bioc_passage

autocorpus.bioc_passage

BioC Passage builder script.

Classes

BioCPassage(offset, infons, text, sentences=list(), annotations=list(), relations=list()) dataclass

Represents a BioC passage.

Functions

from_dict(passage, offset) classmethod

Create a BioCPassage from a passage dict and an offset.

Parameters:

Name	Type	Description	Default
passage	dict[str, Any]	dict containing info about passage	required
offset	int	Passage offset	required

Returns:

Type	Description
BioCPassage	BioCPassage object

Source code in autocorpus/bioc_passage.py

@classmethod
def from_dict(cls, passage: dict[str, Any], offset: int) -> BioCPassage:
    """Create a BioCPassage from a passage dict and an offset.

    Args:
        passage: dict containing info about passage
        offset: Passage offset

    Returns:
        BioCPassage object
    """
    infons = {k: v for k, v in passage.items() if k not in _DEFAULT_KEYS}

    # TODO: Doesn't account for subsubsection headings which might exist
    if heading := passage.get("section_heading", None):
        infons["section_title_1"] = heading
    if subheading := passage.get("subsection_heading", None):
        infons["section_title_2"] = subheading
    for i, section_type in enumerate(passage["section_type"]):
        infons[f"iao_name_{i + 1}"] = section_type["iao_name"]
        infons[f"iao_id_{i + 1}"] = section_type["iao_id"]

    return cls(offset, infons, passage["body"])

from_title(title, offset) classmethod

Create a BioCPassage from a title and offset.

Parameters:

Name	Type	Description	Default
title	str	Passage title	required
offset	int	Passage offset	required

Returns:

Type	Description
BioCPassage	BioCPassage object

Source code in autocorpus/bioc_passage.py

@classmethod
def from_title(cls, title: str, offset: int) -> BioCPassage:
    """Create a BioCPassage from a title and offset.

    Args:
        title: Passage title
        offset: Passage offset

    Returns:
        BioCPassage object
    """
    infons = {"iao_name_1": "document title", "iao_id_1": "IAO:0000305"}
    return cls(offset, infons, title)