Developing scikit-learn estimators#

Whether you are proposing an estimator for inclusion in scikit-learn, developing a separate package compatible with scikit-learn, or implementing custom components for your own projects, this chapter details how to develop objects that safely interact with scikit-learn pipelines and model selection tools.

This section details the public API you should use and implement for a scikit-learn compatible estimator. Inside scikit-learn itself, we experiment and use some private tools and our goal is always to make them public once they are stable enough, so that you can also use them in your own projects.

APIs of scikit-learn objects#

There are two major types of estimators. You can think of the first group as simple estimators, which consists most estimators, such as LogisticRegression or RandomForestClassifier. And the second group are meta-estimators, which are estimators that wrap other estimators. Pipeline and GridSearchCV are two examples of meta-estimators.

Here we start with a few vocabulary, and then we illustrate how you can implement your own estimators.

Elements of the scikit-learn API are described more definitively in the Glossary of Common Terms and API Elements.

Different objects#

The main objects in scikit-learn are (one class can implement multiple interfaces):

Estimator:

The base object, implements a fit method to learn from data, either:

estimator = estimator.fit(data, targets)

or:

estimator = estimator.fit(data)
Predictor:

For supervised learning, or some unsupervised problems, implements:

prediction = predictor.predict(data)

Classification algorithms usually also offer a way to quantify certainty of a prediction, either using decision_function or predict_proba:

probability = predictor.predict_proba(data)
Transformer:

For modifying the data in a supervised or unsupervised way (e.g. by adding, changing, or removing columns, but not by adding or removing rows). Implements:

new_data = transformer.transform(data)

When fitting and transforming can be performed much more efficiently together than separately, implements:

new_data = transformer.fit_transform(data)
Model:

A model that can give a goodness of fit measure or a likelihood of unseen data, implements (higher is better):

score = model.score(data)

Estimators#

The API has one predominant object: the estimator. An estimator is an object that fits a model based on some training data and is capable of inferring some properties on new data. It can be, for instance, a classifier or a regressor. All estimators implement the fit method:

estimator.fit(X, y)

Out of all the methods that an estimator implements, fit is usually the one you want to implement yourself. Other methods such as set_params, get_params, etc. are implemented in BaseEstimator, which you should inherit from. You might need to inherit from more mixins, which we will explain later.

Instantiation#

This concerns the creation of an object. The object’s __init__ method might accept constants as arguments that determine the estimator’s behavior (like the alpha constant in SGDClassifier). It should not, however, take the actual training data as an argument, as this is left to the fit() method:

clf2 = SGDClassifier(alpha=2.3)
clf3 = SGDClassifier([[1, 2], [2, 3]], [-1, 1]) # WRONG!

Ideally, the arguments accepted by __init__ should all be keyword arguments with a default value. In other words, a user should be able to instantiate an estimator without passing any arguments to it. In some cases, where there are no sane defaults for an argument, they can be left without a default value. In scikit-learn itself, we have very few places, only in some meta-estimators, where the sub-estimator(s) argument is a required argument.

Most arguments correspond to hyperparameters describing the model or the optimisation problem the estimator tries to solve. Other parameters might define how the estimator behaves, e.g. defining the location of a cache to store some data. These initial arguments (or parameters) are always remembered by the estimator. Also note that they should not be documented under the “Attributes” section, but rather under the “Parameters” section for that estimator.

In addition, every keyword argument accepted by __init__ should correspond to an attribute on the instance. Scikit-learn relies on this to find the relevant attributes to set on an estimator when doing model selection.

To summarize, an __init__ should look like:

def __init__(self, param1=1, param2=2):
    self.param1 = param1
    self.param2 = param2

There should be no logic, not even input validation, and the parameters should not be changed; which also means ideally they should not be mutable objects such as lists or dictionaries. If they’re mutable, they should be copied before being modified. The corresponding logic should be put where the parameters are used, typically in fit. The following is wrong:

def __init__(self, param1=1, param2=2, param3=3):
    # WRONG: parameters should not be modified
    if param1 > 1:
        param2 += 1
    self.param1 = param1
    # WRONG: the object's attributes should have exactly the name of
    # the argument in the constructor
    self.param3 = param2

The reason for postponing the validation is that if __init__ includes input validation, then the same validation would have to be performed in set_params, which is used in algorithms like GridSearchCV.

Also it is expected that parameters with trailing _ are not to be set inside the __init__ method. More details on attributes that are not init arguments come shortly.

