lexnlp.extract.en.constraints: Extracting constraint statements

The lexnlp.extract.en.constraints module contains methods that allow for the extraction of constraint statements from text. Statements that are covered by default in this module are:

• after

• at least

• at most

• before

• equal to

• exactly

• first of

• greater

• greater of

• greater than

• greater than or equal to

• greatest of

• last of

• least of

• lesser

• lesser of

• lesser than

• less than

• less than or equal to

• maximum of

• maximum

• minimum of

• minimum

• more than

• more than or equal to

• no earlier than

• no later than

• no less than

• no more than

• not equal to

• not to exceed

• earlier than

• later than

• within

• exceed

• exceeds

• prior to

• highest

• least

• smallest among

The full list of current unit test cases can be found here: https://github.com/LexPredict/lexpredict-lexnlp/tree/master/test_data/lexnlp/extract/en/tests/test_constraints

Extracting constraints

Example

>>> import lexnlp.extract.en.constraints
>>> text = "This will occur at most three times."
>>> print(list(lexnlp.extract.en.constraints.get_constraints(text)))
[('at most', 'this will occur', '')]

>>> import lexnlp.extract.en.conditions
>>> text = "The rate shall be no less than 50 bps."
>>> print(list(lexnlp.extract.en.constraints.get_constraints(text)))
[('no less than', 'the rate shall be', '')]

Customizing constraint statement extraction

Constraint statement extraction can be customized. There are two key module variables that store the default configuration and one function used to create a matching instance:

• CONSTRAINT_PHRASES: This List stores the “trigger” phrases that are used to identify constraint statements. They are typically adverbial or prepositional phrases.

• CONSTRAINT_PATTERN_TEMPLATE: This String stores the regular expression pattern that drives matching in this module.

Note

For more examples and information about natural language constraints, see the linguistic resources below:

• https://www.ijcai.org/Proceedings/16/Papers/111.pdf

• https://www.sciencedirect.com/science/article/pii/S131915781200002X

The default behavior of this module can be customized by overriding the value of RE_CONSTRAINT with a new regular expression created using create_constraint_pattern above. The example below demonstrates a simple addition of a new phrase:

>>> # Out of the box behavior
>>> import lexnlp.extract.en.constraints
>>> text = "The applicable rate shall be the smallest among thing A and thing B at time T."
>>> print(list(lexnlp.extract.en.constraints.get_constraints(text)))
[]

>>> # Customize the `RE_CONSTRAINT` variable by adding a new phrase
>>> import regex as re
>>> my_constraint_phrases = lexnlp.extract.en.constraints.CONSTRAINT_PHRASES
>>> my_constraint_phrases.append("smallest among")
>>> CONSTRAINT_PATTERN = lexnlp.extract.en.constraints.create_constraint_pattern(lexnlp.extract.en.constraints.CONSTRAINT_PATTERN_TEMPLATE, my_constraint_phrases)
>>> lexnlp.extract.en.constraints.RE_CONSTRAINT = re.compile(CONSTRAINT_PATTERN, re.IGNORECASE | re.UNICODE | re.DOTALL | re.MULTILINE | re.VERBOSE)

>>> # Run the `get_constraints` method again to test
>>> print(list(lexnlp.extract.en.constraints.get_constraints(text)))
[('smallest among', 'the applicable rate shall be the', '')]