disable populating log_context based on commit_sha/commit_id

[matt-codecov]

apparently populating the log context was slamming the commit table. so this PR stops using the commit table when populating log context

previously:

• if commit_id or commit_sha+repo_id were available, query the commits table to populate every field
• everything is populated this way, including commit_id (the DB PK for the commit)

now:

• commit_id and commit_sha will be set if available but not used to query anything else
• if repo_id is available, that will take care of populating the repo + owner

it's possible this will just move the problem to the repos table. if that is the case, it's solvable but will take a little elbow grease:

• transform populate_from_sqlalchemy() in helpers/log_context.py into populate_log_context(kwargs) -> (Commit, Repository, Owner) in tasks/base.py
• add a run_impl() stub to BaseCodecovTask which takes owner, commit, and repo kwargs, make all task subclasses accept the same (with default None) values
• in several followup PRs, update task subclasses to use the passed-in models instead of re-fetching them

follow-up issue for model reuse: codecov/engineering-team#3140

[codecov-staging]

Codecov Report

All modified and coverable lines are covered by tests ✅

✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!

[codecov-qa]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
1767	1	1766	1

View the top 1 failed tests by shortest run time

helpers/tests/unit/test_log_context.py::::test_as_dict

Stack Traces | 0.042s run time

dbsession = <sqlalchemy.orm.session.Session object at 0x7f12185b2210>
mocker = <pytest_mock.plugin.MockFixture object at 0x7f1218c73890>

    def test_as_dict(dbsession, mocker):
        owner, repo, commit = create_db_records(dbsession)
        log_context = LogContext(commit_id=commit.id_, task_name="foo", task_id="bar")
        log_context.populate_from_sqlalchemy(dbsession)
    
        mock_span = mocker.Mock()
        mock_span.trace_id = 123
        mocker.patch("helpers.log_context.get_current_span", return_value=mock_span)
    
        # `_populated_from_db` is a dataclass field that we want to strip
        # `sentry_trace_id` is a property that we want to include
>       assert log_context.as_dict() == {
            "task_name": "foo",
            "task_id": "bar",
            "commit_id": None,
            "commit_sha": commit.commitid,
            "repo_id": repo.repoid,
            "repo_name": repo.name,
            "owner_id": owner.ownerid,
            "owner_username": owner.username,
            "owner_service": owner.service,
            "owner_plan": owner.plan,
            "sentry_trace_id": 123,
            "checkpoints_data": {},
        }
E       AssertionError: assert {'checkpoints...d': None, ...} == {'checkpoints...id': 191, ...}
E         
E         Omitting 4 identical items, use -vv to show
E         Differing items:
E         {'commit_sha': None} != {'commit_sha': 'c5b67303452bbff57cc1f49984339cde39eb1db5'}
E         {'owner_username': None} != {'owner_username': 'codecove2e'}
E         {'owner_service': None} != {'owner_service': 'github'}
E         {'repo_name': None} != {'repo_name': 'example-python'}...
E         
E         ...Full output truncated (5 lines hidden), use '-vv' to show

.../tests/unit/test_log_context.py:139: AssertionError

To view more test analytics, go to the Test Analytics Dashboard
📢 Thoughts on this report? Let us know!

[codecov-public-qa]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
1767	1	1766	1

View the top 1 failed tests by shortest run time

helpers/tests/unit/test_log_context.py::::test_as_dict

Stack Traces | 0.042s run time

dbsession = <sqlalchemy.orm.session.Session object at 0x7f12185b2210>
mocker = <pytest_mock.plugin.MockFixture object at 0x7f1218c73890>

    def test_as_dict(dbsession, mocker):
        owner, repo, commit = create_db_records(dbsession)
        log_context = LogContext(commit_id=commit.id_, task_name="foo", task_id="bar")
        log_context.populate_from_sqlalchemy(dbsession)
    
        mock_span = mocker.Mock()
        mock_span.trace_id = 123
        mocker.patch("helpers.log_context.get_current_span", return_value=mock_span)
    
        # `_populated_from_db` is a dataclass field that we want to strip
        # `sentry_trace_id` is a property that we want to include
>       assert log_context.as_dict() == {
            "task_name": "foo",
            "task_id": "bar",
            "commit_id": None,
            "commit_sha": commit.commitid,
            "repo_id": repo.repoid,
            "repo_name": repo.name,
            "owner_id": owner.ownerid,
            "owner_username": owner.username,
            "owner_service": owner.service,
            "owner_plan": owner.plan,
            "sentry_trace_id": 123,
            "checkpoints_data": {},
        }
E       AssertionError: assert {'checkpoints...d': None, ...} == {'checkpoints...id': 191, ...}
E         
E         Omitting 4 identical items, use -vv to show
E         Differing items:
E         {'commit_sha': None} != {'commit_sha': 'c5b67303452bbff57cc1f49984339cde39eb1db5'}
E         {'owner_username': None} != {'owner_username': 'codecove2e'}
E         {'owner_service': None} != {'owner_service': 'github'}
E         {'repo_name': None} != {'repo_name': 'example-python'}...
E         
E         ...Full output truncated (5 lines hidden), use '-vv' to show

.../tests/unit/test_log_context.py:139: AssertionError

To view more test analytics, go to the Test Analytics Dashboard
📢 Thoughts on this report? Let us know!

[github-actions]

✅ All tests successful. No failed tests were found.

📣 Thoughts on this report? Let Codecov know! | Powered by Codecov

[adrian-codecov]

So the intent here is to reenables the log context without the commit?

[matt-codecov]

yes

the stated problem is that populating the log context slammed the commit table. turns out if we don't query the commit table the only thing we are likely to lose is the DB PK of the commit which we don't tend to use anyway. so this PR stops querying the commit table

this may just move the problem, and we should maybe do the followup described in the PR description anyway

[matt-codecov]

followup task for the model reuse codecov/engineering-team#3140