null prediction when using future regressor and len(training set) == n_lags + n_forecasts

[adi250491]

Prerequisites

• Put an X between the brackets on this line if you have done all of the following:
  • Reproduced the problem in a new virtualenv with only neuralprophet installed, directly from github:

    git clone <copied link from github>
    cd neural_prophet
    pip install .

  • Checked the Answered Questions on the Github Discussion board: https://github.com/ourownstory/neural_prophet/discussions
    If you have the same question but the Answer does not solve your issue, please continue the conversation there.
  • Checked that your issue isn't already filed: https://github.com/ourownstory/neural_prophet/issues
    If you have the same issue but there is a twist to your situation, please add an explanation there.
  • Considered whether your bug might actually be solvable by getting a question answered:
    • Please post a package use question
    • Please post a forecasting best practice question
    • Please post an idea or feedback

Describe the bug

When using future regressor, and the training dataset length is exaclty equal to n_lags + n_forecasts, the model.predict methods returns a prediction dataframe with null values on all yhat columns.

The same training data works when future regressor is not used.

Please run the included code to reproduce the error.

To Reproduce

Steps to reproduce the behavior:

import pandas as pd
import numpy as np
import neuralprophet
demo_df = pd.DataFrame(
   {
    "ds": pd.date_range("2022-04-01", "2022-04-30"),
   "y": np.random.random(30),
   "reg": np.random.random(30),
   })

demo_model = neuralprophet.NeuralProphet(n_forecasts=15, n_lags=15)
demo_model.set_plotting_backend('matplotlib')
demo_model.add_future_regressor("reg")
demo_model.fit(demo_df)
prediction = demo_model.predict(demo_df) #<== THIS IS THE BLANK DATAFRAME
demo_model.plot(prediction) 


demo_model2 = neuralprophet.NeuralProphet(n_forecasts=15, n_lags=15)
demo_model2.set_plotting_backend('matplotlib')
# demo_model.add_future_regressor("reg")
demo_model2.fit(demo_df[['ds','y']])
prediction2 = demo_model2.predict(demo_df[['ds','y']])  #<== THIS IS THE EXPECTED BEAHVIOUR
demo_model2.plot(prediction2)

Expected behavior
The model.predict method should return a dataframe with predicted values for the n_forecasts dates

What actually happens
The model.predict method returns a dataframe with null values in all yhat columns

Environment (please complete the following information):

• Python environment Python: 3.11.9, venv
• NeuralProphet version and install method: v0.8.0 installed from PYPI with pip install neuralprophet