autogluon.tabular 1.2.1b20250311__py3-none-any.whl → 1.2.1b20250312__py3-none-any.whl

autogluon/tabular/configs/hyperparameter_configs.py

@@ -107,7 +107,6 @@ hyperparameter_config_dict = dict(
         "XGB": {},
         # 'FASTAI': {},  # FastAI gets killed if the dataset is large (400K rows).
         "AG_AUTOMM": {},
-        "VW": {},
     },
     # Hyperparameters intended to find an interpretable model which doesn't sacrifice predictive accuracy
     interpretable={

autogluon/tabular/models/__init__.py

@@ -22,6 +22,5 @@ from .tabpfn.tabpfn_model import TabPFNModel
 from .tabpfnmix.tabpfnmix_model import TabPFNMixModel
 from .tabular_nn.torch.tabular_nn_torch import TabularNeuralNetTorchModel
 from .text_prediction.text_prediction_v1_model import TextPredictorModel
-from .vowpalwabbit.vowpalwabbit_model import VowpalWabbitModel
 from .xgboost.xgboost_model import XGBoostModel
 from .xt.xt_model import XTModel

autogluon/tabular/register/_ag_model_register.py

@@ -25,7 +25,6 @@ from ..models import (
     TabPFNModel,
     TabularNeuralNetTorchModel,
     TextPredictorModel,
-    VowpalWabbitModel,
     XGBoostModel,
     XTModel,
 )
@@ -51,7 +50,6 @@ REGISTERED_MODEL_CLS_LST = [
     TabPFNModel,
     TabPFNMixModel,
     FastTextModel,
-    VowpalWabbitModel,
     GreedyWeightedEnsembleModel,
     SimpleWeightedEnsembleModel,
     RuleFitModel,

autogluon/tabular/version.py

@@ -1,4 +1,4 @@
 """This is the autogluon version file."""
 
-__version__ = "1.2.1b20250311"
+__version__ = "1.2.1b20250312"
 __lite__ = False

