module search_rank.search_engine_vectors

Short summary

module mlinsights.search_rank.search_engine_vectors

Implements a way to get close examples based on the output of a machine learned model.

source on GitHub

Classes

class	truncated documentation
SearchEngineVectors	Implements a kind of local search engine which looks for similar results assuming they are vectors. The class is …

Static Methods

staticmethod	truncated documentation
read_zip	Restore the features, the metadata to a SearchEngineVectors.

Methods

method	truncated documentation
__init__
__repr__	usual
_first_pass	Finds the closest n_neighbors.
_fit_knn	Fits the nearest neighbors.
_is_iterable	Tells if an objet is an iterator or not.
_prepare_fit	Stores data in the class itself.
_second_pass	Reorders the closest n_neighbors.
fit	Every vector comes with a list of metadata.
kneighbors	Searches for neighbors close to X.
to_zip	Saves the features and the metadata into a zipfile. The function does not save the k-nn.

Documentation

Implements a way to get close examples based on the output of a machine learned model.

source on GitHub

class mlinsights.search_rank.search_engine_vectors.SearchEngineVectors(**pknn)

Bases: object

Implements a kind of local search engine which looks for similar results assuming they are vectors. The class is using sklearn.neighborsNearestNeighbors to find the nearest neighbors of a vector and follows the same API. The class populates members:

• features_: vectors used to compute the neighbors

• knn_: parameters for the sklearn.neighborsNearestNeighbors

• metadata_: metadata, can be None

source on GitHub

Parameters

pknn – list of parameters, see sklearn.neighborsNearestNeighbors

source on GitHub

__init__(**pknn)

Parameters

pknn – list of parameters, see sklearn.neighborsNearestNeighbors

source on GitHub

__repr__()

usual

source on GitHub

_first_pass(X, n_neighbors=None)

Finds the closest n_neighbors.

Parameters

• X – features

• n_neighbors – number of neighbors to get (default is the value passed to the constructor)

Returns

dist, ind

dist is an array representing the lengths to points, ind contains the indices of the nearest points in the population matrix.

source on GitHub

_fit_knn()

Fits the nearest neighbors.

source on GitHub

_is_iterable(data)

Tells if an objet is an iterator or not.

source on GitHub

_prepare_fit(data=None, features=None, metadata=None, transform=None)

Stores data in the class itself.

Parameters

• data – a dataframe or None if the the features and the metadata are specified with an array and a dictionary

• features – features columns or an array

• metadata – data

• transform – transform each vector before using it

transform is a function whose signature:

def transform(vec, many):
    # Many tells is the functions receives many vectors
    # or just one (many=False).

Function transform is applied only if data is not None.

source on GitHub

_second_pass(X, dist, ind)

Reorders the closest n_neighbors.

Parameters

• X – features

• dist – array representing the lengths to points

• ind – indices of the nearest points in the population matrix

Returns

score, ind

score is an array representing the lengths to points, ind contains the indices of the nearest points in the population matrix.

source on GitHub

fit(data=None, features=None, metadata=None)

Every vector comes with a list of metadata.

Parameters

• data – a dataframe or None if the the features and the metadata are specified with an array and a dictionary

• features – features columns or an array

• metadata – data

source on GitHub

kneighbors(X, n_neighbors=None)

Searches for neighbors close to X.

Parameters

X – features

Returns

score, ind, meta

score is an array representing the lengths to points, ind contains the indices of the nearest points in the population matrix, meta is the metadata

source on GitHub

static read_zip(zipfilename, **kwargs)

Restore the features, the metadata to a SearchEngineVectors.

Parameters

• zipfilename – a zipfile.ZipFile or a filename

• zname – a filename in th zipfile

• kwargs – parameters for pandas.read_csv

Returns

SearchEngineVectors

It only works for Python 3.6+.

source on GitHub

to_zip(zipfilename, **kwargs)

Saves the features and the metadata into a zipfile. The function does not save the k-nn.

Parameters

• zipfilename – a zipfile.ZipFile or a filename

• kwargs – parameters for pandas.to_csv (for the metadata)

Returns

zipfilename

The function relies on function to_zip. It only works for Python 3.6+.

source on GitHub