deepchecks - npm Package Compare versions

Comparing version 0.18.1 to 0.19.0

deepchecks.egg-info/PKG-INFO

 Metadata-Version: 2.1
 Name: deepchecks
-Version: 0.18.1
+Version: 0.19.0
 Summary: Package for validating your machine learning model and data
@@ -9,3 +9,3 @@ Home-page: https://github.com/deepchecks/deepchecks
 License: UNKNOWN
-Download-URL: https://github.com/deepchecks/deepchecks/releases/download/0.18.1/deepchecks-0.18.1.tar.gz
+Download-URL: https://github.com/deepchecks/deepchecks/releases/download/0.19.0/deepchecks-0.19.0.tar.gz
 Project-URL: Documentation, https://docs.deepchecks.com
@@ -12,0 +12,0 @@ Project-URL: Bug Reports, https://github.com/deepchecks/deepchecks

deepchecks.egg-info/requires.txt

@@ -35,3 +35,3 @@ pandas<2.2.0,>=1.1.5
 ipykernel>=5.3.0
-ipywidgets<8,>=7.6.5
+ipywidgets>=7.6.5
 jupyter-server>=2.7.2
@@ -43,6 +43,8 @@
 umap-learn
-transformers<4.37.0,>=4.0.0
+transformers>=4.0.0
+huggingface_hub
+sentence_transformers>=3.0.0
 
 [nlp-properties]
-fasttext>=0.8.0
+fasttext<0.9.3,>=0.8.0
 
@@ -61,3 +63,3 @@ [nlp:python_version < "3.7"]
 opencv-python>=4.5.5.62
-albumentations>=1.1.0
+albumentations<1.4.0,>=1.1.0
 imgaug>=0.4.0
@@ -64,0 +66,0 @@ seaborn>=0.1.0

deepchecks/core/display.py

@@ -22,3 +22,4 @@ # ----------------------------------------------------------------------------
 
-from IPython.core.display import display, display_html
+from IPython.core.display import display_html
+from IPython.display import display
 from ipywidgets import Widget
@@ -109,3 +110,3 @@
         elif as_widget is True:
-            display_html(self.widget_serializer.serialize(
+            display(self.widget_serializer.serialize(
                 output_id=unique_id,
@@ -112,0 +113,0 @@ **kwargs

deepchecks/core/serialization/check_result/widget.py

@@ -252,5 +252,7 @@ # ----------------------------------------------------------------------------
         children = []
+        titles = []  # Create a list to store titles
 
-        for i, (name, display) in enumerate(item.items()):
-            tab.set_title(i, name)
+        # First collect all children and titles
+        for name, display in item.items():
+            titles.append(name)  # Add title to list
             children.append(VBox(children=cls.handle_display(
@@ -263,3 +265,6 @@ display,
 
+        # Set all properties at once
         tab.children = children
+        tab.titles = tuple(titles)  # Set all titles at once
+
         style = '<style>.jupyter-widgets.widget-tab > .p-TabBar .p-TabBar-tab {min-width: fit-content;}</style>'
@@ -266,0 +271,0 @@ return VBox(children=[

deepchecks/core/serialization/suite_result/widget.py

@@ -118,7 +118,8 @@ # ----------------------------------------------------------------------------
         ])
-        return Accordion(
+        accordion = Accordion(
             children=[content],
-            _titles={'0': self.value.name},
             selected_index='0'
         )
+        accordion.set_title(0, self.value.name)
+        return accordion
 
@@ -163,5 +164,5 @@ def prepare_summary(
             children=children,
-            _titles={'0': title},
             selected_index=None
         ))
+        accordion.set_title(0, title)
         return VBox(children=(
@@ -205,5 +206,5 @@ # by putting `section_anchor` before the results accordion
                 children=(HTML(value='<p>No outputs to show.</p>'),),
-                _titles={'0': title},
                 selected_index=None
             ))
+            accordion.set_title(0, title)
         else:
@@ -232,5 +233,5 @@ section_id = f'{output_id}-section-{get_random_string()}'
                 children=(VBox(children=children),),
-                _titles={'0': title},
                 selected_index=None
             ))
+            accordion.set_title(0, title)
 