{autogluon.tabular-1.2.1b20250311.dist-info → autogluon.tabular-1.2.1b20250312.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.tabular
-Version: 1.2.1b20250311
+Version: 1.2.1b20250312
 Summary: Fast and Accurate ML in 3 Lines of Code
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community
@@ -41,18 +41,18 @@ Requires-Dist: scipy<1.16,>=1.5.4
 Requires-Dist: pandas<2.3.0,>=2.0.0
 Requires-Dist: scikit-learn<1.5.3,>=1.4.0
 Requires-Dist: networkx<4,>=3.0
-Requires-Dist: autogluon.core==1.2.1b20250311
-Requires-Dist: autogluon.features==1.2.1b20250311
+Requires-Dist: autogluon.core==1.2.1b20250312
+Requires-Dist: autogluon.features==1.2.1b20250312
 Provides-Extra: all
 Requires-Dist: xgboost<2.2,>=2.0; extra == "all"
-Requires-Dist: autogluon.core[all]==1.2.1b20250311; extra == "all"
+Requires-Dist: torch<2.6,>=2.2; extra == "all"
+Requires-Dist: catboost<1.3,>=1.2; extra == "all"
+Requires-Dist: einops<0.9,>=0.7; extra == "all"
+Requires-Dist: fastai<2.8,>=2.3.1; extra == "all"
 Requires-Dist: spacy<3.8; extra == "all"
+Requires-Dist: autogluon.core[all]==1.2.1b20250312; extra == "all"
 Requires-Dist: huggingface-hub[torch]; extra == "all"
-Requires-Dist: fastai<2.8,>=2.3.1; extra == "all"
-Requires-Dist: einops<0.9,>=0.7; extra == "all"
 Requires-Dist: numpy<2.0.0,>=1.25; extra == "all"
-Requires-Dist: torch<2.6,>=2.2; extra == "all"
-Requires-Dist: catboost<1.3,>=1.2; extra == "all"
 Requires-Dist: lightgbm<4.7,>=4.0; extra == "all"
 Provides-Extra: catboost
 Requires-Dist: numpy<2.0.0,>=1.25; extra == "catboost"
@@ -66,7 +66,7 @@ Requires-Dist: imodels<1.4.0,>=1.3.10; extra == "imodels"
 Provides-Extra: lightgbm
 Requires-Dist: lightgbm<4.7,>=4.0; extra == "lightgbm"
 Provides-Extra: ray
-Requires-Dist: autogluon.core[all]==1.2.1b20250311; extra == "ray"
+Requires-Dist: autogluon.core[all]==1.2.1b20250312; extra == "ray"
 Provides-Extra: skex
 Requires-Dist: scikit-learn-intelex<2025.1,>=2024.0; extra == "skex"
 Provides-Extra: skl2onnx
@@ -82,7 +82,6 @@ Requires-Dist: torch<2.6,>=2.2; extra == "tabpfnmix"
 Requires-Dist: huggingface-hub[torch]; extra == "tabpfnmix"
 Requires-Dist: einops<0.9,>=0.7; extra == "tabpfnmix"
 Provides-Extra: tests
-Requires-Dist: tabpfn<0.2,>=0.1.11; extra == "tests"
 Requires-Dist: torch<2.6,>=2.2; extra == "tests"
 Requires-Dist: huggingface-hub[torch]; extra == "tests"
 Requires-Dist: einops<0.9,>=0.7; extra == "tests"
@@ -92,9 +91,6 @@ Requires-Dist: onnxruntime<1.20.0,>=1.17.0; extra == "tests"
 Requires-Dist: onnxruntime-gpu<1.20.0,>=1.17.0; extra == "tests"
 Requires-Dist: onnx<1.18.0,>=1.13.0; platform_system != "Windows" and extra == "tests"
 Requires-Dist: onnx<1.16.2,>=1.13.0; platform_system == "Windows" and extra == "tests"
-Requires-Dist: vowpalwabbit<9.10,>=9; (python_version < "3.11" and sys_platform != "darwin") and extra == "tests"
-Provides-Extra: vowpalwabbit
-Requires-Dist: vowpalwabbit<9.10,>=9; (python_version < "3.11" and sys_platform != "darwin") and extra == "vowpalwabbit"
 Provides-Extra: xgboost
 Requires-Dist: xgboost<2.2,>=2.0; extra == "xgboost"
 

{autogluon.tabular-1.2.1b20250311.dist-info → autogluon.tabular-1.2.1b20250312.dist-info}/RECORD

