Skip to content

xpd api

Show a public symbol's parameters as the source writes them, with its summary, members and guarding rules Use xpd api to read how to call a library symbol without importing the library. It prints one symbol, method, rule or recipe in full, with each parameter as the source writes it, the first paragraph of the docstring, the public methods and the rules that may apply to it. The name is one that xpd find returns, or a method or property of a class.

Terminal window
xpd api [OPTIONS] <NAME>

xpd api --help ends with these lines.

Examples:
xpd api sklego.model_selection.TimeGapSplit
xpd api mapie.regression.TimeSeriesRegressor.update
xpd api pandas.DataFrame.merge --format json

<NAME>

A dotted name, such as pandas.DataFrame or pandas.DataFrame.merge.

Required.

--format <FORMAT>

Print text, or JSON with the entry's parameters, members and rules.

Default: text.

Values: text, json.

--index <INDEX>

Additional index generated from local source.

Terminal window
$ xpd api sklego.meta.GroupedPredictor
sklego.meta.GroupedPredictor(estimator, groups, shrinkage=None, use_global_model=True, check_X=True, shrinkage_kwargs=None) [class]
`GroupedPredictor` is a meta-estimator that fits a separate estimator for each group in the input data.
decision_function(X) Predict confidence scores for samples in `X`.
fit(X, y) Fit one estimator for each group of training data `X` and `y`.
predict(X) Predict target values on new data `X` by predicting on each group. If a group is not found during `.predict()` and `use_global_model=True` the fallback estimator will be used. If `use_global_model=False` a `ValueError` will be raised.
predict_proba(X) Predict probabilities on new data `X`.
Rules: sklego/grouped-shrinkage-arguments

Add a member's name to its class to read one method or property. A property has no parameter list.

Terminal window
$ xpd api arch.unitroot.KPSS.pvalue
arch.unitroot.KPSS.pvalue [property]
The p-value for the test statistic.

--format json prints a list with the one entry. parameters holds the names, and parameters_written holds each parameter as the source writes it, with its type annotation and default. A stub file writes ... for a default, which says only that the parameter is optional.

Terminal window
$ xpd api sklearn.model_selection.TimeSeriesSplit.split --format json
[{"kind":"symbol","package":"sklearn","name":"sklearn.model_selection.TimeSeriesSplit.split","callable":"method","parameters":["X","y","groups"],"parameters_written":["X","y=None","groups=None"],"summary":"Generate indices to split data into training and test set.","members":[],"rules":[]}]

The output says how to call the symbol, and not what each parameter means, so read the library's documentation for that. A rule link means the last part of the symbol's name appears in the rule's patterns, so the rule may apply.

Exit code Meaning
0 The entry was printed
2 No entry has that name, or xpd could not run
Terminal window
$ xpd api no.such.name
xpd: no entry named no.such.name