@@ -237,0 +238,0 @@ return VBox(children=(

deepchecks/nlp/checks/data_integrity/unknown_tokens.py

@@ -27,2 +27,3 @@ # ----------------------------------------------------------------------------
 from deepchecks.nlp.text_data import TextData
+from deepchecks.nlp.utils.text_properties_models import check_nltk_resource
 from deepchecks.utils.numbers import round_sig
@@ -183,6 +184,6 @@ from deepchecks.utils.strings import format_list, format_percent, truncate_string
             # Choose tokenizer based on availability of nltk
-            if nltk.download('punkt', quiet=True):
+            if check_nltk_resource('punkt_tab'):
                 tokenize = nltk.word_tokenize
             else:
-                warnings.warn('nltk punkt is not available, using str.split instead to identify individual words. '
+                warnings.warn('nltk punkt_tab is not available, using str.split instead to identify individual words. '
                               'Please check your internet connection.')
@@ -189,0 +190,0 @@ tokenize = str.split

deepchecks/nlp/utils/text_properties_models.py

@@ -21,3 +21,4 @@ # ----------------------------------------------------------------------------
 import requests
-from nltk import corpus
+from nltk import corpus, data
+from nltk import download as nltk_download
 from transformers.utils import logging as transformers_logging
@@ -28,2 +29,29 @@
 
+def check_nltk_resource(resource_name: str, resource_path: Optional[str] = None) -> bool:
+    """Check if NLTK resource exists and download if it doesn't.
+
+    Parameters
+    ----------
+    resource_name : str
+        Name of the NLTK resource to check/download
+    resource_path : str, optional
+        Path to check for resource existence. If None, defaults to 'tokenizers/{resource_name}'
+
+    Returns
+    -------
+    bool
+        True if resource exists or was successfully downloaded, False otherwise
+    """
+    try:
+        # Default to tokenizers path if not specified
+        path_to_check = resource_path if resource_path else f'tokenizers/{resource_name}'
+        data.find(path_to_check)
+        return True
+    except LookupError:
+        try:
+            return nltk_download(resource_name, quiet=True)
+        except Exception:  # pylint: disable=broad-except
+            return False
+
+
 def get_create_model_storage(models_storage: Union[pathlib.Path, str, None] = None):
@@ -30,0 +58,0 @@ """Get the models storage directory and create it if needed."""