@@ -1,10 +1,10 @@
-autogluon.tabular-1.2.1b20250311-py3.9-nspkg.pth,sha256=cQGwpuGPqg1GXscIwt-7PmME1OnSpD-7ixkikJ31WAY,554
+autogluon.tabular-1.2.1b20250312-py3.9-nspkg.pth,sha256=cQGwpuGPqg1GXscIwt-7PmME1OnSpD-7ixkikJ31WAY,554
 autogluon/tabular/__init__.py,sha256=2OXpJCvENRHubBTYNIPpHX93WWuFZzsJBtTZbNVHVas,400
-autogluon/tabular/version.py,sha256=ARUvydDaVcSCiD0-UKqmcnkkW1BazMjdDfNgY3_HLRE,91
+autogluon/tabular/version.py,sha256=jJrTQDH93aen31K7aA8rJsXAtKx1fUM2eMDvRjqIPVs,91
 autogluon/tabular/configs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/configs/config_helper.py,sha256=Pb2aW9Z9w77pYKPRVZ3nBzHY3KJaiEJSJ747zZcJIVk,21132
 autogluon/tabular/configs/feature_generator_presets.py,sha256=EV5Ym8VW15q92MwOUpTi7wZFS2QooM51fLg3RdUsn-M,1223
-autogluon/tabular/configs/hyperparameter_configs.py,sha256=SHR56xcHTslf5dULbl34TgcuSqZdwgI-k3zEbBg4cFE,18053
+autogluon/tabular/configs/hyperparameter_configs.py,sha256=hp8J7g5GY3Couz929f1ItawobCw-isLTZJBcLoJY348,18035
 autogluon/tabular/configs/presets_configs.py,sha256=2Jlq1X9sVmVlyUxWsZpDV7ma2TncH5Y2HXDML7x2gYc,6810
 autogluon/tabular/configs/zeroshot/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/configs/zeroshot/zeroshot_portfolio_2023.py,sha256=oKO_2nEpI_EiLaUGmNN-3kPBIp5ATndbCOaVZ1m0048,29911
@@ -16,7 +16,7 @@ autogluon/tabular/experimental/plot_leaderboard.py,sha256=BN_kB-zmOZNUYWyI7z9pF6
 autogluon/tabular/learner/__init__.py,sha256=Hhmk5WpKQHohVmI-veOaKMelKJpIdzeXrmw_DPn3DTU,63
 autogluon/tabular/learner/abstract_learner.py,sha256=0kf0huvg0nphe-lrdKtNTzdIFr14jzJPsfZDRBkKo3g,55253
 autogluon/tabular/learner/default_learner.py,sha256=hjdKbcFtIQxQ3-k1LiGOo-w5sLxIIQAyFLs3-R35aw0,24781
-autogluon/tabular/models/__init__.py,sha256=tDVqwVG9q2ctLWRouyXeYs5NiSBnOnwh3anAfZyM3jg,1099
+autogluon/tabular/models/__init__.py,sha256=fZDKUKiD9hDzEyFXXbt7_b4yADK9peREdP8QoukWukQ,1036
 autogluon/tabular/models/_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/models/_utils/rapids_utils.py,sha256=gbej9Hjn4alCWZuGN9sOLXMMAyWbgHPThTsp2feS39o,1038
 autogluon/tabular/models/_utils/torch_utils.py,sha256=dxs_KMMAOmNkRNjYf_hrzqaHIfkqn1xoKRKqCFbQ1Rk,537
@@ -130,9 +130,6 @@ autogluon/tabular/models/tabular_nn/utils/data_preprocessor.py,sha256=ypXqtxdt1q
 autogluon/tabular/models/tabular_nn/utils/nn_architecture_utils.py,sha256=tttzR5EtYcFa6sIrUG9wyegdYmYE5DPK_CiLF1-L3c8,2875
 autogluon/tabular/models/text_prediction/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/models/text_prediction/text_prediction_v1_model.py,sha256=PBN7F98qgEAO6U76rV_hxZfAmKr_XpVKjElOdBvfX8c,1090
-autogluon/tabular/models/vowpalwabbit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-autogluon/tabular/models/vowpalwabbit/vowpalwabbit_model.py,sha256=h_j33hnsuxDM0mfExpmejO2pu5lIpTm1uIEirS2OXXI,11802
-autogluon/tabular/models/vowpalwabbit/vowpalwabbit_utils.py,sha256=jZ0STjvqwKw8jJDeoo5yAXTvgwFvY8Fsz6OqSif_JGI,3677
 autogluon/tabular/models/xgboost/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/models/xgboost/callbacks.py,sha256=uynimXya07XQMBkDvec-7mXK6OfMGP6M8MiVYu8OVRI,7008
