Fix #160: Improve default time axis formatting using ConciseDateFormatter

[Amityush-lgtm]

PR Description

Fixes #160

This PR improves the default time axis formatting for spectrogram plots by introducing Matplotlib's ConciseDateFormatter and AutoDateLocator for more readable and concise time labels on the x-axis.

What Changed:

• Introduced _TimeAxisMixin in radiospectra/mixins.py to share time axis setup logic between different plotting methods.
• Converted self.times (astropy.time.Time) to native Python datetime objects before plotting. This allows Matplotlib to use its built-in ConciseDateFormatter and AutoDateLocatornatively.
• Modified PcolormeshPlotMixin.plot() and NonUniformImagePlotMixin.plotim() to use the new mixin and datetime-based plotting.
• Added tests to verify the formatter and x-label generation.

Visual Proofs:

AI Assistance Disclosure

AI tools were used for:

• Code generation (e.g., when writing an implementation or fixing a bug)
• Test/benchmark generation
• Documentation (including examples)
• Research and understanding
• No AI tools were used

Regardless of AI use, the human contributor remains fully responsible for correctness, design choices, licensing compatibility, and long-term maintainability.

Note:

For the code given in the #58 (comment), there was a slight change made in the code for using datetime:

axes.set_xlim(Time("2020-07-11T02:15").datetime, Time("2020-07-11T02:50").datetime)

@samaloney, Let me know your thoughts on this! Happy to make changes.

[Amityush-lgtm]

I was unsatisfied with my current approach in using native datetime since we have to manually convert when using axes.set_xlim so i started exploring the other ways that avoids it. I am now confused between two alternatives to improve default time axis formatting without requiring manual user conversion.

• First approach, Since you preferred to use time_support() from astropy and i started to find ways to keep it intact so for that i need to add an MJD aware ConciseDateFormatter adapter.
• Second approach, to use Matplotlib date values self.times.plot_date internally and adding a small custom date converter so the axis can still accept astropy.time.Time.

@samaloney, Let me know your thoughts on this, which one should i prefer or any other alternatives you have. Happy to make changes!

[samaloney]

Those plots look much better! So it sounds like there is an incompatibly between time_support and ConciseDateFormatter, it does seem like time_support takes some customisation support away.

the adapter approach sounds interesting would try to hook into the matplotlib.units API and remove the need to the time_support?

If you do try this just make sure to not remove the commits for this working approach above so it would be easy to compare.

[Amityush-lgtm]

@samaloney I have implemented the matplotlib.units approach. It works with both set_xlim(Time()) and set_xlim(Time().datetime). Let me know your thoughts on this.

[Amityush-lgtm]

These are plots made using this method

[samaloney]

I haven't forgotten this will have another look at this later this week.

[Copilot]

Pull request overview

This PR addresses #160 by improving spectrogram time-axis labeling, switching the default x-axis formatting to Matplotlib’s date locator/formatter machinery for more readable, concise time ticks while preserving support for setting limits with both astropy.time.Time and datetime.

Changes:

• Added a shared _TimeAxisMixin and a custom Matplotlib date converter to support astropy.time.Time inputs on the x-axis.
• Updated PcolormeshPlotMixin.plot() and NonUniformImagePlotMixin.plotim() to plot using Matplotlib date-number values and apply AutoDateLocator + ConciseDateFormatter.
• Expanded test coverage to validate date conversion behavior, axis limit compatibility, leap-second handling, formatter selection, and x-label content.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
radiospectra/spectrogram/tests/test_spectrogrambase.py	Adds/updates tests asserting plot-date conversion, set_xlim compatibility (Time/datetime), leap-second handling, and ConciseDateFormatter usage.
radiospectra/mixins.py	Introduces _TimeAxisMixin plus a Time-aware DateConverter, and updates plotting to use plot_date + concise time-axis formatting.
changelog/163.feature.rst	Adds a news fragment describing the improved default time-axis formatting behavior.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[samaloney]

Ok so it would be better to no have to change the plot methods so I think take a similar approach to astropy here and make a context manager but use matplotlib dates and just document that this won't handle leap seconds or nanosecond time resolution the for the entire age of the universe 🤣

class ConciseAstropyConverter(munits.ConversionInterface):
     # implement requrired methods use AutoDateLocator
     # and ConciseDateFormatter

@contextmanager
def concise_time_support():
    # Save the original converter if it exists
    # Register our concise-compatible converter
    # Restore 

[Amityush-lgtm]

Thanks @samaloney, I'll implement a custom matplotlib.units converter with a context manager and concise_time_support() and also add note about the lack of nanosecond resolution 🙂. On it!

[Amityush-lgtm]

@samaloney, I have implemented ConciseAstropyConverter and the concise_time_support() context manager.
Let me know your thoughts on this!