deepchecks/nlp/utils/text_properties.py

@@ -25,4 +25,2 @@ # ----------------------------------------------------------------------------
 import torch.cuda
-from nltk import corpus
-from nltk import download as nltk_download
 from nltk import sent_tokenize, word_tokenize
@@ -34,3 +32,4 @@ from tqdm import tqdm
 from deepchecks.nlp.utils.text import cut_string, hash_text, normalize_text, remove_punctuation
-from deepchecks.nlp.utils.text_properties_models import get_cmudict_dict, get_fasttext_model, get_transformer_pipeline
+from deepchecks.nlp.utils.text_properties_models import (check_nltk_resource, get_cmudict_dict, get_fasttext_model,
+                                                         get_transformer_pipeline)
 from deepchecks.utils.function import run_available_kwargs
@@ -69,4 +68,4 @@ from deepchecks.utils.strings import SPECIAL_CHARACTERS, format_list
     if hash_key not in sentences_cache:
-        if not nltk_download('punkt', quiet=True):
-            _warn_if_missing_nltk_dependencies('punkt', 'property')
+        if not check_nltk_resource('punkt_tab'):
+            _warn_if_missing_nltk_dependencies('punkt_tab', 'property')
             return None
@@ -220,3 +219,6 @@ sentences_cache[hash_key] = sent_tokenize(text)
             if len(text_to_use) == 0:
-                text_to_use = cut_string(sentences[0], MAX_CHARS)
+                if len(sentences) > 0:
+                    text_to_use = cut_string(sentences[0], MAX_CHARS)
+                else:
+                    text_to_use = None
             text_list_to_predict.append(text_to_use)
@@ -231,7 +233,10 @@ else:
                 for text in text_list_to_predict:
-                    try:
-                        v = classifier(text)[0]
-                        results.append(output_formatter(v))
-                    except Exception:  # pylint: disable=broad-except
+                    if text is None:
                         results.append(np.nan)
+                    else:
+                        try:
+                            v = classifier(text)[0]
+                            results.append(output_formatter(v))
+                        except Exception:  # pylint: disable=broad-except
+                            results.append(np.nan)
                 return results  # Return the results if prediction is successful
@@ -348,6 +353,5 @@
         return np.nan
-    if not nltk_download('punkt', quiet=True):
-        _warn_if_missing_nltk_dependencies('punkt', 'Lexical Density')
+    if not check_nltk_resource('punkt_tab'):
+        _warn_if_missing_nltk_dependencies('punkt_tab', 'Lexical Density')
         return np.nan
-
     all_words = _split_to_words_with_cache(text)
@@ -364,6 +368,7 @@ if len(all_words) == 0:
         return np.nan
-    if not nltk_download('averaged_perceptron_tagger', quiet=True):
-        _warn_if_missing_nltk_dependencies('averaged_perceptron_tagger', 'Unique Noun Count')
+    if not check_nltk_resource('averaged_perceptron_tagger_eng',
+                               'taggers/averaged_perceptron_tagger_eng'):
+        _warn_if_missing_nltk_dependencies('averaged_perceptron_tagger_eng',
+                                           'Unique Noun Count')
         return np.nan
-
     unique_words_with_tags = set(textblob.TextBlob(text).tags)
@@ -383,6 +388,7 @@ return sum(1 for (_, tag) in unique_words_with_tags if tag.startswith('N'))
     if cmudict_dict is None:
-        if not nltk_download('cmudict', quiet=True):
+        if not check_nltk_resource('cmudict', 'corpora/cmudict'):
             _warn_if_missing_nltk_dependencies('cmudict', 'Reading Ease')
             return np.nan
-        cmudict_dict = corpus.cmudict.dict()
+        else:
+            cmudict_dict = get_cmudict_dict()
     text_sentences = _sample_for_property(text, mode='sentences', limit=DEFAULT_SENTENCE_SAMPLE_SIZE,
@@ -454,10 +460,11 @@ return_as_list=True)
         return np.nan
-    if not nltk_download('punkt', quiet=True):
-        _warn_if_missing_nltk_dependencies('punkt', 'Unique Syllables Count')
+    if not check_nltk_resource('punkt_tab'):
+        _warn_if_missing_nltk_dependencies('punkt_tab', 'Unique Syllables Count')
         return np.nan
     if cmudict_dict is None:
-        if not nltk_download('cmudict', quiet=True):
+        if not check_nltk_resource('cmudict', 'corpora/cmudict'):
             _warn_if_missing_nltk_dependencies('cmudict', 'Unique Syllables Count')
             return np.nan
-        cmudict_dict = corpus.cmudict.dict()
+        else:
+            cmudict_dict = get_cmudict_dict()
 
@@ -491,4 +498,4 @@ text = remove_punctuation(text.lower())
         return np.nan
-    if not nltk_download('punkt', quiet=True):
-        _warn_if_missing_nltk_dependencies('punkt', 'Sentences Count')
+    if not check_nltk_resource('punkt_tab'):
+        _warn_if_missing_nltk_dependencies('punkt_tab', 'Sentences Count')
         return np.nan
@@ -502,10 +509,11 @@ return len(_split_to_sentences_with_cache(text))
         return np.nan
-    if not nltk_download('punkt', quiet=True):
-        _warn_if_missing_nltk_dependencies('punkt', 'Average Syllable Length')
+    if not check_nltk_resource('punkt_tab'):
+        _warn_if_missing_nltk_dependencies('punkt_tab', 'Average Syllable Length')
         return np.nan
     if cmudict_dict is None:
-        if not nltk_download('cmudict', quiet=True):
+        if not check_nltk_resource('cmudict', 'corpora/cmudict'):
             _warn_if_missing_nltk_dependencies('cmudict', 'Average Syllable Length')
             return np.nan
-        cmudict_dict = corpus.cmudict.dict()
+        else:
+            cmudict_dict = get_cmudict_dict()
     sentence_count = len(_split_to_sentences_with_cache(text))
@@ -557,3 +565,3 @@ text = remove_punctuation(text.lower())
         {'name': 'Unique Noun Count', 'method': unique_noun_count, 'output_type': 'numeric'},
-    )
+)
 