 autogluon/tabular/models/xgboost/xgboost_model.py,sha256=tTSnTzEot2JB0qEvhki4h3RdaLjEpfMs-jWKsxlJWO4,14304
@@ -146,7 +143,7 @@ autogluon/tabular/predictor/__init__.py,sha256=zCMgjxQlWpDWnr1l1xjBCiK3rWC3N3RoD
 autogluon/tabular/predictor/interpretable_predictor.py,sha256=5UeKgnMFsfY65tiO3kxfHBPr03lyswLrgdtjPhI0Y7Q,6934
 autogluon/tabular/predictor/predictor.py,sha256=jOkpypHAPrL2nsI4iypVkZV90TpMORK-G_Ixr3Kw3XQ,357182
 autogluon/tabular/register/__init__.py,sha256=7CLOTWIUho0wi4eAwhYJ5Y0PfvNCWKnRwlw3bwYoTNE,93
-autogluon/tabular/register/_ag_model_register.py,sha256=tMr-QgxgCE49tdThdSFOZaJg2D9ckDh6fiR5K4cRtvk,1564
+autogluon/tabular/register/_ag_model_register.py,sha256=JNnmL6cwL_zvObRmyuRbwYsCxDT-qrVF5PY8dzJ5U9k,1518
 autogluon/tabular/register/_model_register.py,sha256=jqSg0d89dXAAcp-OT4II90ce994ByKMMzAYmpkyaRbI,6824
 autogluon/tabular/testing/__init__.py,sha256=XrEGLmMdmRT6QHNR13M9wna57LO4O3Q4tt27Ca8omAc,79
 autogluon/tabular/testing/fit_helper.py,sha256=gVHTdAsp_lSZ_qbwjXM7aA5fI32zHj3_zXwEXC9C_ds,19586
@@ -160,11 +157,11 @@ autogluon/tabular/trainer/model_presets/presets.py,sha256=bTPGPyz07a7GG6327yO6ry
 autogluon/tabular/trainer/model_presets/presets_distill.py,sha256=MnFC2GJc6RmDBNAGbsO2XMfo3PjR8cUrZoilWW8gTYQ,3295
 autogluon/tabular/tuning/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autogluon/tabular/tuning/feature_pruner.py,sha256=9iNku8gVbYEkjuKlyITPJDicsNkoraaQOlINQq9iZlQ,6877
-autogluon.tabular-1.2.1b20250311.dist-info/LICENSE,sha256=CeipvOyAZxBGUsFoaFqwkx54aPnIKEtm9a5u2uXxEws,10142
-autogluon.tabular-1.2.1b20250311.dist-info/METADATA,sha256=ePXfAObdxt2URjcT0647Bv92vDrnGlPbJSj68hz6Cl0,14386
-autogluon.tabular-1.2.1b20250311.dist-info/NOTICE,sha256=7nPQuj8Kp-uXsU0S5so3-2dNU5EctS5hDXvvzzehd7E,114
-autogluon.tabular-1.2.1b20250311.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
-autogluon.tabular-1.2.1b20250311.dist-info/namespace_packages.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
-autogluon.tabular-1.2.1b20250311.dist-info/top_level.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
-autogluon.tabular-1.2.1b20250311.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
-autogluon.tabular-1.2.1b20250311.dist-info/RECORD,,
+autogluon.tabular-1.2.1b20250312.dist-info/LICENSE,sha256=CeipvOyAZxBGUsFoaFqwkx54aPnIKEtm9a5u2uXxEws,10142
+autogluon.tabular-1.2.1b20250312.dist-info/METADATA,sha256=er3--iygQUQAhMpOQ3Pd_WJnlXRKkGr7go0ueJRdfrA,14069
+autogluon.tabular-1.2.1b20250312.dist-info/NOTICE,sha256=7nPQuj8Kp-uXsU0S5so3-2dNU5EctS5hDXvvzzehd7E,114
+autogluon.tabular-1.2.1b20250312.dist-info/WHEEL,sha256=tZoeGjtWxWRfdplE7E3d45VPlLNQnvbKiYnx7gwAy8A,92
+autogluon.tabular-1.2.1b20250312.dist-info/namespace_packages.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
+autogluon.tabular-1.2.1b20250312.dist-info/top_level.txt,sha256=giERA4R78OkJf2ijn5slgjURlhRPzfLr7waIcGkzYAo,10
+autogluon.tabular-1.2.1b20250312.dist-info/zip-safe,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
+autogluon.tabular-1.2.1b20250312.dist-info/RECORD,,

