[Major] Separate plotting dependencies to enable running without plotting

[kneureither]

🔬 Background

Poetry can now be installed in a minimal mode by using poetry install --only main. Minimal mode does only include the main dependencies, specified under [tool.poetry.dependencies]. Matplotlib could also be excluded in the future, by moving it into the [tool.poetry.group.plotting.dependencies] in pyproject.toml.
All tests are passed, except test_model_performance.py which saves svg files.
fixes issue #1548.

🔮 Key changes

• Add a poetry group for plotting (for now it only includes plotly, matplotlib could also be added in the same manner)
• Catch plotly import in files (forecaster.py, plot_forecast_plotly.py, plot_model_parameters_plotly.py) and display warning, but proceed without plotting anything.
• status of plotly import is handled by plotting_backend which is set to "no-backend-installed"

📋 Review Checklist

• I have performed a self-review of my own code.
• I have commented my code, added docstrings and data types to function definitions.
• I have added pytests to check whether my feature / fix works.

[kneureither]

Can you give some feedback on the approach @ourownstory? Is that how you envisioned this to work? I would add further libraries to the plotting group and deal with them in the same manner as with plotly. I used a specific group instead of specifying the optional = true tag on a package level, because it is more user-friendly. What do you think about using the plotting_backend variable to check whether the corresponding modules are installed?