medkit.io._brat_utils

medkit.io._brat_utils#

Attributes#

`GROUPING_ENTITIES`
`GROUPING_RELATIONS`
`logger`

Classes#

`BratEntity`	A simple entity annotation data structure.
`BratRelation`	A simple relation data structure.
`BratAttribute`	A simple attribute data structure.
`BratNote`	A simple note data structure.
`Grouping`	A grouping data structure for entities of type And-Group and Or-Group.
`BratAugmentedEntity`	An augmented entity data structure with its relations and attributes.
`BratDocument`
`RelationConf`	Configuration data structure of a BratRelation.
`AttributeConf`	Configuration data structure of a BratAttribure.
`BratAnnConfiguration`	A data structure to represent 'annotation.conf' in brat documents.

Functions#

`ensure_attr_value`(→ str)	Ensure that the attribue value is a string.
`parse_file`(→ BratDocument)	Read an annotation file to get the Entities, Relations and Attributes in it.
`parse_string`(→ BratDocument)	Read a string containing all annotations and extract Entities, Relations and Attributes.
`_parse_entity`(→ BratEntity)	Parse the brat entity string into an Entity structure.
`_parse_relation`(→ BratRelation)	Parse the annotation string into a Relation structure.
`_parse_attribute`(→ BratAttribute)	Parse the annotation string into an Attribute structure.
`_parse_note`(→ BratNote)	Parse the annotation string into an Note structure.

Module Contents#

medkit.io._brat_utils.GROUPING_ENTITIES#

medkit.io._brat_utils.GROUPING_RELATIONS#

medkit.io._brat_utils.logger#

class medkit.io._brat_utils.BratEntity#

A simple entity annotation data structure.

uid: str#

type: str#

span: list[tuple[int, int]]#

text: str#

property start: int#

property end: int#

to_str() → str#

class medkit.io._brat_utils.BratRelation#

A simple relation data structure.

uid: str#

type: str#

subj: str#

obj: str#

to_str() → str#

class medkit.io._brat_utils.BratAttribute#

A simple attribute data structure.

uid: str#

type: str#

target: str#

value: str = None#

to_str() → str#

class medkit.io._brat_utils.BratNote#

A simple note data structure.

uid: str#

target: str#

value: str#

type: str = 'AnnotatorNotes'#

to_str() → str#

medkit.io._brat_utils.ensure_attr_value(attr_value: Any) → str#: Ensure that the attribue value is a string.

class medkit.io._brat_utils.Grouping#

A grouping data structure for entities of type And-Group and Or-Group.

uid: str#

type: str#

items: list[BratEntity]#

property text#

class medkit.io._brat_utils.BratAugmentedEntity#

An augmented entity data structure with its relations and attributes.

uid: str#

type: str#

span: tuple[tuple[int, int], Ellipsis]#

text: str#

relations_from_me: tuple[BratRelation, Ellipsis]#

relations_to_me: tuple[BratRelation, Ellipsis]#

attributes: tuple[BratAttribute, Ellipsis]#

property start: int#

property end: int#

class medkit.io._brat_utils.BratDocument#

entities: dict[str, BratEntity]#

relations: dict[str, BratRelation]#

attributes: dict[str, BratAttribute]#

notes: dict[str, BratNote]#

groups: dict[str, Grouping] = None#

get_augmented_entities() → dict[str, BratAugmentedEntity]#

class medkit.io._brat_utils.RelationConf#

Bases: NamedTuple

Configuration data structure of a BratRelation.

type: str#

arg1: str#

arg2: str#

class medkit.io._brat_utils.AttributeConf#

Bases: NamedTuple

Configuration data structure of a BratAttribure.

from_entity: bool#

type: str#

value: str#

class medkit.io._brat_utils.BratAnnConfiguration(top_values_by_attr: int = 50)#

A data structure to represent ‘annotation.conf’ in brat documents.

This is necessary to generate a valid annotation project in brat. An ‘annotation.conf’ has four sections. The section ‘events’ is not supported in medkit, so the section is empty.

_entity_types: set[str]#

_rel_types_arg_1: dict[str, set[str]]#

_rel_types_arg_2: dict[str, set[str]]#

_attr_entity_values: dict[str, list[str]]#

_attr_relation_values: dict[str, list[str]]#

top_values_by_attr#

property entity_types: list[str]#

property rel_types_arg_1: dict[str, list[str]]#

property rel_types_arg_2: dict[str, list[str]]#

property attr_relation_values: dict[str, list[str]]#

property attr_entity_values: dict[str, list[str]]#

add_entity_type(type: str)#

add_relation_type(relation_conf: RelationConf)#

add_attribute_type(attr_conf: AttributeConf)#

to_str() → str#

static _attribute_to_str(type: str, values: list[str], from_entity: bool) → str#

static _relation_to_str(type: str, arg_1_types: list[str], arg_2_types: list[str]) → str#

medkit.io._brat_utils.parse_file(ann_path: str | pathlib.Path, detect_groups: bool = False) → BratDocument#

Read an annotation file to get the Entities, Relations and Attributes in it.

All other lines are ignored.

Parameters:

ann_pathstr or Path: The path to the annotation file to be processed.
detect_groupsbool, default=False: If set to True, the function will also parse the group of entities according to some specific keywords. By default, it is set to False.

Returns:

Document: The dataclass object containing entities, relations and attributes

medkit.io._brat_utils.parse_string(ann_string: str, detect_groups: bool = False) → BratDocument#

Read a string containing all annotations and extract Entities, Relations and Attributes.

All other lines are ignored.

Parameters:

ann_stringstr: The string containing all brat annotations
detect_groupsbool, default=False: If set to True, the function will also parse the group of entities according to some specific keywords. By default, it is set to False.

Returns:

Document: The dataclass object containing entities, relations and attributes

medkit.io._brat_utils._parse_entity(entity_id: str, entity_content: str) → BratEntity#

Parse the brat entity string into an Entity structure.

Parameters:

entity_idstr

The ID defined in the brat annotation (e.g., ‘T12’)

entity_contentstr

The string content for this ID to parse: (e.g., ‘Temporal-Modifier 116 126thistory of’)

Returns:

BratEntity: The dataclass object representing the entity

Raises:

ValueError: Raises when the entity can’t be parsed

medkit.io._brat_utils._parse_relation(relation_id: str, relation_content: str) → BratRelation#

Parse the annotation string into a Relation structure.

Parameters:

relation_idstr: The ID defined in the brat annotation (e.g., ‘R12’)
relation_contentstr: The relation text content. (e.g., ‘Modified-By Arg1:T8 Arg2:T6t’)

Returns:

BratRelation: The dataclass object representing the relation

Raises:

ValueError: Raises when the relation can’t be parsed

medkit.io._brat_utils._parse_attribute(attribute_id: str, attribute_content: str) → BratAttribute#

Parse the annotation string into an Attribute structure.

Parameters:

attribute_idstr: The attribute ID defined in the annotation. (e.g., ‘A1’)
attribute_contentstr: The attribute text content. (e.g., ‘Tense T19 Past-Ended’)

Returns:

BratAttribute: The dataclass object representing the attribute

Raises:

ValueError: Raises when the attribute can’t be parsed

medkit.io._brat_utils._parse_note(note_id: str, note_content: str) → BratNote#

Parse the annotation string into an Note structure.

Parameters:

note_idstr: The note ID defined in the annotation. (e.g., ‘#1’)
note_contentstr: The note text content. (e.g., ‘AnnotatorNotes T10 C0011849’)

Returns:

BratNote: The dataclass object representing the note

Raises:

ValueError: Raises when the note can’t be parsed