autogluon/tabular/models/vowpalwabbit/vowpalwabbit_model.py

@@ -1,286 +0,0 @@
-from __future__ import annotations
-
-import logging
-import os
-import time
-
-import numpy as np
-import pandas as pd
-
-from autogluon.common.features.types import (
-    R_CATEGORY,
-    R_FLOAT,
-    R_INT,
-    R_OBJECT,
-    S_IMAGE_PATH,
-    S_TEXT_AS_CATEGORY,
-    S_TEXT_NGRAM,
-    S_TEXT_SPECIAL,
-)
-from autogluon.common.utils.try_import import try_import_vowpalwabbit
-from autogluon.core.constants import (
-    BINARY,
-    MULTICLASS,
-    PROBLEM_TYPES_CLASSIFICATION,
-    PROBLEM_TYPES_REGRESSION,
-    REGRESSION,
-)
-from autogluon.core.models import AbstractModel
-from autogluon.core.utils.exceptions import TimeLimitExceeded
-
-from .vowpalwabbit_utils import VWFeaturesConverter
-
-logger = logging.getLogger(__name__)
-
-
-class VowpalWabbitModel(AbstractModel):
-    """
-    VowpalWabbit Model: https://vowpalwabbit.org/
-
-    VowpalWabbit Command Line args: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Command-line-arguments
-
-    """
-    ag_key = "VW"
-    ag_name = "VowpalWabbit"
-    ag_priority = 10
-
-    model_internals_file_name = "model-internals.pkl"
-
-    # Ref: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Loss-functions
-    CLASSIFICATION_LOSS_FUNCTIONS = ["logistic", "hinge"]
-    REGRESSION_LOSS_FUNCTIONS = ["squared", "quantile", "poisson", "classic"]
-
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self._load_model = None  # Used for saving and loading internal model file
-
-    # The `_preprocess` method takes the input data and transforms it to the internal representation usable by the model.
-    # `_preprocess` is called by `preprocess` and is used during model fit and model inference.
-    def _preprocess(self, X: pd.DataFrame, is_train=False, **kwargs) -> pd.Series:
-        X = super()._preprocess(X, **kwargs)
-        if is_train:
-            self._features_converter = VWFeaturesConverter()
-            self._feature_metadata_dict = self._feature_metadata.to_dict()
-        # self._feature_metadata contains the information related to features metadata.
-        X_series = self._features_converter.convert_features_to_vw_format(X, self._feature_metadata_dict)
-        return X_series
-
-    # The `_fit` method takes the input training data (and optionally the validation data) and trains the model.
-    def _fit(
-        self,
-        X: pd.DataFrame,
-        y: pd.Series,
-        time_limit=None,
-        verbosity=2,
-        **kwargs,  # training data  # training labels
-    ):  # kwargs includes many other potential inputs, refer to AbstractModel documentation for details
-        time_start = time.time()
-        try_import_vowpalwabbit()
-        import vowpalwabbit
-
-        seed = 0  # Random seed
-
-        # Valid self.problem_type values include ['binary', 'multiclass', 'regression', 'quantile', 'softclass']
-        if self.problem_type not in PROBLEM_TYPES_REGRESSION + PROBLEM_TYPES_CLASSIFICATION:
-            raise TypeError(f"Vowpal Wabbit does not support {self.problem_type}")
-
-        # Certain parameters like passes are passed as hyperparameters but are not used
-        # while initialising the model.
-        # passes: Used as epochs
-
-        params = self._get_model_params()
-        params["loss_function"] = params.get("loss_function", self._get_default_loss_function())
-        passes = params.pop("passes")
-
-        # Make sure to call preprocess on X near the start of `_fit`.
-        # This is necessary because the data is converted via preprocess during predict, and needs to be in the same format as during fit.
-        X_series = self.preprocess(X, is_train=True)
-
-        self._validate_loss_function(loss_function=params["loss_function"])
-
-        # VW expects label from 1 to N for Binary and Multiclass classification problems
-        # AutoGluon does label encoding from 0 to N-1, hence we increment the value of y by 1
-        if self.problem_type != REGRESSION:
-            y = y.apply(lambda row: row + 1)