@@ -571,3 +579,3 @@ ALL_PROPERTIES: Tuple[TextProperty, ...] = \
         {'name': 'Average Syllable Length', 'method': average_syllable_length, 'output_type': 'numeric'},
-    ) + DEFAULT_PROPERTIES
+) + DEFAULT_PROPERTIES
 
@@ -761,7 +769,8 @@ LONG_RUN_PROPERTIES = ('Toxicity', 'Fluency', 'Formality', 'Unique Noun Count')
     if properties_requiring_cmudict:
-        if not nltk_download('cmudict', quiet=True):
+        if not check_nltk_resource('cmudict', 'corpora/cmudict'):
             _warn_if_missing_nltk_dependencies('cmudict', format_list(properties_requiring_cmudict))
             for prop in properties_requiring_cmudict:
                 calculated_properties[prop] = [np.nan] * len(raw_text)
-        kwargs['cmudict_dict'] = get_cmudict_dict(use_cache=cache_models)
+        else:
+            kwargs['cmudict_dict'] = get_cmudict_dict(use_cache=cache_models)
 
@@ -768,0 +777,0 @@ if 'Toxicity' in properties_types and 'toxicity_classifier' not in kwargs:

deepchecks/nlp/utils/text.py

@@ -89,3 +89,3 @@ # ----------------------------------------------------------------------------
         return text
-    if nltk.download('punkt', quiet=True):
+    if nltk.download('punkt', quiet=True) and nltk.download('punkt_tab', quiet=True):
         tokenize = word_tokenize
@@ -92,0 +92,0 @@ else:

deepchecks/ppscore.py

@@ -39,4 +39,4 @@ """PPS (Predictive Power Score) module."""
 import pandas as pd
-from pandas.api.types import (is_bool_dtype, is_categorical_dtype, is_datetime64_any_dtype, is_numeric_dtype,
-                              is_object_dtype, is_string_dtype, is_timedelta64_dtype)
+from pandas.api.types import (is_bool_dtype, is_datetime64_any_dtype, is_numeric_dtype, is_object_dtype,
+                              is_string_dtype, is_timedelta64_dtype)
 from sklearn import preprocessing, tree
@@ -228,3 +228,3 @@ from sklearn.metrics import f1_score, mean_absolute_error
 
-def _dtype_represents_categories(series) -> bool:
+def _dtype_represents_categories(series: pd.Series) -> bool:
     """Determine if the dtype of the series represents categorical values."""
@@ -235,3 +235,3 @@ return (
             or is_string_dtype(series)
-            or is_categorical_dtype(series)
+            or isinstance(series.dtype, pd.CategoricalDtype)
     )
@@ -238,0 +238,0 @@

deepchecks/tabular/checks/data_integrity/mixed_nulls.py

@@ -17,3 +17,2 @@ # ----------------------------------------------------------------------------
 import pandas as pd
-from pandas.api.types import is_categorical_dtype
 
@@ -109,3 +108,3 @@ from deepchecks.core import CheckResult, ConditionCategory, ConditionResult
             column_data = df[column_name]
-            if is_categorical_dtype(column_data) is True:
+            if isinstance(column_data.dtype, pd.CategoricalDtype):
                 # NOTE:
@@ -112,0 +111,0 @@ # 'pandas.Series.value_counts' and 'pandas.Series.apply'

deepchecks/tabular/datasets/regression/airbnb.py

@@ -93,4 +93,6 @@ # ----------------------------------------------------------------------------
 
-_TRAIN_DATA_URL = 'https://drive.google.com/uc?export=download&id=1UWkr1BQlyyUkbsW5hHIFTr-x0evZE3Ie'
-_TEST_DATA_URL = 'https://drive.google.com/uc?export=download&id=1lfpWVtDktrnsLUzCN1tkRc1jRbguEz3a'
+_TRAIN_DATA_URL = ('https://raw.githubusercontent.com/deepchecks/deepchecks-datasets/'
+                   '8dd24134239b9df5d2a3a13cdce38cc22caaaaf4/airbnb_ref_data.csv')
+_TEST_DATA_URL = ('https://raw.githubusercontent.com/deepchecks/deepchecks-datasets/'
+                  '8dd24134239b9df5d2a3a13cdce38cc22caaaaf4/airbnb_prod_data.csv')
 _target = 'price'
@@ -97,0 +99,0 @@ _predictions = 'predictions'

deepchecks/tabular/metric_utils/scorers.py

@@ -264,2 +264,3 @@ # ----------------------------------------------------------------------------
                 self.predictions = pd.Series(predictions, index=data.index)
+                self._estimator_type = 'classifier'
 
