CHANGES.rst

Changes

.. currentmodule:: skrub

Ongoing development

Skrub is a very recent package. It is currently undergoing fast development and backward compatibility is not ensured.

Major changes

Minor changes

skrub release 0.1.0

Major changes

• :class:`TargetEncoder` has been removed in favor of :class:`sklearn.preprocessing.TargetEncoder`, available since scikit-learn 1.3.
• :class:`Joiner` and :func:`fuzzy_join` support several ways of rescaling distances; match_score has been replaced by max_dist; bugs which prevented the Joiner to consistently vectorize inputs and accept or reject matches across calls to transform have been fixed. :pr:`821` by :user:`Jérôme Dockès <jeromedockes>`.
• :class:`InterpolationJoiner` was added to join two tables by using machine-learning to infer the matching rows from the second table. :pr:`742` by :user:`Jérôme Dockès <jeromedockes>`.
• Pipelines including :class:`TableVectorizer` can now be grid-searched, since we can now call set_params on the default transformers of :class:`TableVectorizer`. :pr:`814` by :user:`Vincent Maladiere <Vincent-Maladiere>`
• :func:`to_datetime` is now available to support pandas.to_datetime over dataframes and 2d arrays. :pr:`784` by :user:`Vincent Maladiere <Vincent-Maladiere>`
• Some parameters of :class:`Joiner` have changed. The goal is to harmonize parameters across all estimator that perform join(-like) operations, as discussed in #751. :pr:`757` by :user:`Jérôme Dockès <jeromedockes>`.
• :func:`dataframe.pd_join`, :func:`dataframe.pd_aggregate`, :func:`dataframe.pl_join` and :func:`dataframe.pl_aggregate` are now available in the dataframe submodule. :pr:`733` by :user:`Vincent Maladiere <Vincent-Maladiere>`
• :class:`FeatureAugmenter` is renamed to :class:`Joiner`. :pr:`674` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :func:`fuzzy_join` and :class:`FeatureAugmenter` can now join on datetime columns. :pr:`552` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :class:`Joiner` now supports joining on multiple column keys. :pr:`674` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• The signatures of all encoders and functions have been revised to enforce cleaner calls. This means that some arguments that could previously be passed positionally now have to be passed as keywords. :pr:`514` by :user:`Lilian Boulard <LilianBoulard>`.
• Parallelized the :class:`GapEncoder` column-wise. Parameters n_jobs and verbose added to the signature. :pr:`582` by :user:`Lilian Boulard <LilianBoulard>`
• Introducing :class:`AggJoiner`, a transformer performing aggregation on auxiliary tables followed by left-joining on a base table. :pr:`600` by :user:`Vincent Maladiere <Vincent-Maladiere>`.
• Introducing :class:`AggTarget`, a transformer performing aggregation on the target y, followed by left-joining on a base table. :pr:`600` by :user:`Vincent Maladiere <Vincent-Maladiere>`.
• Added the :class:`SelectCols` and :class:`DropCols` transformers that allow selecting a subset of a dataframe's columns inside of a pipeline. :pr:`804` by :user:`Jérôme Dockès <jeromedockes>`.

