`cobra_db.model`

Module Contents

Classes

`BaseObj`	Defines the interface to load the object from dict and how to export it to dict.
`Embedded`	Base class for an embedded doc (that doesn't get its own _id)
`Metadata`	Stores information of when the db doc was created, modified, and the data model
`Entity`	Base class for all entities that are stored in the db with an _id, i.e. that
`Source`	Base class for a pointer to more information.
`DicomEntity`	Base class for all entities that contain dicom_tags.
`EntitySource`	Pointer to another Entity instance in the database.
`FileSource`	Pointer to a file in the filesystem.
`Patient`	Represents a unique person that has gone through at least one study.
`RadiologicalStudy`	Aggregation of multiple dicom instances by the combination of patient and date.
`RadiologicalSeries`	Base class for all entities that contain dicom_tags.
`ImageMetadata`	Represents one DICOM image i.e. a DICOM instance with PixelData.
`Annotation`	Abstract class for any annotation that is stored in cobra_db. Should be subclassed
`Annotator`	Base class for all entities that are stored in the db with an _id, i.e. that

class cobra_db.model.BaseObj[source]

Defines the interface to load the object from dict and how to export it to dict.

static expand(value)[source]

_to_dict()[source]

to_dict()[source]: expand the nested objects and remove all keys with None value

abstract classmethod from_dict(d: dict)[source]

class cobra_db.model.Embedded[source]

Bases: BaseObj

Base class for an embedded doc (that doesn’t get its own _id)

classmethod from_dict(d: dict)[source]

class cobra_db.model.Metadata[source]

Bases: Embedded

Stores information of when the db doc was created, modified, and the data model version. Should be used in any doc with _id that is stored in the database.

model_version: str

created: datetime.datetime

modified: datetime.datetime

project_name: str

classmethod create()[source]

class cobra_db.model.Entity[source]

Bases: BaseObj

Base class for all entities that are stored in the db with an _id, i.e. that require a Dao. Usually parent of a @dataclass class

_id: bson.ObjectId

_metadata: Metadata

__repr__()[source]: Return repr(self).

__post_init__()[source]: Automatically add db_metadata if missing

classmethod from_dict(obj_dict: dict)[source]

class cobra_db.model.Source[source]: Base class for a pointer to more information.

class cobra_db.model.DicomEntity[source]

Bases: Entity

Base class for all entities that contain dicom_tags.

dicom_tags: dict

static optional(ds: pydicom.dataset.Dataset, tag_name: str, parser: Callable = None, default=None)[source]: Get tag value from dataset, if missing return default value.

static choose_first(ds, *tag_names, parser, default)[source]: Choose first tag that does not return None,

abstract classmethod from_dataset(ds: pydicom.dataset.Dataset)[source]: Abstract method. How to create an instance from a pydicom Dataset

get_tag(keyword: str, default=None)[source]

get_acquisition_datetime()[source]

class cobra_db.model.EntitySource[source]

Bases: Embedded, Source

Pointer to another Entity instance in the database. Should only be used when there’s ambiguity about the referenced class/collection.

name: str

description: str

_entity_type: Union[Type[Entity], str]

_id: bson.ObjectId

__post_init__()[source]

to_dict()[source]: expand the nested objects and remove all keys with None value

class cobra_db.model.FileSource[source]

Bases: Embedded, Source

Pointer to a file in the filesystem. rel_path is relative to the place where drive_name is mounted. This is because drive_name can be mounted in different paths according to the machine and we don’t have another way of referencing a drive. filename is automatically stored for easy indexing and querying without having to manipulate the rel_path to extract the filename.

drive_name: str

rel_path: str

filename: str

get_local_filepath(mount_paths: dict) → str[source]: Returns the filepath according to mount_paths

__post_init__()[source]

classmethod from_mount_paths(filepath: str, mount_paths: dict)[source]

create an instance from the filepath and the mount_paths.

