[fix]: smooth architecture and double check

[piotrlaczkowski]

After adding new features we have some issued to be solved:

(1) Multiple casting to float in the numeric varibles:

features_specs = {
    # ======= NUMERICAL Features =========================
    "feat1": FeatureType.FLOAT_NORMALIZED,
    "feat2": FeatureType.FLOAT_RESCALED,
    # ======= CATEGORICAL Features ========================
    "feat3": FeatureType.STRING_CATEGORICAL,
    "feat4": FeatureType.INTEGER_CATEGORICAL,
    # ======= TEXT Features ========================
    "feat5": FeatureType.TEXT,
}

ppr = PreprocessingModel(
    path_data="data/my_data.csv",
    features_specs=features_specs,
    overwrite_stats=True,
    use_distribution_aware=True,  # Enable Distribution-Aware Encoding
    # distribution_aware_bins=1000, # Set number of bins for finer data encoding
)
# construct the preprocessing pipelines
ppr.build_preprocessor()

ppr.plot_model("model_architecture.png")

gives:

(2) Features selection layer (need to add to Text Features and Dates features

features_specs2 = {
    # ======= NUMERICAL Features =========================
    # _using the FeatureType
    "feat1": FeatureType.FLOAT_NORMALIZED,
    "feat2": FeatureType.FLOAT_RESCALED,
    # _using the NumericalFeature with custom attributes
    "feat3": NumericalFeature(
        name="feat3",
        feature_type=FeatureType.FLOAT_DISCRETIZED,
        bin_boundaries=[0.0, 1.0, 2.0],
    ),
    "feat4": NumericalFeature(
        name="feat4",
        feature_type=FeatureType.FLOAT,
    ),
    # directly by string name
    "feat5": "float",
    # ======= CATEGORICAL Features ========================
    # _using the FeatureType
    "feat6": FeatureType.STRING_CATEGORICAL,
    # _using the CategoricalFeature with custom attributes
    "feat7": CategoricalFeature(
        name="feat7",
        feature_type=FeatureType.INTEGER_CATEGORICAL,
        embedding_size=100,
    ),
    # ======= TEXT Features ========================
    "feat8": TextFeature(
        name="feat8",
        max_tokens=100,
        stop_words=["stop", "next"],
    ),
    # ======= DATE Features ========================
    "feat10": DateFeature(
        name="feat10",
        feature_type=FeatureType.DATE,
        date_format="%Y-%m-%d",
        output_format="year",
    ),
    # ======== CUSTOM PIPELINE ========================
    "feat9": Feature(
        name="feat9",
        feature_type=FeatureType.FLOAT_NORMALIZED,
        preprocessors=[
            tf.keras.layers.Rescaling,
            tf.keras.layers.Normalization,
        ],
        # leyers required kwargs
        scale=1,
    ),
}
ppr = PreprocessingModel(
    path_data="data/my_data.csv",
    features_specs=features_specs2,
    overwrite_stats=True,
    # FEATURES SELECTION
    feature_selection_placement="all_features", # Choose between (all_features|numeric|categorical)
    feature_selection_units=32,
    feature_selection_dropout=0.15,
)
# construct the preprocessing pipelines
ppr.build_preprocessor()

gives: