]

Fixed pytest error from test_node_handler.test_get_node_by_id_endpoint_empty_response.
The issue occurred while trying to get node by ID when node does not exist matching provided ID:

pydantic/error_wrappers.py:63: in pydantic.error_wrappers.ValidationError.errors
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = typing.Union[api.models.Regression, api.models.Node], attr = '__pydantic_model__'

    def __getattr__(self, attr):
        if attr in {'__name__', '__qualname__'}:
            return self._name or self.__origin__.__name__
    
        # We are careful for copy and pickle.
        # Also for simplicity we don't relay any dunder names
        if '__origin__' in self.__dict__ and not _is_dunder(attr):
            return getattr(self.__origin__, attr)
>       raise AttributeError(attr)
E       AttributeError: __pydantic_model__

/usr/lib/python3.10/typing.py:984: AttributeError

[gctucker]

@JenySadadia Some checks are still failing. Otherwise I think this is good to be merged and necessary so other PRs can pass checks too.

[JenySadadia]

e2e tests were failing after fastapi version change.
Error logs:

ernelci-api-e2e-tests | _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
kernelci-api-e2e-tests | 
kernelci-api-e2e-tests | self = <api.db.Database object at 0x7f25275935e0>
kernelci-api-e2e-tests | model = <class 'api.models.User'>
kernelci-api-e2e-tests | attributes = {'profile.username': 'test_user'}
kernelci-api-e2e-tests | 
kernelci-api-e2e-tests |     async def find_one_by_attributes(self, model, attributes):
kernelci-api-e2e-tests |         """Find one object with matching attributes without pagination
kernelci-api-e2e-tests |     
kernelci-api-e2e-tests |         The attributes dictionary provides key/value pairs used to find an
kernelci-api-e2e-tests |         object with matching attributes.
kernelci-api-e2e-tests |         """
kernelci-api-e2e-tests |         col = self._get_collection(model)
kernelci-api-e2e-tests | >       obj = await col.find_one(attributes)
kernelci-api-e2e-tests | E       RuntimeError: Task <Task pending name='anyio.from_thread.BlockingPortal._call_func' coro=<BlockingPortal._call_func() running at /home/kernelci/.local/lib/python3.10/site-packages/anyio/from_thread.py:217> cb=[TaskGroup._spawn.<locals>.task_done() at /home/kernelci/.local/lib/python3.10/site-packages/anyio/_backends/_asyncio.py:661]> got Future <Future pending cb=[_chain_future.<locals>._call_check_cancel() at /usr/local/lib/python3.10/asyncio/futures.py:385]> attached to a different loop

The newer TestClient from fastapi is causing issues by using anyio and creating a different event loop.

Fixed it by removing synchronous TestClient and used async one for all the tests.
This documentation also recommends using async test client in this kind of scenarios:https://www.starlette.io/testclient/#asynchronous-tests

[gctucker]

@JenySadadia See also #306 to avoid the issue altogether without changing the code. If this PR #305 is all fine and has no side-effect issues then we can just merge it I guess. If more work is required then we can have #306 merged first to get back to working image builds and fix this properly later with an upgrade to the latest FastAPI and Pydantic.

[JenySadadia]

@JenySadadia See also #306 to avoid the issue altogether without changing the code. If this PR #305 is all fine and has no side-effect issues then we can just merge it I guess. If more work is required then we can have #306 merged first to get back to working image builds and fix this properly later with an upgrade to the latest FastAPI and Pydantic.

Thanks for the other fix. I think we can merge it as it doesn't require any code changes.
If we want to upgrade packages in the future, it would need some code rework and we can refer to this PR for it.

[gctucker]

OK so I'm adding the skip label so it doesn't interfere with the staging runs. I guess we could have a GitHub issue about upgrading to the latest FastAPI with all related dependencies, and then resume progress on this PR.

[gctucker]

Replaced with #321 to do a full upgrade for the initial production rollout.