Latest version: v0.4.1

PyUp actively tracks 485,906 Python packages for vulnerabilities to keep your Python environments secure.

Scan your dependencies

Page 1 of 2




This release adds a dashboard functionality based on plotly dash that allows you to interactively build decision rules using parallel plots of your data.

There are many, many breaking changes, so almost all `0.2.*` code will no longer work.


New Features
- new `estimator.parallel_coordinates(X, y, rule_id)` method to plot a parallel
coordinates plot of data entering rule `rule_id`.
- new rules: `MultiRangeAndRule` and `MultiRangeOrRule`.
- new nodes: `MultiRangeAndNode` and `MultiRangeOrNode`.

Bug Fixes
- Fixes bugs with `replace_rule` and `append_rule`

- `append_rule` now also inserts in the correct position when `rule_id` is
inside a `CaseWhen` rule


Breaking Changes
- Custom rules are now defined with `__rule__` method that returns a boolean mask
instead of with `predict(X)` method.
- `DummyRule` is now called `PredictionRule`

New Features
- each rule now gets assigned a `rule_id`, which is displayed when you call
- new `score_rules(X, y)` method that shows performance of individual rules
- new `get_igraph()` method, that returns an igraph Graph object of the rules
- new `plot()` method that returns a plotly figure of the rules
- new `get_rule(rule_id)`, `replace_rule(rule_id, new_rule)` and `append_rule(rule_id, new_rule)` methods
- new `get_rule_params(rule_id)` and `set_rule_params(rule_id, **params)` methods
- new `get_rule_input(rule_id, X, y)` and `get_rule_leftover(rule_id, X, y)` to get the specific data
that either flows into a rule, or the unlabeled data that flows out of a rule.
This helps in constructing new rules as you can target it to the data
that would appear in that part of the rule graph.

- data is now split up and only non-labeled data is passed to downstream rules.


Page 1 of 2