In-database BIDS file information

patch-bids-tables.sql

@@ -0,0 +1,52 @@
+CREATE TABLE `bids_dataset` (
+  `ID` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
+  `Path` VARCHAR(255) NOT NULL,
+  `InsertTime` DATETIME NOT NULL,
+  PRIMARY KEY (`ID`),
+  UNIQUE KEY `bids_dataset_path_unique` (`Path`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;
+
+CREATE TABLE `bids_file` (
+  `ID` INT(10) UNSIGNED NOT NULL AUTO_INCREMENT,
+  `DatasetID` INT(10) UNSIGNED NOT NULL,
+  `Path` VARCHAR(255) NOT NULL,
+  `SourcePath` VARCHAR(255) NULL,
+  `InsertTime` DATETIME NOT NULL,
+  `Blake2bHash` CHAR(128) NOT NULL,
+  `Derivative` TINYINT(1) NOT NULL,
+  PRIMARY KEY (`ID`),
+  UNIQUE KEY `bids_file_path_unique` (`Path`),
+  KEY `bids_file_dataset_id_fk_idx` (`DatasetID`),
+  CONSTRAINT `bids_file_dataset_id_fk`
+    FOREIGN KEY (`DatasetID`) REFERENCES `bids_dataset` (`ID`) ON DELETE CASCADE
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;
+
+ALTER TABLE `files`
+  ADD COLUMN `BidsInfoID` INT(10) UNSIGNED NULL,
+  ADD KEY `files_bids_info_id_fk_idx` (`BidsInfoID`),
+  ADD CONSTRAINT `files_bids_info_id_fk`
+    FOREIGN KEY (`BidsInfoID`) REFERENCES `bids_file` (`ID`) ON DELETE SET NULL;
+
+ALTER TABLE `physiological_file`
+  ADD COLUMN `BidsInfoID` INT(10) UNSIGNED NULL,
+  ADD KEY `physiological_file_bids_info_id_fk_idx` (`BidsInfoID`),
+  ADD CONSTRAINT `physiological_file_bids_info_id_fk`
+    FOREIGN KEY (`BidsInfoID`) REFERENCES `bids_file` (`ID`) ON DELETE SET NULL;
+
+ALTER TABLE `physiological_event_file`
+  ADD COLUMN `BidsInfoID` INT(10) UNSIGNED NULL,
+  ADD KEY `physiological_event_file_bids_info_id_fk_idx` (`BidsInfoID`),
+  ADD CONSTRAINT `physiological_event_file_bids_info_id_fk`
+    FOREIGN KEY (`BidsInfoID`) REFERENCES `bids_file` (`ID`) ON DELETE SET NULL;
+
+ALTER TABLE `meg_ctf_head_shape_file`
+  ADD COLUMN `BidsInfoID` INT(10) UNSIGNED NULL,
+  ADD COLUMN `InsertTime` DATETIME NOT NULL,
+  ADD KEY `meg_ctf_head_shape_file_bids_info_id_fk_idx` (`BidsInfoID`),
+  ADD CONSTRAINT `meg_ctf_head_shape_file_bids_info_id_fk`
+    FOREIGN KEY (`BidsInfoID`) REFERENCES `bids_file` (`ID`) ON DELETE SET NULL;
+
+ALTER TABLE `physiological_file`
+  DROP FOREIGN KEY `physiological_file_head_shape_file_id_fk`,
+  ADD CONSTRAINT `physiological_file_head_shape_file_id_fk`
+    FOREIGN KEY (`HeadShapeFileID`) REFERENCES `meg_ctf_head_shape_file` (`ID`) ON DELETE SET NULL;

python/lib/db/models/bids_dataset.py

@@ -0,0 +1,30 @@
+from datetime import datetime
+from pathlib import Path
+
+from sqlalchemy.orm import Mapped, mapped_column
+
+from lib.db.base import Base
+from lib.db.decorators.string_path import StringPath
+
+
+class DbBidsDataset(Base):
+    """
+    A LORIS BIDS dataset.
+    """
+
+    __tablename__ = 'bids_dataset'
+
+    id: Mapped[int] = mapped_column('ID', primary_key=True, autoincrement=True)
+    """
+    The ID of this BIDS dataset.
+    """
+
+    path: Mapped[Path] = mapped_column('Path', StringPath, unique=True)
+    """
+    The path of this BIDS dataset, relative to the LORIS data directory.
+    """
+
+    insert_time: Mapped[datetime] = mapped_column('InsertTime', default=datetime.now)
+    """
+    The time at which this BIDS dataset was created in LORIS.
+    """