Parameters:

filepath – _description_
mount_paths – _description_

class cobra_db.model.Patient[source]

Bases: Entity

Represents a unique person that has gone through at least one study. https://dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C.2.2.html https://dicom.nema.org/medical/dicom/current/output/chtml/part03/sect_C.2.3.html

anon_id: str

hidden: bool

birth_date: datetime.datetime

__repr__()[source]: Return repr(self).

classmethod from_dataset(ds: pydicom.dataset.Dataset)[source]: Create an instance of the entity from a DICOM dataset

__eq__(other: Patient) → bool[source]: Return self==value.

class cobra_db.model.RadiologicalStudy[source]

Bases: DicomEntity

Aggregation of multiple dicom instances by the combination of patient and date.

study_uid: List[str]

date: datetime.datetime

series_count: int

patient_id: bson.ObjectId

sop_class: List[str]

modality: List[cobra_db.enums.Modality]

description: str

accession_number: List[str]

patient_age: int

patient_weight: float

patient_height: float

patient_pregnancy_status: cobra_db.enums.PregnancyStatus

manufacturer: List[str]

manufacturer_model_name: List[str]

detector_id: List[str]

detector_type: List[str]

device_serial_number: List[str]

software_versions: List[str]

date_of_last_detector_calibration: List[str]

breast_implant_present: List[str]

__repr__()[source]: Return repr(self).

classmethod from_dataset(ds: pydicom.dataset.Dataset)[source]: Create an instance of the entity from a DICOM dataset

classmethod from_dict(d: dict)[source]: Convert a mongodb dict into a RadiologicalStudy object

__eq__(other: RadiologicalStudy) → bool[source]: Return self==value.

class cobra_db.model.RadiologicalSeries[source]

Bases: DicomEntity

Base class for all entities that contain dicom_tags.

study_id: bson.ObjectId

date: datetime.datetime

series_uid: str

description: str

image_count: int

image_shape: Tuple[int, int]

protocol_name: str

__repr__()[source]: Return repr(self).

classmethod from_dataset(ds: pydicom.dataset.Dataset)[source]: Create an instance of the entity from a DICOM dataset

classmethod from_images_metadata(images_metadata: List[ImageMetadata])[source]

__eq__(other: RadiologicalSeries) → bool[source]: Return self==value.

class cobra_db.model.ImageMetadata[source]

Bases: DicomEntity

Represents one DICOM image i.e. a DICOM instance with PixelData. Only the dicom headers are stored. While the pixel data is not stored.

dicom_tags: dict

file_source: FileSource

study_id: bson.ObjectId

series_id: bson.ObjectId

aka_file_sources: List[FileSource]

classmethod from_dataset(ds: pydicom.dataset.Dataset, mount_paths: dict)[source]

Create one instance from a pydicom Dataset.

Parameters:

ds – pydicom Dataset
mount_paths – dict that maps the drive_names to the mount path of the drives that are in use.

to_dataset() → pydicom.dataset.Dataset[source]: Creates a pydicom Dataset with only the metadata in dicom_tags. Useful to use pydicom interface when required.

get_checksum(mount_paths: dict) → str[source]

Obtain the md5 hexdigest hash of the original file. This helps finding duplicated images. Warning: md5 is not cryptographycally secure.

Parameters:: mount_paths – a dictionary mapping the drive_name to the mount path in the local computer.
Returns:: string hexdigest with the hash of the file.

get_local_filepath(mount_paths: dict) → str[source]

Returns the filepath of the file_source according to mount_paths.

Parameters:: mount_paths – a dictionary mapping the drive_name to the mount path in the local computer.
Returns:: string hexdigest with the hash of the file.

get_pixel_data(mount_paths) → numpy.ndarray[source]

Load the file and extract the pixel_array from the pydicom dataset.

Parameters:: mount_paths – a dictionary mapping the drive_name to the mount path in the local computer.
Returns:: the content of pydicom.Dataset.pixel_array

get_file_size(mount_paths) → int[source]: Get the size in bytes of the file

__repr__()[source]: Return repr(self).

class cobra_db.model.Annotation[source]

Bases: Entity

Abstract class for any annotation that is stored in cobra_db. Should be subclassed

annotation_type: List[Literal[text, semantic_segmentation, binary, categorical]]

annotator_email: str

date: datetime.datetime

possible_labels: List[str]

file_format: str

aggregation_level: Literal[patient, study, series, image]

referenced_entity_uid: str

patient_anon_id: str

study_date: datetime.datetime

file_source: FileSource

project_name: str

class cobra_db.model.Annotator[source]

Bases: Entity

Base class for all entities that are stored in the db with an _id, i.e. that require a Dao. Usually parent of a @dataclass class

email: str

specialization: str

first_name: str

last_name: str

cobra_db.model

Module Contents

Classes

`cobra_db.model`