DataVec API

DataVec extends pandas.Series with linear-algebra friendly behaviour. The reference below is generated directly from the docstrings, so it stays in sync with the code.

Bases: Series

Column vector with labeled index for linear-algebra operations.

__getitem__(key)

v.getitem(k) == v[k]

v = DataVec({'a':1,'b':2}) v['a'] 1

__init__(data=None, **kwargs)

Create a DataVec.

Inherit from :meth: pd.Series.__init__.

Additional Parameters

idxnames (List of) name(s) for levels of index.

dg(sparse=True)

Return the diagonal matrix diag(v).

drop_vestigial_levels()

Drop index levels that don't vary.

inv()

Inverse of a vector defined for 1-vector case.

norm(ord=None, **kwargs)

Vector norm ‖v‖_ord (defaults to the Euclidean norm).

outer(other)

Outer product of two series (vectors).

proj(other)

Projection of self on other.

random(size, distribution='normal', *, index=None, idxnames=None, name=None, rng=None, **dist_kwargs) classmethod

Draw a random vector with optional labelled index.

Parameters

size : int Length of the vector. distribution : random-spec, default "normal" Name/tuple/callable describing the distribution. Built-in shorthands: - "normal" / "gaussian" (loc, scale) - "uniform" (low, high) - "chi2" / "chisquare" (df) - "exponential" (scale) - "bernoulli" (p) - "binomial" (n, p) - "poisson" (lam) - "standard_normal" A tuple like ("chi2", 3) maps to the appropriate parameters. Custom callables must accept size= and return an array. index : Sequence or pandas.Index, optional Values for the index; defaults to 0..size-1. idxnames : str or sequence of str, optional Name(s) applied to the index levels. name : str, optional Vector name; defaults to an auto-generated vec_* identifier. rng : numpy.random.Generator | int | None, optional RNG or seed used by the draw (falls back to default_rng()). **dist_kwargs Additional keyword arguments forwarded to the distribution.

Returns

DataVec Labelled random vector of shape (size,).

resid(other)

Residual from projection of self on other.