sparknlp.annotator.matcher.rule_based_matcher#

Classes#

class RuleBasedMatcher[source]#

Rule-based token matcher over multiple Spark NLP annotation columns.

Rules are supplied as JSON or JSONL. Each token pattern can combine attributes from token text, POS, lemmas, NER, dependency metadata, or custom mapped annotation columns. Matches are emitted as CHUNK annotations.

Input Annotation types	Output Annotation type
DOCUMENT, TOKEN, ...	CHUNK

inputAnnotatorTypes[source]#

outputAnnotatorType = 'chunk'[source]#

rules[source]#

rulesResource[source]#

setRules(value)[source]#

Sets inline JSON/JSONL rule definitions.

Parameters:

valuestr, dict, or list

JSON/JSONL string, one rule dict, or a list of rule dicts.

setRulesResource(path, read_as=ReadAs.TEXT, options={'format': 'text'})[source]#

Sets an external JSON or JSONL rules resource.

class RuleBasedMatcherModel(classname='com.johnsnowlabs.nlp.annotators.matcher.RuleBasedMatcherModel', java_model=None)[source]#

Instantiated model of RuleBasedMatcher.

name = 'RuleBasedMatcherModel'[source]#

inputAnnotatorTypes[source]#

outputAnnotatorType = 'chunk'[source]#

inputColumnTypes[source]#

rulesJson[source]#