Minor changes

• :class:`DatetimeEncoder` doesn't remove constant features anymore. It also supports an 'errors' argument to raise or coerce errors during transform, and a 'add_total_seconds' argument to include the number of seconds since Epoch. :pr:`784` by :user:`Vincent Maladiere <Vincent-Maladiere>`
• Scaling of matching_score in :func:`fuzzy_join` is now between 0 and 1; it used to be between 0.5 and 1. Moreover, the division by 0 error that occurred when all rows had a perfect match has been fixed. :pr:`802` by :user:`Jérôme Dockès <jeromedockes>`.
• :class:`TableVectorizer` is now able to apply parallelism at the column level rather than the transformer level. This is the default for univariate transformers, like :class:`MinHashEncoder`, and :class:`GapEncoder`. :pr:`592` by :user:`Leo Grinsztajn <LeoGrin>`
• inverse_transform in :class:`SimilarityEncoder` now works as expected; it used to raise an exception. :pr:`801` by :user:`Jérôme Dockès <jeromedockes>`.
• :class:`TableVectorizer` propagate the n_jobs parameter to the underlying transformers except if the underlying transformer already set explicitly n_jobs. :pr:`761` by :user:`Leo Grinsztajn <LeoGrin>`, :user:`Guillaume Lemaitre <glemaitre>`, and :user:`Jerome Dockes <jeromedockes>`.
• Parallelized the :func:`deduplicate` function. Parameter n_jobs added to the signature. :pr:`618` by :user:`Jovan Stojanovic <jovan-stojanovic>` and :user:`Lilian Boulard <LilianBoulard>`
• Functions :func:`datasets.fetch_ken_embeddings`, :func:`datasets.fetch_ken_table_aliases` and :func:`datasets.fetch_ken_types` have been renamed. :pr:`602` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Make pyarrow an optional dependencies to facilitate the integration with pyodide. :pr:`639` by :user:`Guillaume Lemaitre <glemaitre>`.
• Bumped minimal required Python version to 3.10. :pr:`606` by :user:`Gael Varoquaux <GaelVaroquaux>`
• Bumped minimal required versions for the dependencies: - numpy >= 1.23.5 - scipy >= 1.9.3 - scikit-learn >= 1.2.1 - pandas >= 1.5.3 :pr:`613` by :user:`Lilian Boulard <LilianBoulard>`
• You can now pass column-specific transformers to :class:`TableVectorizer` using the specific_transformers argument. :pr:`583` by :user:`Lilian Boulard <LilianBoulard>`.
• Do not support 1-D array (and pandas Series) in :class:`TableVectorizer`. Pass a 2-D array (or a pandas DataFrame) with a single column instead. This change is for compliance with the scikit-learn API. :pr:`647` by :user:`Guillaume Lemaitre <glemaitre>`
• Fixes a bug in :class:`TableVectorizer` with remainder: it is now cloned if it's a transformer so that the same instance is not shared between different transformers. :pr:`678` by :user:`Guillaume Lemaitre <glemaitre>`
• :class:`GapEncoder` speedup :pr:`680` by :user:`Leo Grinsztajn <LeoGrin>`
  • Improved :class:`GapEncoder`'s early stopping logic. The parameters tol and min_iter have been removed. The parameter max_no_improvement can now be used to control the early stopping. :pr:`663` by :user:`Simona Maggio <simonamaggio>` :pr:`593` by :user:`Lilian Boulard <LilianBoulard>` :pr:`681` by :user:`Leo Grinsztajn <LeoGrin>`
  • Implementation improvement leading to a ~x5 speedup for each iteration.
  • Better default hyperparameters: batch_size now defaults to 1024, and max_iter_e_steps to 1.
• Removed the most_frequent and k-means strategies from the :class:`SimilarityEncoder`. These strategy were used for scalability reasons, but we recommend using the :class:`MinHashEncoder` or the :class:`GapEncoder` instead. :pr:`596` by :user:`Leo Grinsztajn <LeoGrin>`
• Removed the similarity argument from the :class:`SimilarityEncoder` constructor, as we only support the ngram similarity. :pr:`596` by :user:`Leo Grinsztajn <LeoGrin>`
• Added the analyzer parameter to the :class:`SimilarityEncoder` to allow word counts for similarity measures. :pr:`619` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• skrub now uses modern type hints introduced in PEP 585. :pr:`609` by :user:`Lilian Boulard <LilianBoulard>`
• Some bug fixes for :class:`TableVectorizer` ( :pr:`579`):
  • check_is_fitted now looks at "transformers_" rather than "columns_"
  • the default of the remainder parameter in the docstring is now "passthrough" instead of "drop" to match the implementation.
  • uint8 and int8 dtypes are now considered as numerical columns.
• Removed the leading "<" and trailing ">" symbols from KEN entities and types. :pr:`601` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Add get_feature_names_out method to :class:`MinHashEncoder`. :pr:`616` by :user:`Leo Grinsztajn <LeoGrin>`
• Removed requests from the requirements. :pr:`613` by :user:`Lilian Boulard <LilianBoulard>`
• :class:`TableVectorizer` now handles mixed types columns without failing by converting them to string before type inference. :pr:`623`by :user:`Leo Grinsztajn <LeoGrin>`
• Moved the default storage location of data to the user's home folder. :pr:`652` by :user:`Felix Lefebvre <flefebv>` and :user:`Gael Varoquaux <GaelVaroquaux>`
• Fixed bug when using :class:`TableVectorizer`'s transform method on categorical columns with missing values. :pr:`644` by :user:`Leo Grinsztajn <LeoGrin>`
• :class:`TableVectorizer` never output a sparse matrix by default. This can be changed by increasing the sparse_threshold parameter. :pr:`646` by :user:`Leo Grinsztajn <LeoGrin>`
• :class:`TableVectorizer` doesn't fail anymore if an infered type doesn't work during transform. The new entries not matching the type are replaced by missing values. :pr:`666` by :user:`Leo Grinsztajn <LeoGrin>`