@@ -266,0 +267,0 @@ def predict(self, data: pd.DataFrame) -> np.ndarray:

deepchecks/utils/abstracts/weak_segment_abstract.py

@@ -325,3 +325,4 @@ # ----------------------------------------------------------------------------
                 if scorer is not None and dummy_model is not None and label_col is not None:
-                    leaf_data, leaf_labels = leaf_filter.filter(data_for_search, label_col_for_search)
+                    leaf_data = leaf_filter.filter(data_for_search)
+                    leaf_labels = label_col_for_search.loc[leaf_data.index]
                     leaf_score = scorer.run_on_data_and_label(dummy_model, leaf_data, leaf_labels)
@@ -345,3 +346,3 @@ else:  # if no scorer is provided, use the average loss_per_sample of samples in the leaf as the score
         grid_searcher = GridSearchCV(DecisionTreeRegressor(random_state=random_state),
-                                     scoring=neg_worst_segment_score, param_grid=search_space, n_jobs=-1, cv=3)
+                                     scoring=neg_worst_segment_score, param_grid=search_space, n_jobs=1, cv=3)
         try:
@@ -348,0 +349,0 @@ grid_searcher.fit(data_for_search, score_per_sample_for_search)

deepchecks/utils/performance/partition.py

@@ -14,3 +14,3 @@ # ----------------------------------------------------------------------------
 from copy import deepcopy
-from typing import Callable, List, Optional, Tuple, Union
+from typing import Callable, List
 
@@ -50,16 +50,9 @@ import numpy as np
 
-    def filter(self, dataframe: pd.DataFrame, label_col: Optional[pd.Series] = None) -> \
-            Union[Tuple[pd.DataFrame, pd.Series], pd.DataFrame]:
+    def filter(self, dataframe: pd.DataFrame) -> pd.DataFrame:
         """Run the filter on given dataframe. Return rows in data frame satisfying the filter properties."""
-        if label_col is not None:
-            dataframe['temp_label_col'] = label_col
         for func in self.filter_functions:
             dataframe = dataframe.loc[func(dataframe)]
+        return dataframe
 
-        if label_col is not None:
-            return dataframe.drop(columns=['temp_label_col']), dataframe['temp_label_col']
-        else:
-            return dataframe
 
-
 class DeepchecksBaseFilter(DeepchecksFilter):
@@ -219,2 +212,3 @@ """Extend DeepchecksFilter class for feature range based filters.
     """
+    # pylint: disable=function-redefined,multiple-statements
     column = dataset.data[column_name]
@@ -233,6 +227,6 @@ if column_name in dataset.numerical_features:
             if end == percentile_values[-1]:
-                f = lambda df, a=start, b=end: (df[column_name] >= a) & (df[column_name] <= b)
+                def f(df, a=start, b=end): return (df[column_name] >= a) & (df[column_name] <= b)
                 label = f'[{format_number(start)} - {format_number(end)}]'
             else:
-                f = lambda df, a=start, b=end: (df[column_name] >= a) & (df[column_name] < b)
+                def f(df, a=start, b=end): return (df[column_name] >= a) & (df[column_name] < b)
                 label = f'[{format_number(start)} - {format_number(end)})'
@@ -239,0 +233,0 @@

PKG-INFO

 Metadata-Version: 2.1
 Name: deepchecks
-Version: 0.18.1
+Version: 0.19.0
 Summary: Package for validating your machine learning model and data
@@ -9,3 +9,3 @@ Home-page: https://github.com/deepchecks/deepchecks
 License: UNKNOWN
-Download-URL: https://github.com/deepchecks/deepchecks/releases/download/0.18.1/deepchecks-0.18.1.tar.gz
+Download-URL: https://github.com/deepchecks/deepchecks/releases/download/0.19.0/deepchecks-0.19.0.tar.gz
 Project-URL: Documentation, https://docs.deepchecks.com
@@ -12,0 +12,0 @@ Project-URL: Bug Reports, https://github.com/deepchecks/deepchecks

