(app) Run the flow only if the state has updated 1/2

[tchaton]

What does this PR do?

This PR introduces a breaking change where the flow will execute only if the state changes. This is required in order to move toward a serverless expression of the flow. This still triggers if the flow modifies its own state which we might want to disable in the future, but I believe this is required for now.

Does your PR introduce any breaking changes? If yes, please list them.

Yes, this used to print infinitely on master, but now this will print only once as no state was modified.

class Flow(LightningFlow):
    def run(self):
        logger.info("Hello World")

app = LightningApp(Flow())

Parts of Lightning-AI/lightning-app#938

Before submitting

• Was this discussed/approved via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

cc @Borda

[lantiga]

Great that we're getting to this, really important.

It checks out, I've mostly got a question around the scheduler test.

[lantiga]

Just a comment on the scheduler frequency. We should keep it at a second granularity (so 0.5 timeout is fine) and document it, and just patch it during testing to make tests faster.