• Dataset fetcher :func:`datasets.fetch_employee_salaries` now has a parameter overload_job_titles to allow overloading the job titles (employee_position_title) with the column underfilled_job_title, which provides some more information about the job title. :pr:`581` by :user:`Lilian Boulard <LilianBoulard>`

• Fix bugs which was triggered when extract_until was "year", "month", "microseconds" or "nanoseconds", and add the option to set it to None to only extract total_time, the time from epoch. :class:`DatetimeEncoder`. :pr:`743` by :user:`Leo Grinsztajn <LeoGrin>`

Before skrub: dirty_cat

Skrub was born from the dirty_cat package.

Dirty-cat release 0.4.1

Major changes

• :func:`fuzzy_join` and :class:`FeatureAugmenter` can now join on numerical columns based on the euclidean distance. :pr:`530` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :func:`fuzzy_join` and :class:`FeatureAugmenter` can perform many-to-many joins on lists of numerical or string key columns. :pr:`530` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :func:`GapEncoder.transform` will not continue fitting of the instance anymore. It makes functions that depend on it (:func:`~GapEncoder.get_feature_names_out`, :func:`~GapEncoder.score`, etc.) deterministic once fitted. :pr:`548` by :user:`Lilian Boulard <LilianBoulard>`
• :func:`fuzzy_join` and :class:`FeatureAugmenter` now perform joins on missing values as in pandas.merge but raises a warning. :pr:`522` and :pr:`529` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Added :func:`get_ken_table_aliases` and :func:`get_ken_types` for exploring KEN embeddings. :pr:`539` by :user:`Lilian Boulard <LilianBoulard>`.

Minor changes

• Improvement of date column detection and date format inference in :class:`TableVectorizer`. The format inference now tries to find a format which works for all non-missing values of the column, and only tries pandas default inference if it fails. :pr:`543` by :user:`Leo Grinsztajn <LeoGrin>` :pr:`587` by :user:`Leo Grinsztajn <LeoGrin>`

Dirty-cat Release 0.4.0

Major changes

• SuperVectorizer is renamed as :class:`TableVectorizer`, a warning is raised when using the old name. :pr:`484` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• New experimental feature: joining tables using :func:`fuzzy_join` by approximate key matching. Matches are based on string similarities and the nearest neighbors matches are found for each category. :pr:`291` by :user:`Jovan Stojanovic <jovan-stojanovic>` and :user:`Leo Grinsztajn <LeoGrin>`
• New experimental feature: :class:`FeatureAugmenter`, a transformer that augments with :func:`fuzzy_join` the number of features in a main table by using information from auxiliary tables. :pr:`409` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Unnecessary API has been made private: everything (files, functions, classes) starting with an underscore shouldn't be imported in your code. :pr:`331` by :user:`Lilian Boulard <LilianBoulard>`
• The :class:`MinHashEncoder` now supports a n_jobs parameter to parallelize the hashes computation. :pr:`267` by :user:`Leo Grinsztajn <LeoGrin>` and :user:`Lilian Boulard <LilianBoulard>`.
• New experimental feature: deduplicating misspelled categories using :func:`deduplicate` by clustering string distances. This function works best when there are significantly more duplicates than underlying categories. :pr:`339` by :user:`Moritz Boos <mjboos>`.

