Time Series Multiple forecast - names

[ngupta23]

Is your feature request related to a problem? Please describe.
With the March release of time series module, we have the ability to plot multiple forecasts in a single plot. This works ok if we have different models (such as from the output of compare_models). But sometimes, we want to compare different versions of the same model in which case, we get the same legend

from pycaret.datasets import get_data
from pycaret.time_series import TSForecastingExperiment

#### Load data and simulate missing values ----
data = get_data("airline")

#### Setup Forecasting Experiment ----
exp = TSForecastingExperiment()
exp.setup(data=data, fh=12, session_id=42)

model = exp.create_model("exp_smooth")
tuned = exp.tune_model(model)
final = exp.finalize_model(tuned)

exp.plot_model([model, tuned, final])

Describe the solution you'd like
Allow user to pass the model names through a generic argument called labels which can possibly be repurposed for various plot types as per the needs of the plots.

exp.plot_model([model, tuned, final], labels = ["baseline", "tuned", "finalized"])

Also add checks to make sure that if labels are passed, then they are the same length as the estimators

Describe alternatives you've considered
None available at this time

[ngupta23]

Aligned in 2022/0409 meeting with @moezali1 and @Yard1

Use data_kwargs = {"labels": ["baseline", "tuned", "finalized"]} instead of adding an additional argument.