Fix squeeze dimension in test logits calculation

[JulianAssmann]

Fixes #27.

Issue
SAP_RPT_OSS_Regressor.predict() raised an error when called with a single-row DataFrame during regression tasks:

The root cause is In sap_rpt_oss/model/torch_model.py, line 361, where .squeeze() is called without specifying a dimension. When there is only one test sample, this removes all size-1 dimensions, collapsing the tensor to a 0-d scalar. The subsequent np.concatenate call then fails because it cannot concatenate zero-dimensional arrays.

Fix

# Before
test_logits = logits[test_mask].squeeze()

# After
test_logits = logits[test_mask].squeeze(-1)

Reproduce bug and verify fix

import numpy as np
import pandas as pd
from sklearn.datasets import fetch_california_housing
from sklearn.model_selection import train_test_split
import sap_rpt_oss

# Load data
data = fetch_california_housing()
X = pd.DataFrame(data.data, columns=data.feature_names)
y = data.target

X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=500, random_state=42)

# Train model
model = sap_rpt_oss.SAP_RPT_OSS_Regressor()
model.fit(X_train, y_train)

# This works (multiple samples)
preds = model.predict(X_test.iloc[:5])
print(f"Multiple samples: {preds}")

# This works after the fix
single_pred = model.predict(X_test.iloc[[0]])
print(f"Single sample: {single_pred}")

[cla-assistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[cla-assistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}