python/lib/db/models/bids_file.py

@@ -0,0 +1,58 @@
+from datetime import datetime
+from pathlib import Path
+
+from sqlalchemy import ForeignKey
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+import lib.db.models.bids_dataset as db_bids_dataset
+from lib.db.base import Base
+from lib.db.decorators.int_bool import IntBool
+from lib.db.decorators.string_path import StringPath
+
+
+class DbBidsFile(Base):
+    """
+    A file within a LORIS BIDS dataset.
+    """
+
+    __tablename__ = 'bids_file'
+
+    id: Mapped[int] = mapped_column('ID', primary_key=True, autoincrement=True)
+    """
+    The ID of this BIDS file.
+    """
+
+    dataset_id: Mapped[int] = mapped_column('DatasetID', ForeignKey('bids_dataset.ID', ondelete='CASCADE'))
+    """
+    The ID of the BIDS dataset to which this file belongs.
+    """
+
+    path: Mapped[Path] = mapped_column('Path', StringPath, unique=True)
+    """
+    The path of this file relative to its LORIS BIDS dataset.
+    """
+
+    source_path: Mapped[Path | None] = mapped_column('SourcePath', StringPath)
+    """
+    The source path of this file relative to the BIDS dataset from which it was imported.
+    """
+
+    insert_time: Mapped[datetime] = mapped_column('InsertTime', default=datetime.now)
+    """
+    The time at which this BIDS dataset was created in LORIS.
+    """
+
+    blake2b_hash: Mapped[str] = mapped_column('Blake2bHash')
+    """
+    The BLAKE2b hash of this file.
+    """
+
+    derivative: Mapped[bool] = mapped_column('Derivative', IntBool)
+    """
+    Whether this file is a BIDS derivative.
+    """
+
+    dataset: Mapped['db_bids_dataset.DbBidsDataset'] = relationship('DbBidsDataset')
+    """
+    The BIDS dataset to which this file belongs.
+    """

python/lib/db/models/file.py

@@ -4,6 +4,7 @@
 from sqlalchemy import ForeignKey
 from sqlalchemy.orm import Mapped, mapped_column, relationship
 
+import lib.db.models.bids_file as db_bids_file
 import lib.db.models.dicom_archive as db_dicom_archive
 import lib.db.models.file_parameter as db_file_parameter
 import lib.db.models.session as db_session
@@ -40,6 +41,16 @@ class DbFile(Base):
     acquisition_order_per_modality : Mapped[int | None]   = mapped_column('AcqOrderPerModality')
     acquisition_date               : Mapped[date | None]  = mapped_column('AcquisitionDate')
 
+    bids_info_id: Mapped[int | None] = mapped_column('BidsInfoID', ForeignKey('bids_file.ID', ondelete='SET NULL'))
+    """
+    The ID of the BIDS information of this file, if any.
+    """
+
+    bids_info: Mapped['db_bids_file.DbBidsFile | None'] = relationship('DbBidsFile')
+    """
+    The BIDS information of this file, if any.
+    """
+
     session: Mapped['db_session.DbSession'] = relationship('DbSession', back_populates='files')
     """
     The session to which this file belongs.

python/lib/db/models/meg_ctf_head_shape_file.py

@@ -1,7 +1,10 @@
+from datetime import datetime
 from pathlib import Path
 
+from sqlalchemy import ForeignKey
 from sqlalchemy.orm import Mapped, mapped_column, relationship
 
+import lib.db.models.bids_file as db_bids_file
 import lib.db.models.meg_ctf_head_shape_point as db_meg_ctf_head_shape_point
 from lib.db.base import Base
 from lib.db.decorators.string_path import StringPath
@@ -20,11 +23,21 @@ class DbMegCtfHeadShapeFile(Base):
     ID of the head shape file.
     """
 
+    bids_info_id: Mapped[int | None] = mapped_column('BidsInfoID', ForeignKey('bids_file.ID', ondelete='SET NULL'))
+    """
+    The ID of the BIDS information of this head shape file, if any.
+    """
+
     path: Mapped[Path] = mapped_column('Path', StringPath)
     """
     Path of the head shape file relative to the LORIS data directory.
     """
 