-        y = y.astype(str) + " "
-
-        # Concatenate y and X to get the training data in VW format
-        final_training_data = y + X_series
-        final_training_data = final_training_data.tolist()
-
-        extra_params = {
-            "cache_file": "train.cache",
-            "holdout_off": True,
-        }
-
-        if verbosity <= 3:
-            extra_params["quiet"] = True
-
-        # Initialize the model
-        if self.problem_type in PROBLEM_TYPES_CLASSIFICATION:
-            # Ref: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Predicting-probabilities#multi-class---oaa
-            extra_params["oaa"] = self.num_classes
-            extra_params["probabilities"] = True
-        self.model = vowpalwabbit.Workspace(**params, **extra_params)
-
-        time_start_fit = time.time()
-        if time_limit is not None:
-            time_limit_fit = time_limit - (time_start_fit - time_start) - 0.3  # Account for 0.3s overhead
-            if time_limit_fit <= 0:
-                raise TimeLimitExceeded
-        else:
-            time_limit_fit = None
-
-        # Train the model
-        np.random.seed(seed)
-        epoch = 0
-
-        for epoch in range(1, passes + 1):
-            # TODO: Add Early Stopping support via validation
-            self._train_single_epoch(training_data=final_training_data)
-            if time_limit_fit is not None and epoch < passes:
-                time_fit_used = time.time() - time_start_fit
-                time_fit_used_per_epoch = time_fit_used / epoch
-                time_left = time_limit_fit - time_fit_used
-                if time_left <= (time_fit_used_per_epoch * 2):
-                    logger.log(30, f"\tEarly stopping due to lack of time. Fit {epoch}/{passes} passes...")
-                    break
-
-        self.params_trained["passes"] = epoch
-
-    def _train_single_epoch(self, training_data):
-        row_order = np.arange(0, len(training_data))
-        row_order = np.random.permutation(row_order)
-        for row_i in row_order:
-            row = training_data[row_i]
-            self.model.learn(row)
-
-    def _validate_loss_function(self, loss_function):
-        # Ref: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Loss-functions
-        if loss_function:
-            if self.problem_type in PROBLEM_TYPES_CLASSIFICATION:
-                assert loss_function in self.CLASSIFICATION_LOSS_FUNCTIONS, (
-                    f"For {self.problem_type} problem, VW supports: {self.CLASSIFICATION_LOSS_FUNCTIONS}. " f"Got loss_function:{loss_function}"
-                )
-            elif self.problem_type in PROBLEM_TYPES_REGRESSION:
-                assert loss_function in self.REGRESSION_LOSS_FUNCTIONS, (
-                    f"For {self.problem_type} problem, VW supports: {self.REGRESSION_LOSS_FUNCTIONS}. " f"Got loss_function:{loss_function}"
-                )
-
-    def _get_default_loss_function(self) -> str:
-        # Ref: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Loss-functions
-        if self.problem_type in PROBLEM_TYPES_CLASSIFICATION:
-            return "logistic"
-        else:
-            return "squared"
-
-    def save(self, path: str = None, verbose=True) -> str:
-        """
-        AutoGluon by default saves the complete Abstract Model in a pickle file format.
-        This includes the internal self.model which is the actual model.
-        However, saving VW model in pickle is not possible.
-        Hence, we dump the Abstract Model by setting setting self.model as None
-        and save self.model as a separate internal file using that model's saving mechanism
-
-        :param path: path where model is to be saved
-        :param verbose: verbosity
-        :return: path where model is saved
-        """
-
-        self._load_model = self.model is not None
-        __model = self.model
-        self.model = None
-        path = super().save(path=path, verbose=verbose)
-        self.model = __model
-        # Export model