Fitting#

The next thing you will probably want to do is to estimate some parameters in the model. This is implemented in the fit() method, and it’s where the training happens. For instance, this is where you have the computation to learn or estimate coefficients for a linear model.

The fit() method takes the training data as arguments, which can be one array in the case of unsupervised learning, or two arrays in the case of supervised learning. Other metadata that come with the training data, such as sample_weight, can also be passed to fit as keyword arguments.

Note that the model is fitted using X and y, but the object holds no reference to X and y. There are, however, some exceptions to this, as in the case of precomputed kernels where this data must be stored for use by the predict method.

Parameters

X

array-like of shape (n_samples, n_features)

y

array-like of shape (n_samples,)

kwargs

optional data-dependent parameters

The number of samples, i.e. X.shape[0] should be the same as y.shape[0]. If this requirement is not met, an exception of type ValueError should be raised.

y might be ignored in the case of unsupervised learning. However, to make it possible to use the estimator as part of a pipeline that can mix both supervised and unsupervised transformers, even unsupervised estimators need to accept a y=None keyword argument in the second position that is just ignored by the estimator. For the same reason, fit_predict, fit_transform, score and partial_fit methods need to accept a y argument in the second place if they are implemented.

The method should return the object (self). This pattern is useful to be able to implement quick one liners in an IPython session such as:

y_predicted = SGDClassifier(alpha=10).fit(X_train, y_train).predict(X_test)

Depending on the nature of the algorithm, fit can sometimes also accept additional keywords arguments. However, any parameter that can have a value assigned prior to having access to the data should be an __init__ keyword argument. Ideally, fit parameters should be restricted to directly data dependent variables. For instance a Gram matrix or an affinity matrix which are precomputed from the data matrix X are data dependent. A tolerance stopping criterion tol is not directly data dependent (although the optimal value according to some scoring function probably is).

When fit is called, any previous call to fit should be ignored. In general, calling estimator.fit(X1) and then estimator.fit(X2) should be the same as only calling estimator.fit(X2). However, this may not be true in practice when fit depends on some random process, see random_state. Another exception to this rule is when the hyper-parameter warm_start is set to True for estimators that support it. warm_start=True means that the previous state of the trainable parameters of the estimator are reused instead of using the default initialization strategy.

Estimated Attributes#

According to scikit-learn conventions, attributes which you’d want to expose to your users as public attributes and have been estimated or learned from the data must always have a name ending with trailing underscore, for example the coefficients of some regression estimator would be stored in a coef_ attribute after fit has been called. Similarly, attributes that you learn in the process and you’d like to store yet not expose to the user, should have a leading underscure, e.g. _intermediate_coefs. You’d need to document the first group (with a trailing underscore) as “Attributes” and no need to document the second group (with a leading underscore).

The estimated attributes are expected to be overridden when you call fit a second time.

Universal attributes#

Estimators that expect tabular input should set a n_features_in_ attribute at fit time to indicate the number of features that the estimator expects for subsequent calls to predict or transform. See SLEP010 for details.

Similarly, if estimators are given dataframes such as pandas or polars, they should set a feature_names_in_ attribute to indicate the features names of the input data, detailed in SLEP007. Using validate_data would automatically set these attributes for you.

Rolling your own estimator#

If you want to implement a new estimator that is scikit-learn compatible, there are several internals of scikit-learn that you should be aware of in addition to the scikit-learn API outlined above. You can check whether your estimator adheres to the scikit-learn interface and standards by running check_estimator on an instance. The parametrize_with_checks pytest decorator can also be used (see its docstring for details and possible interactions with pytest):

>>> from sklearn.utils.estimator_checks import check_estimator
>>> from sklearn.tree import DecisionTreeClassifier
>>> check_estimator(DecisionTreeClassifier())  # passes

The main motivation to make a class compatible to the scikit-learn estimator interface might be that you want to use it together with model evaluation and selection tools such as GridSearchCV and Pipeline.

Before detailing the required interface below, we describe two ways to achieve the correct interface more easily.

And you can check that the above estimator passes all common checks:

>>> from sklearn.utils.estimator_checks import check_estimator
>>> check_estimator(TemplateClassifier())  # passes

get_params and set_params#

All scikit-learn estimators have get_params and set_params functions.

The get_params function takes no arguments and returns a dict of the __init__ parameters of the estimator, together with their values.