+    insert_time: Mapped[datetime] = mapped_column('InsertTime', default=datetime.now)
+    """
+    The time at which this head shape file was created in LORIS.
+    """
+
     blake2b_hash: Mapped[str] = mapped_column('Blake2bHash')
     """
     Blake2B hash of the head shape file, which may be used to check that the on-disk file data
@@ -35,3 +48,8 @@ class DbMegCtfHeadShapeFile(Base):
     """
     3D points present in the head shape file.
     """
+
+    bids_info: Mapped['db_bids_file.DbBidsFile | None'] = relationship('DbBidsFile')
+    """
+    The BIDS information of this head shape file, if any.
+    """

python/lib/db/models/physio_event_file.py

@@ -4,6 +4,7 @@
 from sqlalchemy import ForeignKey
 from sqlalchemy.orm import Mapped, mapped_column, relationship
 
+import lib.db.models.bids_file as db_bids_file
 import lib.db.models.imaging_file_type as db_imaging_file_type
 import lib.db.models.physio_event_parameter as db_physio_event_parameter
 import lib.db.models.physio_file as db_physio_file
@@ -24,8 +25,18 @@ class DbPhysioEventFile(Base):
     last_update    : Mapped[datetime]    = mapped_column('LastUpdate', default=datetime.now)
     last_written   : Mapped[datetime]    = mapped_column('LastWritten', default=datetime.now)
 
+    bids_info_id: Mapped[int | None] = mapped_column('BidsInfoID', ForeignKey('bids_file.ID', ondelete='SET NULL'))
+    """
+    The ID of the BIDS information of this event file, if any.
+    """
+
     physio_file       : Mapped['db_physio_file.DbPhysioFile | None']                     = relationship('DbPhysioFile')
     project           : Mapped['db_project.DbProject | None']                            = relationship('DbProject')
     imaging_file_type : Mapped['db_imaging_file_type.DbImagingFileType | None']          = relationship('DbImagingFileType')
     task_events       : Mapped[list['db_physio_task_event.DbPhysioTaskEvent']]           = relationship('DbPhysioTaskEvent', back_populates='event_file')
     event_parameters  : Mapped[list['db_physio_event_parameter.DbPhysioEventParameter']] = relationship('DbPhysioEventParameter', back_populates='event_file')
+
+    bids_info: Mapped['db_bids_file.DbBidsFile | None'] = relationship('DbBidsFile')
+    """
+    The BIDS information of this event file, if any.
+    """

python/lib/db/models/physio_file.py

@@ -4,6 +4,7 @@
 from sqlalchemy import ForeignKey
 from sqlalchemy.orm import Mapped, mapped_column, relationship
 
+import lib.db.models.bids_file as db_bids_file
 import lib.db.models.meg_ctf_head_shape_file as db_meg_ctf_head_shape_file
 import lib.db.models.physio_channel as db_physio_channel
 import lib.db.models.physio_event_archive as db_physio_event_archive
@@ -33,7 +34,12 @@ class DbPhysioFile(Base):
     index            : Mapped[int | None]      = mapped_column('Index')
     parent_id        : Mapped[int | None]      = mapped_column('ParentID')
 
-    head_shape_file_id: Mapped[int | None] = mapped_column('HeadShapeFileID', ForeignKey('meg_ctf_head_shape_file.ID'))
+    bids_info_id: Mapped[int | None] = mapped_column('BidsInfoID', ForeignKey('bids_file.ID', ondelete='SET NULL'))
+    """
+    The ID of the BIDS information of this file, if any.
+    """
+
+    head_shape_file_id: Mapped[int | None] = mapped_column('HeadShapeFileID', ForeignKey('meg_ctf_head_shape_file.ID', ondelete='SET NULL'))
     """
     ID of the head shape file associated to this file, which is only present for MEG CTF files.
     """
@@ -48,6 +54,11 @@ class DbPhysioFile(Base):
     event_files   : Mapped[list['db_physio_event_file.DbPhysioEventFile']]         = relationship('DbPhysioEventFile', back_populates='physio_file')
     task_events   : Mapped[list['db_physio_task_event.DbPhysioTaskEvent']]         = relationship('DbPhysioTaskEvent', back_populates='physio_file')
 
+    bids_info: Mapped['db_bids_file.DbBidsFile | None'] = relationship('DbBidsFile')
+    """
+    The BIDS information of this file, if any.
+    """
+
     head_shape_file: Mapped['db_meg_ctf_head_shape_file.DbMegCtfHeadShapeFile | None'] = relationship('DbMegCtfHeadShapeFile')
     """
     The head shape file associated to this file, which is only present for MEG CTF files.

