Pin tk<9 for py3.14

[AdamRJensen]

• I am familiar with the contributing guidelines
• I attest that all AI-generated material has been vetted for accuracy and is in compliance with the pvlib license
• Tests added
• New code is fully documented. Includes numpydoc compliant docstrings, examples, and comments where necessary.
• Pull request is nearly complete and ready for detailed review.
• Maintainer: Appropriate GitHub Labels (including remote-data) and Milestone are assigned to the Pull Request and linked Issue.

Test alternative solution to #2875. This alternative solution was discoved by using Claude FYI.

[kandersolar]

@AdamRJensen nice find! Did you find a bug report or some other context?

[kandersolar]

I have made some progress towards understanding this. First, simplify the reproducer down to:

micromamba create -y -n test_env -c anaconda python=3.14 tk=9.0.4

Then, notice that the conda equivalent works just fine (proof):

conda create -y -n test_env -c anaconda python=3.14 tk=9.0.4

Full GH Actions reproducer showing that it works with conda but not micromamba

name: tk-debug

on:
  push:

jobs:
  with-conda:
    runs-on: windows-latest

    steps:
      - name: Checkout source
        uses: actions/checkout@v4

      - name: Set up conda
        uses: conda-incubator/setup-miniconda@v4
      
      - name: Install python and tk
        shell: bash -l {0}
        run: conda create -y -n test_env -c anaconda python=3.14 tk=9.0.4

      - name: List installed package versions
        shell: bash -l {0}
        run: conda list

  with-mamba:
    runs-on: windows-latest

    steps:
      - name: Checkout source
        uses: actions/checkout@v4

      - name: Set up mamba
        uses: mamba-org/setup-micromamba@v3
      
      - name: Install python and tk
        shell: bash -l {0}
        run: micromamba create -y -n test_env -c anaconda python=3.14 tk=9.0.4

      - name: List installed package versions
        shell: bash -l {0}
        run: micromamba list

Finally, read the docs and see that micromamba does not support the anaconda default channels:

https://mamba.readthedocs.io/en/latest/user_guide/troubleshooting.html#using-the-defaults-channels

So: pinning an older tk might fix the immediate problem, but is still unsupported usage. I think we should consider either using micromamba as intended (i.e. using only conda-forge) or switch back to conda so the main channel is supported.

[mikofski]

Only use conda-forge