[AUTOREVERT] tune entry point for cli and event bridge, enabling providing custom running configurations (#7526)

Changes lambda entry point for autorevert to enable better handling
eventbridge parameters.

The changes are:

* Moved entry point from a global `main` to specific ones for CLI and
lambda;
* Moved main execution logic to a reused function;
* Created a general config holder interface dataclass;

this enables to use the lambda with different configs from eventbridge
like:

```
resource "aws_cloudwatch_event_rule" "pytorch_auto_revert_5m" {
  name                = "pytorch-auto-revert-5m"
  description         = "Every 5 minutes with config A"
  schedule_expression = "rate(5 minutes)"
}

resource "aws_cloudwatch_event_target" "pytorch_auto_revert_5m_target" {
  rule      = aws_cloudwatch_event_rule.pytorch_auto_revert_5m.name
  target_id = "pytorch-auto-revert-5m"
  arn       = aws_lambda_function.pytorch_auto_revert.arn

  input = jsonencode({
    mode      = "hud"
    dry_run   = false
    overrides = {
      bisection_limit    = 10
      clickhouse_database = "default"
    }
  })
}
```

Key reasonings on the decisions:
- created a single class for parsed configs that is distinct from
default configs, still allowing validating and maintaining readability
and single use (one is default, another is actual);
- Reused entry points from lambda and cli, so, mostly if not all, can be
fully replicated via cli in case of firefight or debugging;
- broken entry point for cli and lambda: one parses opts and dotfiles,
another parse eventbridge. Separation of concerns and easier to read;
- removed possibility of running without any parameters in CLI to
replicate lambda - there will be multiple configurations, so, this does
not make sense anymore;

---------

Signed-off-by: Jean Schmidt <[email protected]>
Co-authored-by: Ivan Zaitsev <[email protected]>

Author: jeanschmidt

aws/lambda/pytorch-auto-revert/.gitignore

@@ -1,3 +1,4 @@
 *.zip
 deployment/
 venv/
+*.html

aws/lambda/pytorch-auto-revert/Makefile

@@ -13,14 +13,10 @@ venv/bin/python:
 	venv/bin/pip install -r requirements.txt
 
 venv/bin/lintrunner: venv/bin/python
-	venv/bin/pip install lintrunner==0.12.5 boto3-stubs==1.34.51
+	venv/bin/pip install -r dev_requirements.txt
 	# lintrunner only works properly with virtualenv if you activate it first
 	. venv/bin/activate && lintrunner init --config ../../../.lintrunner.toml
 
-.PHONY: run-local
-run-local: venv/bin/python
-	NOTIFY_ISSUE_NUMBER=265 REPO_FULL_NAME=pytorch/pytorch-canary venv/bin/python -m pytorch_auto_revert --dry-run
-
 .PHONY: run-local-workflows
 run-local-workflows: venv/bin/python
 	NOTIFY_ISSUE_NUMBER=265 REPO_FULL_NAME=pytorch/pytorch-canary venv/bin/python -m pytorch_auto_revert --dry-run autorevert-checker Lint trunk pull inductor linux-binary-manywheel --hours 8

aws/lambda/pytorch-auto-revert/dev_requirements.txt

@@ -0,0 +1,3 @@
+boto3-stubs>=1.34.51
+lintrunner==0.12.5
+setuptools>=80.9.0