It take one keyword argument, deep, which receives a boolean value that determines whether the method should return the parameters of sub-estimators (only relevant for meta-estimators). The default value for deep is True. For instance considering the following estimator:

>>> from sklearn.base import BaseEstimator
>>> from sklearn.linear_model import LogisticRegression
>>> class MyEstimator(BaseEstimator):
...     def __init__(self, subestimator=None, my_extra_param="random"):
...         self.subestimator = subestimator
...         self.my_extra_param = my_extra_param

The parameter deep controls control whether or not the parameters of the subestimator should be reported. Thus when deep=True, the output will be:

>>> my_estimator = MyEstimator(subestimator=LogisticRegression())
>>> for param, value in my_estimator.get_params(deep=True).items():
...     print(f"{param} -> {value}")
my_extra_param -> random
subestimator__C -> 1.0
subestimator__class_weight -> None
subestimator__dual -> False
subestimator__fit_intercept -> True
subestimator__intercept_scaling -> 1
subestimator__l1_ratio -> None
subestimator__max_iter -> 100
subestimator__multi_class -> deprecated
subestimator__n_jobs -> None
subestimator__penalty -> l2
subestimator__random_state -> None
subestimator__solver -> lbfgs
subestimator__tol -> 0.0001
subestimator__verbose -> 0
subestimator__warm_start -> False
subestimator -> LogisticRegression()

If the meta-estimator takes multiple sub-estimators, often, those sub-estimators have names (as e.g. named steps in a Pipeline object), in which case the key should become <name>__C, <name>__class_weight, etc.

When deep=False, the output will be:

>>> for param, value in my_estimator.get_params(deep=False).items():
...     print(f"{param} -> {value}")
my_extra_param -> random
subestimator -> LogisticRegression()

On the other hand, set_params takes the parameters of __init__ as keyword arguments, unpacks them into a dict of the form 'parameter': value and sets the parameters of the estimator using this dict. It returns the estimator itself.

The set_params function is used to set parameters during grid search for instance.

Cloning#

As already mentioned that when constructor arguments are mutable, they should be copied before modifying them. This also applies to constructor arguments which are estimators. That’s why meta-estimators such as GridSearchCV create a copy of the given estimator before modifying it.

However, in scikit-learn, when we copy an estimator, we get an unfitted estimator where only the constructor arguments are copied (with some exceptions, e.g. attributes related to certain internal machinery such as metadata routing).

The function responsible for this behavior is clone.

Estimators can customize the behavior of base.clone by overriding the base.BaseEstimator.__sklearn_clone__ method. __sklearn_clone__ must return an instance of the estimator. __sklearn_clone__ is useful when an estimator needs to hold on to some state when base.clone is called on the estimator. For example, FrozenEstimator makes use of this.

Estimator types#

Among simple estimators (as opposed to meta-estimators), the most common types are transformers, classifiers, regressors, and clustering algorithms.

Transformers inherit from TransformerMixin, and implement a transform method. These are estimators which take the input, and transform it in some way. Note that they should never change the number of input samples, and the output of transform should correspond to its input samples in the same given order.

Regressors inherit from RegressorMixin, and implement a predict method. They should accept numerical y in their fit method. Regressors use r2_score by default in their score method.

Classifiers inherit from ClassifierMixin. If it applies, classifiers can implement decision_function to return raw decision values, based on which predict can make its decision. If calculating probabilities is supported, classifiers can also implement predict_proba and predict_log_proba.

Classifiers should accept y (target) arguments to fit that are sequences (lists, arrays) of either strings or integers. They should not assume that the class labels are a contiguous range of integers; instead, they should store a list of classes in a classes_ attribute or property. The order of class labels in this attribute should match the order in which predict_proba, predict_log_proba and decision_function return their values. The easiest way to achieve this is to put:

self.classes_, y = np.unique(y, return_inverse=True)

in fit. This returns a new y that contains class indexes, rather than labels, in the range [0, n_classes).

A classifier’s predict method should return arrays containing class labels from classes_. In a classifier that implements decision_function, this can be achieved with:

def predict(self, X):
    D = self.decision_function(X)
    return self.classes_[np.argmax(D, axis=1)]

The multiclass module contains useful functions for working with multiclass and multilabel problems.

Clustering algorithms inherit from ClusterMixin. Ideally, they should accept a y parameter in their fit method, but it should be ignored. Clustering algorithms should set a labels_ attribute, storing the labels assigned to each sample. If applicale, they can also implement a predict method, returning the labels assigned to newly given samples.

If one needs to check the type of a given estimator, e.g. in a meta-estimator, one can check if the given object implements a transform method for transformers, and otherwise use helper functions such as is_classifier or is_regressor.

