v2.0.0: compatibility with PhyloNetworks v0.17; output named tuple (breaking, #21 fix)

[cecileane]

breaking: dropped functionality to visualize a two-binary trait substitution model
old deprecated function removed

[cecileane]

The tests seem to fail randomly on macos, at the step that installs R. I do not understand why. When I re-run the macos job manually, it sometimes manages to install R (and then the actual tests run well too).

[pbastide]

Sorry, I'm on the move and away from office this end of the week, I cannot look into this in details.
The R install issue seems to come from this line of the setup.
I'm not sure what is causing it however.
Maybe we could try to install qpdf manually before hand to see if we get a more informative error ?
e.g. add something like:

      - name: insall qpdf on macos
        if: matrix.os == 'macos-latest'
        run: brew install qpdf
      - name: Install R
        uses: r-lib/actions/setup-r@v2
        with:
          use-public-rspm: true
          r-version: ${{ matrix.R-version }}

I can try to look more into it next week if needed.

[cecileane]

Thank you! I added the qpdf installation as you suggested, and the tests passed. Great!
The main change is a fix for issue #21: the output is a named tuple instead of a tuple, which is a breaking change.
I would love for someone else to look at the names. There might be better choices than the ones I thought of. You can look at the documentation preview for the list of names, for example, here

[cecileane]

Mmm, this last attempt failed: qpdf was successfully installed with brew, but the setup-r step ignored it, tried to install it by its own mean, and failed.

[pbastide]

For the qpdf issue: it seems that we were not the only one that had the issue, see:
r-lib/actions#948
r-lib/actions#950
r-lib/actions#952
It looks like it has just been fixed by version 2.11.2, see:
r-lib/actions#953.

So I removed the manual installation of qpdf, and hope that the new version of setup-r will for for us.

[pbastide]

I tried to fix some minor typos and punctuation.
The main change is to add that the plot is a "rightwards phylogram", to mirror the terms used in ape::plot.phylo.
But "phylogram" has an other definition on wikipedia, which I'm not sure is suited. So maybe it would be safer to not mention this term (as is was initially) ?

[pbastide]

I'm fine with 1.1.0, but out of curiosity, what made you change your mind and not go to 2.0.0 ?
(As it is a non backward compatible change, see #20 (comment))

[cecileane]

Ah, indeed, it should be 2.0.0. When starting this PR, I merely wanted a version compatible with the new PhyloNetworks v0.17.0. Later I noticed the open issue #21 and thought that it was a good time to act on it, introducing a breaking change.

I also agree with removing "phylogram", but it would be nice to keep the term "rightwards" to be explicit that these plots won't convey the "semidirected" nature of networks like "circular" trees can do.

[pbastide]

Thank you, looks good to me !