-        if self._load_model:
-            file_path = os.path.join(path, self.model_internals_file_name)
-            self.model.save(file_path)
-        self._load_model = None
-        return path
-
-    @classmethod
-    def load(cls, path: str, reset_paths=True, verbose=True):
-        """
-        There are two files which needs to be loaded.
-        First is the Abstract Model pickle dump and second is the internal model file.
-        For VW, based on different problem_type/hyperparams, loading arguments will be different
-        """
-        try_import_vowpalwabbit()
-        import vowpalwabbit
-
-        # Load Abstract Model. This is without the internal model
-        model = super().load(path, reset_paths=reset_paths, verbose=verbose)
-        params = model._get_model_params()
-        # Load the internal model file
-        if model._load_model:
-            file_path = os.path.join(path, cls.model_internals_file_name)
-
-            model_load_params = f" -i {file_path} --quiet"
-            if model.problem_type in PROBLEM_TYPES_CLASSIFICATION:
-                model_load_params += " --probabilities --loss_function=logistic"
-            if params["sparse_weights"]:
-                model_load_params += " --sparse_weights"
-
-            model.model = vowpalwabbit.Workspace(model_load_params)
-        model._load_model = None
-        return model
-
-    def _predict_proba(self, X, **kwargs):
-        # Preprocess the set of X features
-        X = self.preprocess(X, **kwargs)
-
-        y_pred_proba = np.array([self.model.predict(row) for row in X])
-        return self._convert_proba_to_unified_form(y_pred_proba)
-
-    def _get_memory_size(self) -> int:
-        # TODO: Can be improved further to make it more accurate
-        # Returning 5MB as the value
-        return int(5e6)
-
-    # The `_set_default_params` method defines the default hyperparameters of the model.
-    # User-specified parameters will override these values on a key-by-key basis.
-    def _set_default_params(self):
-        default_params = {
-            "passes": 10,  # TODO: Much better if 500+, revisit this if wanting to use VW to get strong results
-            "bit_precision": 32,
-            "ngram": 2,
-            "skips": 1,
-            "learning_rate": 1,
-            "sparse_weights": True,
-        }
-        for param, val in default_params.items():
-            self._set_default_param_value(param, val)
-
-    # The `_get_default_auxiliary_params` method defines various model-agnostic parameters such as maximum memory usage and valid input column dtypes.
-    # For most users who build custom models, they will only need to specify the valid/invalid dtypes to the model here.
-    def _get_default_auxiliary_params(self) -> dict:
-        default_auxiliary_params = super()._get_default_auxiliary_params()
-        # Ignore the below mentioned special types. Only those features that are not of the below mentioned
-        # type are passed to the model for training list are passed features
-        extra_auxiliary_params = dict(
-            valid_raw_types=[R_INT, R_FLOAT, R_CATEGORY, R_OBJECT], ignored_type_group_special=[S_IMAGE_PATH, S_TEXT_NGRAM, S_TEXT_AS_CATEGORY, S_TEXT_SPECIAL]
-        )
-        default_auxiliary_params.update(extra_auxiliary_params)
-        return default_auxiliary_params
-
-    @classmethod
-    def _get_default_ag_args(cls) -> dict:
-        default_ag_args = super()._get_default_ag_args()
-        extra_ag_args = {
-            "valid_stacker": False,
-        }
-        default_ag_args.update(extra_ag_args)
-        return default_ag_args
-
-    @classmethod
-    def supported_problem_types(cls) -> list[str] | None:
-        return ["binary", "multiclass", "regression"]
-
-    def _more_tags(self):
-        # `can_refit_full=True` because best epoch is communicated at end of `_fit`: `self.params_trained['passes'] = epoch`
-        return {"can_refit_full": True}
-
-    @classmethod
-    def _class_tags(cls):
-        return {"handles_text": True}