Estimator Tags#

Note

Scikit-learn introduced estimator tags in version 0.21 as a private API and mostly used in tests. However, these tags expanded over time and many third party developers also need to use them. Therefore in version 1.6 the API for the tags were revamped and exposed as public API.

The estimator tags are annotations of estimators that allow programmatic inspection of their capabilities, such as sparse matrix support, supported output types and supported methods. The estimator tags are an instance of Tags returned by the method __sklearn_tags__. These tags are used in different places, such as is_regressor or the common checks run by check_estimator and parametrize_with_checks, where tags determine which checks to run and what input data is appropriate. Tags can depend on estimator parameters or even system architecture and can in general only be determined at runtime and are therefore instance attributes rather than class attributes. See Tags for more information about individual tags.

It is unlikely that the default values for each tag will suit the needs of your specific estimator. You can change the default values by defining a __sklearn_tags__() method which returns the new values for your estimator’s tags. For example:

class MyMultiOutputEstimator(BaseEstimator):

    def __sklearn_tags__(self):
        tags = super().__sklearn_tags__()
        tags.target_tags.single_output = False
        tags.non_deterministic = True
        return tags

You can create a new subclass of Tags if you wish to add new tags to the existing set. Note that all attributes that you add in a child class need to have a default value. It can be of the form:

from dataclasses import dataclass, asdict

@dataclass
class MyTags(Tags):
    my_tag: bool = True

class MyEstimator(BaseEstimator):
    def __sklearn_tags__(self):
        tags_orig = super().__sklearn_tags__()
        as_dict = {
            field.name: getattr(tags_orig, field.name)
            for field in fields(tags_orig)
        }
        tags = MyTags(**as_dict)
        tags.my_tag = True
        return tags

Developer API for set_output#

With SLEP018, scikit-learn introduces the set_output API for configuring transformers to output pandas DataFrames. The set_output API is automatically defined if the transformer defines get_feature_names_out and subclasses base.TransformerMixin. get_feature_names_out is used to get the column names of pandas output.

base.OneToOneFeatureMixin and base.ClassNamePrefixFeaturesOutMixin are helpful mixins for defining get_feature_names_out. base.OneToOneFeatureMixin is useful when the transformer has a one-to-one correspondence between input features and output features, such as StandardScaler. base.ClassNamePrefixFeaturesOutMixin is useful when the transformer needs to generate its own feature names out, such as PCA.

You can opt-out of the set_output API by setting auto_wrap_output_keys=None when defining a custom subclass:

class MyTransformer(TransformerMixin, BaseEstimator, auto_wrap_output_keys=None):

    def fit(self, X, y=None):
        return self
    def transform(self, X, y=None):
        return X
    def get_feature_names_out(self, input_features=None):
        ...

The default value for auto_wrap_output_keys is ("transform",), which automatically wraps fit_transform and transform. The TransformerMixin uses the __init_subclass__ mechanism to consume auto_wrap_output_keys and pass all other keyword arguments to it’s super class. Super classes’ __init_subclass__ should not depend on auto_wrap_output_keys.

For transformers that return multiple arrays in transform, auto wrapping will only wrap the first array and not alter the other arrays.

See Introducing the set_output API for an example on how to use the API.

Developer API for check_is_fitted#

By default check_is_fitted checks if there are any attributes in the instance with a trailing underscore, e.g. coef_. An estimator can change the behavior by implementing a __sklearn_is_fitted__ method taking no input and returning a boolean. If this method exists, check_is_fitted simply returns its output.

See __sklearn_is_fitted__ as Developer API for an example on how to use the API.

Developer API for html representation#

Warning

The html representation API is experimental and the API is subject to change.

Estimators inheriting from BaseEstimator display a html representation of themselves in interactive programming environments such as Jupyter notebooks. For instance, we can display this html diagram:

from sklearn.base import BaseEstimator

BaseEstimator()

The raw html representation is obtained by invoking the function estimator_html_repr on an estimator instance.

To customize the URL linking to an estimator’s documentation (i.e. when clicking on the “?” icon), override the _doc_link_module and _doc_link_template attributes. In addition, you can provide a _doc_link_url_param_generator method. Set _doc_link_module to the name of the (top level) module that contains your estimator. If the value does not match the top level module name, the html representation will not contain a link to the documentation. For scikit-learn estimators this is set to "sklearn".

The _doc_link_template is used to construct the final URL. By default, it can contain two variables: estimator_module (the full name of the module containing the estimator) and estimator_name (the class name of the estimator). If you need more variables you should implement the _doc_link_url_param_generator method which should return a dictionary of the variables and their values. This dictionary will be used to render the _doc_link_template.

