Support for inverted clocks

[mfhepp]

See #26

This adds support for inverted clock outputs, which is an easy way of obtaining a pair of 0° and 180° phase shifted signals from the same clock.

I added the parameter inverted=False to the two methods for configuring clocks.

So you can now create a pair of clocks like so

    si5351.clock_0.configure_fractional(si5351.pll_a, int_divider, numerator, denominator,
                                        inverted=False)
    si5351.clock_1.configure_fractional(si5351.pll_a, int_divider, numerator, denominator,
                                        inverted=True)

The default behavior is unchanged.

[ladyada]

hiya! thanks so much for submitting a PR! we can review & merge PRs once they have passed continuous integration (CI). that means that code is 'linted' - that means it is formatted and passes basic structure tests, so that we maintain the same text formatting for all new code
if your code isnt passing, check the CI output (click on the red X next to the PR to scroll through the log and find where the error is

here is a tutorial on pylint and black: https://learn.adafruit.com/improve-your-code-with-pylint

and overall how to contribute PRs to circuitpython: https://learn.adafruit.com/contribute-to-circuitpython-with-git-and-github

once you get that green checkmark that indicates CI has passed, please comment reply to this post so we know its time for another review (we may not get notified on CI pass and miss that its time to look!)

[mfhepp]

Thanks! The CI test fails because it does not want more than five arguments for a method. However, I do not see an elegant way of avoiding that in this case:

def configure_fractional(self, pll, divider, numerator, denominator, inverted=False):
          """Configure the clock output with the specified PLL source ...

because pll, divider, numerator, denominator are mandatory for this functionality, and the inverted option is IMO best handled via an optional argument. Adding a method or attribute does not work well, because you need to reset the SI5351 PLL when changing the phase, and an intransparent PLL reset seems bad to me.

[dhalbert]

Add a # pylint: disable=too-many-arguments to the routine in question. It's fine to override these guidelines as necessary.

[tekktrik]

@mfhepp Looks like black is angry again, running pre-commit run --all-files again should fix it!

[mfhepp]

@mfhepp Looks like black is angry again, running pre-commit run --all-files again should fix it!

Thanks - but even after running pre-commit run --all-files, it does not pass.

[dhalbert]

Thanks - but even after running pre-commit run --all-files, it does not pass.

If it is black that was complaining, it will reformat the files, and you can then push a new commit.

If that is what you already did, then there may be a version difference between the black you are using the the one that the CI is using.