pybool_ir.experiments.smooth#

Classes and methods for the paper “Smooth Operators for Effective Systematic Review Queries”.

TODO: This is quite messy. Needs to be cleaned up. NB: See the bottom of this file for experiments and more details, like accessing the data.

Classes

`OracleQueryExperiment`(indexer, collection, ...)	Variant of the QueryExperiment that performs an oracle search to determine an approximately optimal theta value.
`PrecomputedQueryExperiment`(indexer, ...)	Variant of the QueryExperiment that uses precomputed results (i.e., from dense retrieval).
`PredictorQueryExperiment`(indexer, ...)	Variant of the QueryExperiment that uses a pre-trained predictor for estimating the theta parameter.
`QueryExperiment`(indexer, collection, ...)	Base class for running experiments that involve execution of each atomic node in a query.
`SmoothAND`([theta])
`SmoothANDPredictor`(clf[, theta])
`SmoothNOT`([theta])
`SmoothNOTPredictor`(clf[, theta])
`SmoothOR`([theta])
`SmoothORPredictor`(clf[, theta])
`SmoothOperator`(theta)
`SmoothOperatorPredictorMixin`(clf, *args, ...)

class pybool_ir.experiments.smooth.OracleQueryExperiment(indexer: ~pybool_ir.index.index.Indexer, collection: ~pybool_ir.experiments.collections.Collection, query_parser: ~pybool_ir.query.parser.QueryParser = <pybool_ir.query.pubmed.parser.PubmedQueryParser object>, eval_measures: ~typing.List[~ir_measures.measures.base.Measure] | None = None, run_path: ~pathlib.Path | None = None, filter_topics: ~typing.List[str] | None = None, ignore_dates: bool = False, date_field: str = 'dp')#

Bases: QueryExperiment

Variant of the QueryExperiment that performs an oracle search to determine an approximately optimal theta value.

class pybool_ir.experiments.smooth.PrecomputedQueryExperiment(indexer: ~pybool_ir.index.index.Indexer, collection: ~pybool_ir.experiments.collections.Collection, query_parser: ~pybool_ir.query.parser.QueryParser = <pybool_ir.query.pubmed.parser.PubmedQueryParser object>, eval_measures: ~typing.List[~ir_measures.measures.base.Measure] | None = None, run_path: ~pathlib.Path | None = None, filter_topics: ~typing.List[str] | None = None, ignore_dates: bool = False, date_field: str = 'dp')#

Bases: QueryExperiment

Variant of the QueryExperiment that uses precomputed results (i.e., from dense retrieval).

class pybool_ir.experiments.smooth.PredictorQueryExperiment(indexer: ~pybool_ir.index.index.Indexer, collection: ~pybool_ir.experiments.collections.Collection, query_parser: ~pybool_ir.query.parser.QueryParser = <pybool_ir.query.pubmed.parser.PubmedQueryParser object>, eval_measures: ~typing.List[~ir_measures.measures.base.Measure] | None = None, run_path: ~pathlib.Path | None = None, filter_topics: ~typing.List[str] | None = None, ignore_dates: bool = False, date_field: str = 'dp')#

Bases: QueryExperiment

Variant of the QueryExperiment that uses a pre-trained predictor for estimating the theta parameter.

class pybool_ir.experiments.smooth.QueryExperiment(indexer: ~pybool_ir.index.index.Indexer, collection: ~pybool_ir.experiments.collections.Collection, query_parser: ~pybool_ir.query.parser.QueryParser = <pybool_ir.query.pubmed.parser.PubmedQueryParser object>, eval_measures: ~typing.List[~ir_measures.measures.base.Measure] | None = None, run_path: ~pathlib.Path | None = None, filter_topics: ~typing.List[str] | None = None, ignore_dates: bool = False, date_field: str = 'dp')#

Bases: RetrievalExperiment

Base class for running experiments that involve execution of each atomic node in a query.

class pybool_ir.experiments.smooth.SmoothAND(theta: float = 1.0)#: Bases: SmoothOperator

class pybool_ir.experiments.smooth.SmoothANDPredictor(clf: DecisionTreeRegressor, theta: float = 1.0)#: Bases: SmoothOperatorPredictorMixin, SmoothAND

class pybool_ir.experiments.smooth.SmoothNOT(theta: float = 1.0)#: Bases: SmoothOperator

class pybool_ir.experiments.smooth.SmoothNOTPredictor(clf: DecisionTreeRegressor, theta: float = 1.0)#: Bases: SmoothOperatorPredictorMixin, SmoothOR

class pybool_ir.experiments.smooth.SmoothOR(theta: float = 0.0)#: Bases: SmoothOperator

class pybool_ir.experiments.smooth.SmoothORPredictor(clf: DecisionTreeRegressor, theta: float = 0.0)#: Bases: SmoothOperatorPredictorMixin, SmoothOR

class pybool_ir.experiments.smooth.SmoothOperator(theta: float)#: Bases: Operator, ABC

class pybool_ir.experiments.smooth.SmoothOperatorPredictorMixin(clf: BaseDecisionTree, *args, **kwargs)#: Bases: SmoothOperator, ABC