README.md

@@ -18,7 +18,6 @@ <!--
 ![pkgVersion](https://img.shields.io/pypi/v/deepchecks)
-[![Maintainability](https://api.codeclimate.com/v1/badges/970b11794144139975fa/maintainability)](https://codeclimate.com/github/deepchecks/deepchecks/maintainability)
 [![Coverage
 Status](https://coveralls.io/repos/github/deepchecks/deepchecks/badge.svg?branch=main)](https://coveralls.io/github/deepchecks/deepchecks?branch=main) 
 <!-- ALL-CONTRIBUTORS-BADGE:START - Do not remove or modify this section -->
-[![All Contributors](https://img.shields.io/badge/all_contributors-43-orange.svg?style=flat-round)](#https://github.com/deepchecks/deepchecks/blob/main/CONTRIBUTING.rst)
+[![All Contributors](https://img.shields.io/badge/all_contributors-44-orange.svg?style=flat-round)](#https://github.com/deepchecks/deepchecks/blob/main/CONTRIBUTING.rst)
 <!-- ALL-CONTRIBUTORS-BADGE:END --> 
@@ -369,2 +368,3 @@
       <td align="center" valign="top" width="14.28%"><a href="https://www.kaggle.com/rayanaay"><img src="https://avatars.githubusercontent.com/u/55285736?v=4?s=100" width="100px;" alt="AIT ALI YAHIA Rayane"/><br /><sub><b>AIT ALI YAHIA Rayane</b></sub></a><br /><a href="#code-RayanAAY-ops" title="Code">💻</a> <a href="#ideas-RayanAAY-ops" title="Ideas, Planning, & Feedback">🤔</a></td>
+      <td align="center" valign="top" width="14.28%"><a href="https://github.com/chris-santiago"><img src="https://avatars.githubusercontent.com/u/40875112?v=4?s=100" width="100px;" alt="Chris Santiago"/><br /><sub><b>Chris Santiago</b></sub></a><br /><a href="#bug-chris-santiago" title="Bug reports">🐛</a> <a href="#code-chris-santiago" title="Code">💻</a></td>
     </tr>
@@ -371,0 +371,0 @@ </tbody>

requirements/dev-requirements.txt

@@ -24,3 +24,4 @@ twine
 
-catboost
+catboost; python_version >= '3.7'
+catboost<=1.2.2; python_version < '3.7'
 lightgbm
@@ -37,3 +38,4 @@ xgboost<=1.7.5
 numpy>=1.18.5
-opencv-python>=4.1.2
+opencv-python>=4.1.2; python_version > '3.6'
+opencv-python<=4.6.0.66; python_version == '3.6'
 pillow>=7.1.2; python_version < '3.8'
@@ -40,0 +42,0 @@ pillow>=10.0.1; python_version >= '3.8' # not directly required, pinned by Snyk to avoid a vulnerability

requirements/nlp-prop-requirements.txt

@@ -1,1 +0,1 @@
-fasttext>=0.8.0
+fasttext>=0.8.0, <0.9.3

requirements/nlp-requirements.txt

@@ -6,3 +6,5 @@ seqeval>=1.0.0
 umap-learn
-transformers>=4.0.0,<4.37.0
+transformers>=4.0.0
+huggingface_hub
+sentence_transformers>=3.0.0
 tiktoken; python_version >= '3.8'

requirements/requirements.txt

@@ -11,3 +11,3 @@ pandas>=1.1.5,<2.2.0
 ipykernel>=5.3.0; python_version >= '3.8'
-ipywidgets>=7.6.5,<8; python_version >= '3.8'
+ipywidgets>=7.6.5; python_version >= '3.8'
 
@@ -14,0 +14,0 @@ # google colab requirements (python 3.7)

requirements/vision-requirements.txt

 pytorch-ignite>=0.4.8
 opencv-python>=4.5.5.62
-albumentations>=1.1.0
+albumentations>=1.1.0,<1.4.0
 imgaug>=0.4.0
@@ -5,0 +5,0 @@ seaborn>=0.1.0

VERSION

@@ -1,1 +0,1 @@
-0.18.1
+0.19.0

Alert delta unavailable

Currently unable to show alert delta for PyPI packages.

Improved metrics

Total package byte prevSize

19922140

0.01%

Lines of code

70247

0.06%