Minor changes

• Add example Wikipedia embeddings to enrich the data. :pr:`487` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• datasets.fetching: contains a new function :func:`get_ken_embeddings` that can be used to download Wikipedia embeddings and filter them by type.
• datasets.fetching: contains a new function :func:`fetch_world_bank_indicator` that can be used to download indicators from the World Bank Open Data platform. :pr:`291` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Removed example Fitting scalable, non-linear models on data with dirty categories. :pr:`386` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :class:`MinHashEncoder`'s :func:`minhash` method is no longer public. :pr:`379` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• Fetching functions now have an additional argument directory, which can be used to specify where to save and load from datasets. :pr:`432` by :user:`Lilian Boulard <LilianBoulard>`
• Fetching functions now have an additional argument directory, which can be used to specify where to save and load from datasets. :pr:`432` and :pr:`453` by :user:`Lilian Boulard <LilianBoulard>`
• The :class:`TableVectorizer`'s default OneHotEncoder for low cardinality categorical variables now defaults to handle_unknown="ignore" instead of handle_unknown="error" (for sklearn >= 1.0.0). This means that categories seen only at test time will be encoded by a vector of zeroes instead of raising an error. :pr:`473` by :user:`Leo Grinsztajn <LeoGrin>`

Bug fixes

• The :class:`MinHashEncoder` now considers None and empty strings as missing values, rather than raising an error. :pr:`378` by :user:`Gael Varoquaux <GaelVaroquaux>`

Dirty-cat Release 0.3.0

Major changes

• New encoder: :class:`DatetimeEncoder` can transform a datetime column into several numerical columns (year, month, day, hour, minute, second, ...). It is now the default transformer used in the :class:`TableVectorizer` for datetime columns. :pr:`239` by :user:`Leo Grinsztajn <LeoGrin>`

• The :class:`TableVectorizer` has seen some major improvements and bug fixes:

  • Fixes the automatic casting logic in transform.
  • To avoid dimensionality explosion when a feature has two unique values, the default encoder (:class:`~sklearn.preprocessing.OneHotEncoder`) now drops one of the two vectors (see parameter drop="if_binary").
  • fit_transform and transform can now return unencoded features, like the :class:`~sklearn.compose.ColumnTransformer`'s behavior. Previously, a RuntimeError was raised.

  :pr:`300` by :user:`Lilian Boulard <LilianBoulard>`

• Backward-incompatible change in the TableVectorizer: To apply remainder to features (with the *_transformer parameters), the value 'remainder' must be passed, instead of None in previous versions. None now indicates that we want to use the default transformer. :pr:`303` by :user:`Lilian Boulard <LilianBoulard>`

• Support for Python 3.6 and 3.7 has been dropped. Python >= 3.8 is now required. :pr:`289` by :user:`Lilian Boulard <LilianBoulard>`

• Bumped minimum dependencies:

  • scikit-learn>=0.23
  • scipy>=1.4.0
  • numpy>=1.17.3
  • pandas>=1.2.0 :pr:`299` and :pr:`300` by :user:`Lilian Boulard <LilianBoulard>`

• Dropped support for Jaro, Jaro-Winkler and Levenshtein distances.

  • The :class:`SimilarityEncoder` now exclusively uses ngram for similarities, and the similarity parameter is deprecated. It will be removed in 0.5. :pr:`282` by :user:`Lilian Boulard <LilianBoulard>`

Notes

• The transformers_ attribute of the :class:`TableVectorizer` now contains column names instead of column indices for the "remainder" columns. :pr:`266` by :user:`Leo Grinsztajn <LeoGrin>`

Dirty-cat Release 0.2.2

Bug fixes

• Fixed a bug in the :class:`TableVectorizer` causing a :class:`FutureWarning` when using the :func:`get_feature_names_out` method. :pr:`262` by :user:`Lilian Boulard <LilianBoulard>`

Dirty-cat Release 0.2.1

Major changes

• Improvements to the :class:`TableVectorizer`

  • Type detection works better: handles dates, numerics columns encoded as strings, or numeric columns containing strings for missing values.