autogluon/tabular/models/vowpalwabbit/vowpalwabbit_utils.py

@@ -1,93 +0,0 @@
-import pandas as pd
-
-from autogluon.common.features.types import R_CATEGORY, R_FLOAT, R_INT, S_TEXT
-
-
-class VWFeaturesConverter:
-    """
-    Converts features in PandasDataFrame to VW format
-    Ref: https://github.com/VowpalWabbit/vowpal_wabbit/wiki/Input-format
-    """
-
-    PIPE = "|"
-    SPACE = " "
-
-    # TODO: Add support for different namespaces
-
-    def convert_features_to_vw_format(self, X, feature_metadata) -> pd.Series:
-        """
-        Converts features to VW format.
-        :param X: features
-        :param feature_metadata: schema of X
-        :return: Returns a series of features converted to VW format
-        """
-
-        X_out: pd.Series = None
-
-        for feature in feature_metadata:
-            raw_feature, special_feature = feature_metadata[feature]
-            if X_out is None:
-                X_out = (
-                    self.PIPE
-                    + self.SPACE
-                    + self.__generate_namespace_based_on_ml_type(X[feature], raw_feature, special_feature, feature).astype("str")
-                    + self.SPACE
-                )
-            else:
-                X_out += (
-                    "" + self.SPACE + self.__generate_namespace_based_on_ml_type(X[feature], raw_feature, special_feature, feature).astype("str") + self.SPACE
-                )
-        return X_out
-
-    def __generate_namespace_based_on_ml_type(self, input_series, raw_feature, special_feature, feature_name=None):
-        """
-        Based on the type of feature, preprocess/sanify these features so that it is in VW format
-        Only use raw text, numeric integer, numeric decimals, and category
-        Ref: https://github.com/autogluon/autogluon/blob/master/common/src/autogluon/common/features/types.py
-
-        :param input_series: A single feature as Pandas Series
-        :param raw_feature: Raw feature Type
-        :param special_feature: Special Feature Type
-        :param feature_name: Column Name of this feature
-        :return: Preprocessed Feature as a Pandas Series
-        """
-        if S_TEXT in special_feature:
-            return input_series.apply(self.__preprocess_text)
-        elif raw_feature in [R_INT, R_FLOAT]:
-            return input_series.apply(self.__numeric_namespace_generator, args=(feature_name,))
-        elif raw_feature == R_CATEGORY:
-            return input_series.apply(self.__categorical_namespace_generator, args=(feature_name,))
-        else:
-            raise ValueError(
-                f"Received unsupported raw_feature_type '{str(raw_feature)}' special_feature_type '{str(special_feature)}'"
-                f" for feature '{feature_name}' for class {self.__class__.__name__}."
-            )
-
-    def __preprocess_text(self, s) -> str:
-        if pd.isnull(s):
-            return ""
-        s = " ".join(str(s).split())
-        # Added split to remove tabs spaces since tab is used as separator
-        text = self.__sanify(s)
-        return text
-
-    def __sanify(self, s) -> str:
-        """
-        The sanify is performed because : and | are reserved by vowpal wabbit for distinguishing namespaces and numeric
-        data
-        @param s: input string
-        @returns string
-        """
-        return str(s).replace(":", ";").replace("|", "/")
-
-    def __numeric_namespace_generator(self, feature, feature_name) -> str:
-        if pd.isnull(feature):
-            return ""
-        return feature_name + ":" + str(feature)
-
-    def __categorical_namespace_generator(self, feature, feature_name) -> str:
-        if pd.isnull(feature):
-            return ""
-        else:
-            feature = str(feature).replace(" ", "_")
-            return feature_name + "=" + self.__sanify(feature)