pd not found error while check_estimator used for estimator test

[afzal442]

while I was working on simple_linear_reg_forecaster, I encountered an error pd not found; really hard to debug further.
here is my below code which I've been trying out

import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.base import clone
from sktime.forecasting.base import BaseForecaster
from sktime.utils.validation.forecasting import check_fh

class SimpleReg(BaseForecaster):

    _tags = {
        "ignores-exogeneous-X": True,
        "requires-fh-in-fit": False,
        "handles-missing-data": False,
    }
        
    def __init__(self, regressor=None):
        """
        Initialize the SimpleReg forecaster.

        Parameters
        ----------
        regressor : object, optional (default=None)
            Custom regressor to use for linear regression.
            If None, LinearRegression(fit_intercept=True) will be used.
        """
        self.regressor = regressor
        super(SimpleReg, self).__init__()

    def _fit(self, y, X=None, fh=None):
        """
        Fit the forecaster to the training data using linear regression.

        Parameters
        ----------
        y : pd.Series
            Target time series with which to fit the forecaster.
        X : pd.DataFrame, default=None
            Exogenous variables are ignored.
        fh : int, list or np.array, optional (default=None)
            The forecast horizon with the steps ahead to predict.

        Returns
        -------
        self : reference to self
        """
        if self.regressor is None:
            self.regressor_ = LinearRegression(fit_intercept=True)
        else:
            self.regressor_ = clone(self.regressor)

        X_sklearn = _get_X_numpy_int_from_pandas(y.index)

        # Fit regressor
        self.regressor_.fit(X_sklearn, y)
        return self

    def _predict(self, fh=None, X=None):
        """
        Make forecasts for the given forecast horizon.

        Parameters
        ----------
        fh : int, list or np.array
            The forecast horizon with the steps ahead to predict.
        X : pd.DataFrame, default=None
            Exogenous variables (ignored).

        Returns
        -------
        y_pred : pd.Series
            Point predictions for the forecast.
        """
        if not hasattr(self, "regressor_"):
            raise NotFittedError("Model must be fitted before making predictions.")

        fh = check_fh(fh)

        X_sklearn = _get_X_numpy_int_from_pandas(fh)
        y_pred_sklearn = self.regressor_.predict(X_sklearn)
        y_pred = pd.Series(y_pred_sklearn, index=fh)
        y_pred.name = self._y.name
        return y_pred

def _get_X_numpy_int_from_pandas(index):
    """Convert pandas index to numpy array of integers"""
    return np.arange(len(index)).reshape(-1, 1)

anything I am missing here. LMK Thanks

[yarnabrina]

Is it a forecaster you're trying to add? Then I doubt bug is an appropriate label. May be we should move it from issue to discussions.

Can you share how are you trying to run check_estimator? You shared your codes here, but not where it exists nor how are you running the function that leads to the error. I'll suggest to provide the following information:

1. location of this code
2. exact code you are using to test this forecaster
3. the full error traceback that you received

Plus, as bug template suggests, versions of python, sktime and other packages can come in handy, but probably optional in this case.

[afzal442]

oh sorry for the confusion! that's any forecaster which I am trying to compile Linear Reg model to be fit using our extension template.
I ran sktime check_estimator utility to test that abv estimator.

from sktime.utils.estimator_checks import check_estimator
from sktime.forecasting.simple_linear_reg_forecaster import SimpleReg
check_estimator(SimpleReg)

Hope that makes sense now. Thanks

[yarnabrina]

Can you provide location and traceback, as per points 1 and 3 above? It'll be better if you eidt the original question with full details instead of in comments.

Please take a look at MRE for future questions.

[yarnabrina]

@afzal442 Based on your comments about code location in discord, I did create a module sktime/forecasting/simple_linear_reg_forecaster.py and added above code there. Running the tests led to a lot of failures (106 passed, and 156 failed). All the failures complained about AssertionError, and I did not get any error related to pd.

Few things that you should check: are you in the right environment? does your environment have pandas? what does the error traceback show? Can you spot the line where the error originated?

Also, the estimator you are trying to create is covered by TrendForecaster I believe. Please check that one out, and confirm if what you want to add is not a duplicate.

[fkiraly]

agree, it is also labelled discussion by @afzal442 - so, let's move to discussions for now.