ENH: attached rate and motif parameters to cogent3 tree object

[YapengLang]

#27

[rmcar17]

@YapengLang could you add a test which checks for the parameters when build_tree/fit_tree are called?

[YapengLang]

@YapengLang could you add a test which checks for the parameters when build_tree/fit_tree are called?

Sure, will do.

[GavinHuttley]

nice one! few changes, clearly need to understand that numerical issue

[GavinHuttley]

Thanks for the revisions! The test failures due to the float comparisons are all ~1e-6. If you set the tolerance to 1e-5 do they all pass? That level of tolerance for a parameter estimate on limited data seems reasonable.

[YapengLang]

Thanks for the revisions! The test failures due to the float comparisons are all ~1e-6. If you set the tolerance to 1e-5 do they all pass? That level of tolerance for a parameter estimate on limited data seems reasonable.

hi @GavinHuttley , setting abs = 1e-5 or 1e-6 does not solve. The largest difference arises from the GTR fit, where one rate estimate is 3.9724, but the expected value is around 3.9915.

Alternatively, I can try setting rel = 1e-2. This would allow the GTR fit to pass with 3.9724 == 3.9915 ± 4.0e-02, and the other fits would also pass (please see commit 700f398).

[GavinHuttley]

Thanks for clarifying @YapengLang !

[GavinHuttley]

Looks good!

[rmcar17]

Thanks @YapengLang, it will auto-merge once the tests all pass!