  :pr:`238` by :user:`Leo Grinsztajn <LeoGrin>`

• :func:`get_feature_names` becomes :func:`get_feature_names_out`, following changes in the scikit-learn API. :func:`get_feature_names` is deprecated in scikit-learn > 1.0. :pr:`241` by :user:`Gael Varoquaux <GaelVaroquaux>`

• Improvements to the :class:`MinHashEncoder`

  • It is now possible to fit multiple columns simultaneously with the :class:`MinHashEncoder`. Very useful when using for instance the :func:`~sklearn.compose.make_column_transformer` function, on multiple columns.

  :pr:`243` by :user:`Jovan Stojanovic <jovan-stojanovic>`

Bug-fixes

• Fixed a bug that resulted in the :class:`GapEncoder` ignoring the analyzer argument. :pr:`242` by :user:`Jovan Stojanovic <jovan-stojanovic>`
• :class:`GapEncoder`'s get_feature_names_out now accepts all iterators, not just lists. :pr:`255` by :user:`Lilian Boulard <LilianBoulard>`
• Fixed :class:`DeprecationWarning` raised by the usage of distutils.version.LooseVersion. :pr:`261` by :user:`Lilian Boulard <LilianBoulard>`

Notes

• Remove trailing imports in the :class:`MinHashEncoder`.
• Fix typos and update links for website.
• Documentation of the :class:`TableVectorizer` and the :class:`SimilarityEncoder` improved.

Dirty-cat Release 0.2.0

Also see pre-release 0.2.0a1 below for additional changes.

Major changes

• Bump minimum dependencies:

  • scikit-learn (>=0.21.0) :pr:`202` by :user:`Lilian Boulard <LilianBoulard>`
  • pandas (>=1.1.5) ! NEW REQUIREMENT ! :pr:`155` by :user:`Lilian Boulard <LilianBoulard>`

• datasets.fetching - backward-incompatible changes to the example datasets fetchers:

  • The backend has changed: we now exclusively fetch the datasets from OpenML. End users should not see any difference regarding this.
  • The frontend, however, changed a little: the fetching functions stay the same but their return values were modified in favor of a more Pythonic interface. Refer to the docstrings of functions dirty_cat.datasets.fetch_* for more information.
  • The example notebooks were updated to reflect these changes. :pr:`155` by :user:`Lilian Boulard <LilianBoulard>`

• Backward incompatible change to :class:`MinHashEncoder`: The :class:`MinHashEncoder` now only supports two dimensional inputs of shape (N_samples, 1). :pr:`185` by :user:`Lilian Boulard <LilianBoulard>` and :user:`Alexis Cvetkov <alexis-cvetkov>`.

• Update handle_missing parameters:

  • :class:`GapEncoder`: the default value "zero_impute" becomes "empty_impute" (see doc).
  • :class:`MinHashEncoder`: the default value "" becomes "zero_impute" (see doc).

  :pr:`210` by :user:`Alexis Cvetkov <alexis-cvetkov>`.

• Add a method "get_feature_names_out" for the :class:`GapEncoder` and the :class:`TableVectorizer`, since get_feature_names will be depreciated in scikit-learn 1.2. :pr:`216` by :user:`Alexis Cvetkov <alexis-cvetkov>`

Notes

• Removed hard-coded CSV file dirty_cat/data/FiveThirtyEight_Midwest_Survey.csv.