python/lib/db/queries/bids_dataset.py

@@ -0,0 +1,16 @@
+from pathlib import Path
+
+from sqlalchemy import select
+from sqlalchemy.orm import Session as Database
+
+from lib.db.models.bids_dataset import DbBidsDataset
+
+
+def try_get_bids_dataset_with_path(db: Database, path: Path) -> DbBidsDataset | None:
+    """
+    Get a BIDS dataset from the database using its path, or return `None` if no dataset is found.
+    """
+
+    return db.execute(select(DbBidsDataset)
+        .where(DbBidsDataset.path == path)
+    ).scalar_one_or_none()

python/lib/db/queries/bids_file.py

@@ -0,0 +1,18 @@
+from pathlib import Path
+
+from sqlalchemy import select
+from sqlalchemy.orm import Session as Database
+
+from lib.db.models.bids_file import DbBidsFile
+
+
+def try_get_bids_file_with_dataset_id_path(db: Database, dataset_id: int, path: Path) -> DbBidsFile | None:
+    """
+    Get a BIDS file from the database using its dataset ID and path, or return `None` if no file is
+    found.
+    """
+
+    return db.execute(select(DbBidsFile)
+        .where(DbBidsFile.dataset_id == dataset_id)
+        .where(DbBidsFile.path == path)
+    ).scalar_one_or_none()

python/lib/dcm2bids_imaging_pipeline_lib/nifti_insertion_pipeline.py

@@ -676,6 +676,7 @@ def _register_into_files_and_parameter_file(self, nifti_rel_path):
             Path(nifti_rel_path),
             file_type,
             self.session,
+            None,
             self.scan_type,
             self.mri_scanner,
             self.dicom_archive,

python/lib/imaging_lib/file.py

@@ -2,6 +2,7 @@
 from datetime import date, datetime
 from pathlib import Path
 
+from lib.db.models.bids_file import DbBidsFile
 from lib.db.models.dicom_archive import DbDicomArchive
 from lib.db.models.file import DbFile
 from lib.db.models.imaging_file_type import DbImagingFileType
@@ -16,6 +17,7 @@ def register_mri_file(
     file_path: Path,
     file_type: DbImagingFileType,
     session: DbSession,
+    bids_info: DbBidsFile | None,
     scan_type: DbMriScanType | None,
     scanner: DbMriScanner | None,
     dicom_archive: DbDicomArchive | None,
@@ -46,6 +48,7 @@ def register_mri_file(
         echo_number              = echo_number,
         phase_encoding_direction = phase_encoding_direction,
         source_file_id           = None,
+        bids_info_id             = bids_info.id if bids_info is not None else None,
         scan_type_id             = scan_type.id if scan_type is not None else None,
         scanner_id               = scanner.id if scanner is not None else None,
         dicom_archive_id         = dicom_archive.id if dicom_archive is not None else None,

python/lib/physio/events.py

@@ -4,6 +4,7 @@
 from pathlib import Path
 from typing import Any
 
+from lib.db.models.bids_file import DbBidsFile
 from lib.db.models.physio_task_event_hed import DbPhysioTaskEventHed
 from lib.db.models.physio_task_event_opt import DbPhysioTaskEventOpt
 from lib.db.models.physio_task_event import DbPhysioTaskEvent
@@ -49,7 +50,12 @@ def from_file(physio_file: DbPhysioFile) -> 'EventDictFileSource':
         )
 
 
-def insert_event_dict_file(env: Env, source: EventDictFileSource, event_file_path: Path) -> DbPhysioEventFile:
+def insert_event_dict_file(
+    env: Env,
+    bids_info: DbBidsFile,
+    source: EventDictFileSource,
+    event_file_path: Path,
+) -> DbPhysioEventFile:
     """
     Insert an event dictionary file into the LORIS database.
     """
@@ -59,6 +65,7 @@ def insert_event_dict_file(env: Env, source: EventDictFileSource, event_file_pat
         project_id     = source.project.id,
         file_type      = 'json',
         file_path      = event_file_path,
+        bids_info_id   = bids_info.id,
     )
 
     env.db.add(event_dict_file)
@@ -202,7 +209,12 @@ def parse_and_insert_event_dict(
     return tag_dict
 
 
-def insert_events_file(env: Env, physio_file: DbPhysioFile, event_file_path: Path) -> DbPhysioEventFile:
+def insert_events_file(
+    env: Env,
+    physio_file: DbPhysioFile,
+    bids_info: DbBidsFile,
+    event_file_path: Path,
+) -> DbPhysioEventFile:
     """
     Insert an events file into the LORIS database.
     """
@@ -212,6 +224,7 @@ def insert_events_file(env: Env, physio_file: DbPhysioFile, event_file_path: Pat
         project_id     = physio_file.session.project.id,
         file_type      = 'tsv',
         file_path      = event_file_path,
+        bids_info_id   = bids_info.id,
     )
 
     env.db.add(event_dict_file)