clinlp.ie.qualifier¶

Functionality for extracting qualifiers from clinical text.

clinlp.ie.qualifier.context_algorithm¶

The Context Algorithm implemented as a spaCy component.

class clinlp.ie.qualifier.context_algorithm.ContextRuleDirection(*values)¶

Bases: Enum

Direction of a rule, as in the original Context Algorithm.

PRECEDING = 1¶: For triggers that precede entities.

FOLLOWING = 2¶: For triggers that follow entities.

BIDIRECTIONAL = 3¶: For triggers that can both precede and follow entities.

PSEUDO = 4¶: For pseudo triggers.

TERMINATION = 5¶: For termination triggers.

class clinlp.ie.qualifier.context_algorithm.ContextRule(pattern: str | list[dict[str, str]], direction: ContextRuleDirection, qualifier: Qualifier, max_scope: int | None = None)¶

Bases: object

A Context rule, as in the original Context Algorithm.

pattern: str | list[dict[str, str]]¶: The pattern to look for in text. Either a string, or a spaCy pattern (list).

direction: ContextRuleDirection¶: The Context rule direction.

qualifier: Qualifier¶: The qualifier to apply.

max_scope: int | None = None¶: The maximum number of tokens a trigger ranges over, or None for using sentence boundaries.

class clinlp.ie.qualifier.context_algorithm.ContextAlgorithm(nlp: Language, *, phrase_matcher_attr: str = 'TEXT', load_rules: bool = True, rules: str | dict | None = '/home/docs/checkouts/readthedocs.org/user_builds/clinlp/checkouts/latest/src/clinlp/resources/context_rules.json', **kwargs)¶

Bases: QualifierDetector

spaCy pipeline component that implements the Context Algorithm.

For more information, see the original paper: https://doi.org/10.1016%2Fj.jbi.2009.05.002

__init__(nlp: Language, *, phrase_matcher_attr: str = 'TEXT', load_rules: bool = True, rules: str | dict | None = '/home/docs/checkouts/readthedocs.org/user_builds/clinlp/checkouts/latest/src/clinlp/resources/context_rules.json', **kwargs) → None¶

Initialize the Context Algorithm.

Parameters:

nlp – The spaCy language model.
phrase_matcher_attr – The token attribute to match phrases on (e.g. TEXT, ORTH, NORM).
load_rules – Whether to parse any rules. Set this to True to load the default builtin rules. Set this to False to use ContextAlgorithm.add_rules to add ContextRules manually.
rules – A dictionary of rules, or a path to a json containing the rules. See the clinlp.resources dir for an example.

Raises:

ValueError – If no rules are provided and load_rules is set to True.

property qualifier_classes: dict[str, QualifierClass]¶

Obtain the qualifier classes that a QualifierDetector initializes.

These are used to initialize the default qualifiers for each entity.

Returns:: dict[str, QualifierClass] – The qualifier classes.

add_rule(rule: ContextRule) → None¶

Add a rule to the Context Algorithm.

Parameters:: rule – The rule to add.
Raises:: TypeError – If the rule pattern is not a string or a list.

add_rules(rules: list[ContextRule]) → None¶

Add multiple rules to the Context Algorithm.

Parameters:: rules – The rules to add.

clinlp.ie.qualifier.qualifier¶

Reusable components for detecting qualifiers in clinical text.

clinlp.ie.qualifier.qualifier.qualifiers_to_str(ent: Span) → set[str] | None¶

Get qualifier information in string format.

Parameters:: ent – The entity to get qualifiers from.
Returns:: Optional[set[str]] – The qualifiers in string format, e.g. {'Presence.Present', ...}, or None if no qualifiers are present.

clinlp.ie.qualifier.qualifier.qualifiers_to_dict(ent: Span) → list[dict] | None¶

Get qualifier information in dictionary format.

Parameters:: ent – The entity to get qualifiers from.
Returns:: Optional[list[dict]] – The qualifiers in dict format, e.g. [{'Name': 'Presence', 'Value': 'Present', 'is_default': True}, ...], or None if no qualifiers are present.

clinlp.ie.qualifier.qualifier.get_qualifiers(entity: Span) → set[Qualifier]¶

Get the qualifiers for an entity.

Returns:: set[Qualifier] – The qualifiers.

clinlp.ie.qualifier.qualifier.set_qualifiers(entity: Span, qualifiers: set[Qualifier]) → None¶

Set the qualifiers for an entity.

Parameters:

entity – The entity to set qualifiers for.
qualifiers – The qualifiers to set.

class clinlp.ie.qualifier.qualifier.Qualifier(name: str, value: str, is_default: bool | None = None, priority: int = 0, prob: float | None = None)¶

Bases: object

A qualifier for an entity.

A qualifier is a piece of information that provides additional context to an entity. For example, a Presence qualifier with a value of Present or Absent. A qualifier has a fixed value.

name: str¶: The name of the qualifier.

value: str¶: The value of the qualifier.

is_default: bool | None = None¶: Whether the value is the default value.

priority: int = 0¶: The priority of the qualifier.

prob: float | None = None¶: The probability of the qualifier.

to_dict() → dict¶

Convert the qualifier to a dictionary.

Returns:: dict – The qualifier as a dictionary.

class clinlp.ie.qualifier.qualifier.QualifierClass(name: str, values: list[str], default: str | None = None, priorities: dict | None = None)¶

Bases: object

A qualifier class.

A qualifier class defines the set of possible values a qualifier can take on. For example: Presence with values Present and Absent. The qualifier class creates qualifiers, although they can also be created directly.

__init__(name: str, values: list[str], default: str | None = None, priorities: dict | None = None) → None¶

Initialize a qualifier class.

Parameters:

name – The name of the qualifier.
values – The possible values of the qualifier.
default – The default value of the qualifier.
priorities – The priorities of the values. If not provided, the order of the values is used.

Raises:

ValueError – If there are duplicate values.
ValueError – If the default value is not in the provided values.

create(value: str | None = None, **kwargs) → Qualifier¶

Create a qualifier in this qualifier class.

Parameters:: value – The value for the qualifier.
Returns:: Qualifier – The created qualifier.
Raises:: ValueError – If the value is not in the possible values.

class clinlp.ie.qualifier.qualifier.QualifierDetector(spans_key: str = 'ents')¶

Bases: Pipe

Abstract pipeline component for detecting qualifiers in clinical text.

__init__(spans_key: str = 'ents') → None¶

Initialize a qualifier detector.

Parameters:: spans_key – The key for the spans in the Doc object.

abstract property qualifier_classes: dict[str, QualifierClass]¶

Obtain the qualifier classes that a QualifierDetector initializes.

These are used to initialize the default qualifiers for each entity.

Returns:: dict[str, QualifierClass] – The qualifier classes.

static add_qualifier_to_ent(entity: Span, new_qualifier: Qualifier) → None¶

Add a qualifier to an entity.

Preferably, qualifiers should not be added in another way than through this method, to ensure consistency.

Parameters:

entity – The entity to add the qualifier to.
new_qualifier – The qualifier to add.

Raises:

RuntimeError – If the entity does not have initialized qualifiers.

__call__(doc: Doc) → Doc¶

Initialize default qualifiers and run detection.

Parameters:: doc – The document to process.
Returns:: Doc – The processed document.

clinlp.ie.qualifier¶

clinlp.ie.qualifier.context_algorithm¶

clinlp.ie.qualifier.qualifier¶

clinlp.ie.qualifier.transformer¶