autowisp.evaluator module

Class Inheritance Diagram

More convenient interface to asteval interpreters.

class autowisp.evaluator.Evaluator(*data)

Bases: EvaluatorBase

Evaluator for expressions involving fields of numpy array or headers.

__init__(*data)

Get ready to evaluate expressions given data.

Parameters:

data ([dict-like]) – A mapping between variable names that will participate in the expressions to be evaluated and the value that should be used. In case of repeating keys, later entries overwrite earlier ones.

Returns:

None

class autowisp.evaluator.EvaluatorBase

Bases: Interpreter

Asteval interpreter with the symbols all AutoWISP expressions share.

__call__(*args, **kwargs)

Evaluate the expression, raising rather than returning None on error.

Asteval defaults to printing the error and returning None, which only relocates the failure: the None flows on and breaks somewhere unrelated, long after the message about the offending expression has scrolled away. Callers wanting the permissive behaviour can still pass raise_errors=False explicitly.

__init__()

Create an interpreter with AutoWISP’s symbol table.

Sub-classes should add their data to the symbol table after invoking this, so a variable named like one of the aggregates shadows the function rather than the other way around.

Returns:

None

nan_aggregates = ('nanmean', 'nanmedian', 'nanstd', 'nanvar', 'nansum', 'nanprod', 'nanmin', 'nanmax', 'nanpercentile', 'nanquantile', 'nanargmin', 'nanargmax', 'nancumsum', 'nancumprod')

removed_names = ('open', 'print')

class autowisp.evaluator.LightCurveEvaluator(lightcurve, lc_points_selection=None, **lc_substitutions)

Bases: EvaluatorBase

Evaluator for expressions involving lightcurve datasets.

__init__(lightcurve, lc_points_selection=None, **lc_substitutions)

Get ready to evaluate expressions against the given light curve.

property lc_points_selection

Getter for the poins selection property.

property lc_substitutions

Substitutions to apply to resolve lightcurve datasets.

These can be changed after creating the instance.

update_substitutions(new_substitutions)

Like update for dict but resets previously read datasets.

class autowisp.evaluator.LightCurveLookUp(key_prefix, evaluator)

Bases: object

Look up datasets under specific key prefix.

__getattr__(key_suffix)

Return the given attribute, reading dataset if needed.

__init__(key_prefix, evaluator)

Look up datasets with keys like <key_prefix>.<something>.

autowisp.evaluator.eval_iterative_rej_polynomial_fit(x, y, *args, **kwargs)

Allow using iterative_rej_polynomial_fit() in evaluators.

autowisp.evaluator.eval_iterative_rej_smoothing_spline(x, *args, **kwargs)

Allow using iterative_rej_smoothing_spline() in evaluators.

autowisp.evaluator.eval_iterative_rejection_average(*args, **kwargs)

Allow using iterative_rejection_average() in evaluators.

autowisp.evaluator.nanrank(x)

Return each entry’s rank as a fraction of the finite entries.

The smallest of n finite values ranks 1/n and the largest 1; ties share the average of the ranks they span, and NaN stays NaN without counting towards n. This is pandas’ rank(pct=True), which is what the photometric reference merit’s qnt_<name> columns used, so a merit rewritten in terms of this ranks candidates the same way.

Unlike the aggregates in EvaluatorBase.nan_aggregates, the result has one entry per input entry, so it composes element-wise with the values it ranks.

Parameters:

x (array-like) – The values to rank.

Returns:

The ranks, as floats in (0, 1], shaped like

x.

Return type:

numpy.ndarray