Coding guidelines#

The following are some guidelines on how new code should be written for inclusion in scikit-learn, and which may be appropriate to adopt in external projects. Of course, there are special cases and there will be exceptions to these rules. However, following these rules when submitting new code makes the review easier so new code can be integrated in less time.

Uniformly formatted code makes it easier to share code ownership. The scikit-learn project tries to closely follow the official Python guidelines detailed in PEP8 that detail how code should be formatted and indented. Please read it and follow it.

In addition, we add the following guidelines:

  • Use underscores to separate words in non class names: n_samples rather than nsamples.

  • Avoid multiple statements on one line. Prefer a line return after a control flow statement (if/for).

  • Use relative imports for references inside scikit-learn.

  • Unit tests are an exception to the previous rule; they should use absolute imports, exactly as client code would. A corollary is that, if sklearn.foo exports a class or function that is implemented in sklearn.foo.bar.baz, the test should import it from sklearn.foo.

  • Please don’t use import * in any case. It is considered harmful by the official Python recommendations. It makes the code harder to read as the origin of symbols is no longer explicitly referenced, but most important, it prevents using a static analysis tool like pyflakes to automatically find bugs in scikit-learn.

  • Use the numpy docstring standard in all your docstrings.

A good example of code that we like can be found here.

Input validation#

The module sklearn.utils contains various functions for doing input validation and conversion. Sometimes, np.asarray suffices for validation; do not use np.asanyarray or np.atleast_2d, since those let NumPy’s np.matrix through, which has a different API (e.g., * means dot product on np.matrix, but Hadamard product on np.ndarray).

In other cases, be sure to call check_array on any array-like argument passed to a scikit-learn API function. The exact parameters to use depends mainly on whether and which scipy.sparse matrices must be accepted.

For more information, refer to the Utilities for Developers page.

Random Numbers#

If your code depends on a random number generator, do not use numpy.random.random() or similar routines. To ensure repeatability in error checking, the routine should accept a keyword random_state and use this to construct a numpy.random.RandomState object. See sklearn.utils.check_random_state in Utilities for Developers.

Here’s a simple example of code using some of the above guidelines:

from sklearn.utils import check_array, check_random_state

def choose_random_sample(X, random_state=0):
    """Choose a random point from X.

    Parameters
    ----------
    X : array-like of shape (n_samples, n_features)
        An array representing the data.
    random_state : int or RandomState instance, default=0
        The seed of the pseudo random number generator that selects a
        random sample. Pass an int for reproducible output across multiple
        function calls.
        See :term:`Glossary <random_state>`.

    Returns
    -------
    x : ndarray of shape (n_features,)
        A random point selected from X.
    """
    X = check_array(X)
    random_state = check_random_state(random_state)
    i = random_state.randint(X.shape[0])
    return X[i]

If you use randomness in an estimator instead of a freestanding function, some additional guidelines apply.

First off, the estimator should take a random_state argument to its __init__ with a default value of None. It should store that argument’s value, unmodified, in an attribute random_state. fit can call check_random_state on that attribute to get an actual random number generator. If, for some reason, randomness is needed after fit, the RNG should be stored in an attribute random_state_. The following example should make this clear:

class GaussianNoise(BaseEstimator, TransformerMixin):
    """This estimator ignores its input and returns random Gaussian noise.

    It also does not adhere to all scikit-learn conventions,
    but showcases how to handle randomness.
    """

    def __init__(self, n_components=100, random_state=None):
        self.random_state = random_state
        self.n_components = n_components

    # the arguments are ignored anyway, so we make them optional
    def fit(self, X=None, y=None):
        self.random_state_ = check_random_state(self.random_state)

    def transform(self, X):
        n_samples = X.shape[0]
        return self.random_state_.randn(n_samples, self.n_components)

The reason for this setup is reproducibility: when an estimator is fit twice to the same data, it should produce an identical model both times, hence the validation in fit, not __init__.

Numerical assertions in tests#

When asserting the quasi-equality of arrays of continuous values, do use sklearn.utils._testing.assert_allclose.

The relative tolerance is automatically inferred from the provided arrays dtypes (for float32 and float64 dtypes in particular) but you can override via rtol.

When comparing arrays of zero-elements, please do provide a non-zero value for the absolute tolerance via atol.

For more information, please refer to the docstring of sklearn.utils._testing.assert_allclose.