"""This package needs extra-dependencies not installed as core dependencies of medkit.
To install them, use `pip install medkit[edsnlp]`.
"""
from __future__ import annotations
__all__ = ["TNMAttribute"]
import dataclasses
from typing import Any, ClassVar
from edsnlp.pipelines.ner.tnm.model import (
TNM,
Metastasis,
Node,
Prefix,
Specification,
Tumour,
)
from typing_extensions import Self
from medkit.core import Attribute, dict_conv
[docs]
@dataclasses.dataclass
class TNMAttribute(Attribute):
"""Attribute destructuring the fields of a TNM string.
The TNM (Tumour/Node/Metastasis) system is used to describe cancer stages.
This class is replicating EDS-NLP's `TDM` class, making it a medkit
:class:`Attribute`.
Attributes
----------
uid:
Identifier of the attribute
label:
The attribute label, always set to :attr:`TNMAttribute.LABEL`
value:
Normalized string representation of the TNM (ex: "pTxN1M1")
tumour:
Tumour score
tumour_specification:
Tumour specification
tumour_suffix:
Tumour suffix
node:
Node score
node_specification:
Node specification
node_suffix:
Node suffix
metastasis:
Metastasis score
resection_completeness:
Resection completeness (R factor)
version:
Version (ex: "uicc", "accj")
version_year:
Version year
"""
prefix: Prefix | None
tumour: Tumour | None
tumour_specification: Specification | None
tumour_suffix: str | None
node: Node | None
node_specification: Specification | None
node_suffix: str | None
metastasis: Metastasis | None
resection_completeness: int | None
version: str | None
version_year: int | None
LABEL: ClassVar[str] = "TNM"
"""
Label used for all TNM attributes
"""
def __init__(
self,
prefix: Prefix | None = None,
tumour: Tumour | None = None,
tumour_specification: Specification | None = None,
tumour_suffix: str | None = None,
node: Node | None = None,
node_specification: Specification | None = None,
node_suffix: str | None = None,
metastasis: Metastasis | None = None,
resection_completeness: int | None = None,
version: str | None = None,
version_year: int | None = None,
metadata: dict[str, Any] | None = None,
uid: str | None = None,
):
# use EDS-NLP's TNM class to build string representation
value = TNM(
prefix=prefix,
tumour=tumour,
tumour_specification=tumour_specification,
tumour_suffix=tumour_suffix,
node=node,
node_specification=node_specification,
node_suffix=node_suffix,
metastasis=metastasis,
resection_completeness=resection_completeness,
version=version,
version_year=version_year,
).norm()
super().__init__(label=self.LABEL, value=value, metadata=metadata, uid=uid)
self.prefix = prefix
self.tumour = tumour
self.tumour_specification = tumour_specification
self.tumour_suffix = tumour_suffix
self.node = node
self.node_specification = node_specification
self.node_suffix = node_suffix
self.metastasis = metastasis
self.resection_completeness = resection_completeness
self.version = version
self.version_year = version_year
[docs]
def to_brat(self) -> str:
return self.value
[docs]
def to_spacy(self) -> str:
return self.value
def to_dict(self) -> dict[str, Any]:
tnm_dict = {
"uid": self.uid,
"prefix": self.prefix,
"tumour": self.tumour,
"tumour_suffix": self.tumour_suffix,
"tumour_specification": self.tumour_specification,
"node": self.node,
"node_specification": self.node_specification,
"node_suffix": self.node_suffix,
"metastasis": self.metastasis,
"resection_completeness": self.resection_completeness,
"version": self.version,
"version_year": self.version_year,
"metadata": self.metadata,
}
dict_conv.add_class_name_to_data_dict(self, tnm_dict)
return tnm_dict
[docs]
@classmethod
def from_dict(cls, tnm_dict: dict[str, Any]) -> Self:
return cls(
uid=tnm_dict["uid"],
prefix=tnm_dict["prefix"],
tumour=tnm_dict["tumour"],
tumour_suffix=tnm_dict["tumour_suffix"],
tumour_specification=tnm_dict["tumour_specification"],
node=tnm_dict["node"],
node_specification=tnm_dict["node_specification"],
node_suffix=tnm_dict["node_suffix"],
metastasis=tnm_dict["metastasis"],
resection_completeness=tnm_dict["resection_completeness"],
version=tnm_dict["version"],
version_year=tnm_dict["version_year"],
metadata=tnm_dict["metadata"],
)