• Improvements to the :class:`TableVectorizer`

  • Missing values are not systematically imputed anymore
  • Type casting and per-column imputation are now learnt during fitting
  • Several bugfixes

  :pr:`201` by :user:`Lilian Boulard <LilianBoulard>`

Dirty-cat Release 0.2.0a1

Version 0.2.0a1 is a pre-release. To try it, you have to install it manually using:

pip install --pre dirty_cat==0.2.0a1

or from the GitHub repository:

pip install git+https://github.com/dirty-cat/dirty_cat.git

Major changes

• Bump minimum dependencies:
  • Python (>= 3.6)
  • NumPy (>= 1.16)
  • SciPy (>= 1.2)
  • scikit-learn (>= 0.20.0)
• :class:`TableVectorizer`: Added automatic transform through the :class:`TableVectorizer` class. It transforms columns automatically based on their type. It provides a replacement for scikit-learn's :class:`~sklearn.compose.ColumnTransformer` simpler to use on heterogeneous pandas DataFrame. :pr:`167` by :user:`Lilian Boulard <LilianBoulard>`
• Backward incompatible change to :class:`GapEncoder`: The :class:`GapEncoder` now only supports two-dimensional inputs of shape (n_samples, n_features). Internally, features are encoded by independent :class:`GapEncoder` models, and are then concatenated into a single matrix. :pr:`185` by :user:`Lilian Boulard <LilianBoulard>` and :user:`Alexis Cvetkov <alexis-cvetkov>`.

Bug-fixes

• Fix get_feature_names for scikit-learn > 0.21. :pr:`216` by :user:`Alexis Cvetkov <alexis-cvetkov>`

Dirty-cat Release 0.1.1

Major changes

Bug-fixes

• RuntimeWarnings due to overflow in :class:`GapEncoder`. :pr:`161` by :user:`Alexis Cvetkov <alexis-cvetkov>`

Dirty-cat Release 0.1.0

Major changes

• :class:`GapEncoder`: Added online Gamma-Poisson factorization through the :class:`GapEncoder` class. This method discovers latent categories formed via combinations of substrings, and encodes string data as combinations of these categories. To be used if interpretability is important. :pr:`153` by :user:`Alexis Cvetkov <alexis-cvetkov>`

Bug-fixes

• Multiprocessing exception in notebook. :pr:`154` by :user:`Lilian Boulard <LilianBoulard>`

Dirty-cat Release 0.0.7

• MinHashEncoder: Added minhash_encoder.py and fast_hast.py files that implement minhash encoding through the :class:`MinHashEncoder` class. This method allows for fast and scalable encoding of string categorical variables.
• datasets.fetch_employee_salaries: change the origin of download for employee_salaries.
  • The function now return a bunch with a dataframe under the field "data", and not the path to the csv file.
  • The field "description" has been renamed to "DESCR".
• SimilarityEncoder: Fixed a bug when using the Jaro-Winkler distance as a similarity metric. Our implementation now accurately reproduces the behaviour of the python-Levenshtein implementation.
• SimilarityEncoder: Added a handle_missing attribute to allow encoding with missing values.
• TargetEncoder: Added a handle_missing attribute to allow encoding with missing values.
• MinHashEncoder: Added a handle_missing attribute to allow encoding with missing values.

Dirty-cat Release 0.0.6

• SimilarityEncoder: Accelerate SimilarityEncoder.transform, by:
  • computing the vocabulary count vectors in fit instead of transform
  • computing the similarities in parallel using joblib. This option can be turned on/off via the n_jobs attribute of the :class:`SimilarityEncoder`.
• SimilarityEncoder: Fix a bug that was preventing a :class:`SimilarityEncoder` to be created when categories was a list.
• SimilarityEncoder: Set the dtype passed to the ngram similarity to float32, which reduces memory consumption during encoding.

Dirty-cat Release 0.0.5

• SimilarityEncoder: Change the default ngram range to (2, 4) which performs better empirically.
• SimilarityEncoder: Added a most_frequent strategy to define prototype categories for large-scale learning.
• SimilarityEncoder: Added a k-means strategy to define prototype categories for large-scale learning.
• SimilarityEncoder: Added the possibility to use hashing ngrams for stateless fitting with the ngram similarity.
• SimilarityEncoder: Performance improvements in the ngram similarity.
• SimilarityEncoder: Expose a get_feature_names method.