patch bug in IAccountant len functions. Currently returns # of unique noise_multiplier values vs intended # of steps

[eddiestudies]

Types of changes

• [ x ] Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change if dependent on len() == 1)
• Docs change / refactoring / dependency upgrade

Motivation and Context / Related issue

Accountants currently implement len with len(self.history). This means that len(accountant) is 1 or # of noise_multiplier changes vs the documented "Number of optimization steps taken so far" . This is because self.history is compact and the true step count is accumulated into self.history[:][2].

How Has This Been Tested (if it applies)

I added iterative tests over existing accountants. For GDP, noise_multiplier changes raise a ValueError so it is ommitted from the noise_multiplier test.
I moved the fix into the base class since the contract for history is defined there. This should avoid future accountants from running into the same bug.

Checklist

• The documentation is up-to-date with the changes I made.
• I have read the CONTRIBUTING document and completed the CLA (see CONTRIBUTING).
• All tests passed, and additional code has been covered with new tests.

[meta-codesync]

This pull request has been imported. If you are a Meta employee, you can view this in D109374292. (Because this pull request was imported automatically, there will not be any future comments.)