Migrate the DICOM study importer to pathlib.Path (path PR 1) (#1349)

Author: MaximeBICMTL

python/lib/config.py

@@ -1,4 +1,5 @@
 import os
+from pathlib import Path
 from typing import Literal
 
 from lib.db.queries.config import try_get_config_with_setting_name
@@ -26,15 +27,15 @@ def get_patient_id_dicom_header_config(env: Env) -> Literal['PatientID', 'Patien
     return patient_id_dicom_header
 
 
-def get_data_dir_path_config(env: Env) -> str:
+def get_data_dir_path_config(env: Env) -> Path:
     """
     Get the LORIS base data directory path from the in-database configuration, or exit the program
     with an error if that configuration value does not exist or is incorrect.
     """
 
-    data_dir_path = os.path.normpath(_get_config_value(env, 'dataDirBasepath'))
+    data_dir_path = Path(_get_config_value(env, 'dataDirBasepath'))
 
-    if not os.path.isdir(data_dir_path):
+    if not data_dir_path.is_dir():
         log_error_exit(
             env,
             (
@@ -52,20 +53,20 @@ def get_data_dir_path_config(env: Env) -> str:
     return data_dir_path
 
 
-def get_dicom_archive_dir_path_config(env: Env) -> str:
+def get_dicom_archive_dir_path_config(env: Env) -> Path:
     """
     Get the LORIS DICOM archive directory path from the in-database configuration, or exit the
     program with an error if that configuration value does not exist or is incorrect.
     """
 
-    dicom_archive_dir_path = os.path.normpath(_get_config_value(env, 'tarchiveLibraryDir'))
+    dicom_archive_dir_path = Path(_get_config_value(env, 'tarchiveLibraryDir'))
 
-    if not os.path.isdir(dicom_archive_dir_path):
+    if not dicom_archive_dir_path.is_dir():
         log_error_exit(
             env,
             (
                 f"The LORIS DICOM archive directory path configuration value '{dicom_archive_dir_path}' does not refer"
-                " to an existing diretory."
+                " to an existing directory."
             ),
         )
 

python/lib/dcm2bids_imaging_pipeline_lib/nifti_insertion_pipeline.py

@@ -689,7 +689,7 @@ def _create_pic_image(self):
         """
         file_info = {
             'cand_id': self.session.candidate.cand_id,
-            'data_dir_path': self.data_dir,
+            'data_dir_path': str(self.data_dir),
             'file_rel_path': self.assembly_nifti_rel_path,
             'is_4D_dataset': self.json_file_dict['time'] is not None,
             'file_id': self.file_id

python/lib/import_dicom_study/dicom_database.py

@@ -1,5 +1,6 @@
 from datetime import datetime
 from functools import cmp_to_key
+from pathlib import Path
 
 from sqlalchemy.orm import Session as Database
 
@@ -17,14 +18,14 @@ def insert_dicom_archive(
     db: Database,
     dicom_summary: DicomStudySummary,
     dicom_import_log: DicomStudyImportLog,
-    archive_location: str,
+    archive_path: Path,
 ):
     """
     Insert a DICOM archive in the database.
     """
 
     dicom_archive = DbDicomArchive()
-    populate_dicom_archive(dicom_archive, dicom_summary, dicom_import_log, archive_location)
+    populate_dicom_archive(dicom_archive, dicom_summary, dicom_import_log, archive_path)
     dicom_archive.date_first_archived = datetime.now()
     db.add(dicom_archive)
     db.commit()
@@ -37,7 +38,7 @@ def update_dicom_archive(
     dicom_archive: DbDicomArchive,
     dicom_summary: DicomStudySummary,
     dicom_import_log: DicomStudyImportLog,
-    archive_location: str,
+    archive_path: Path,
 ):
     """
     Update a DICOM archive in the database.
@@ -47,7 +48,7 @@ def update_dicom_archive(
     delete_dicom_archive_file_series(db, dicom_archive)
 
     # Update the database record with the new DICOM information.
-    populate_dicom_archive(dicom_archive, dicom_summary, dicom_import_log, archive_location)
+    populate_dicom_archive(dicom_archive, dicom_summary, dicom_import_log, archive_path)
     db.commit()
 
     # Insert the new DICOM files and series.
@@ -58,7 +59,7 @@ def populate_dicom_archive(
     dicom_archive: DbDicomArchive,
     dicom_summary: DicomStudySummary,
     dicom_import_log: DicomStudyImportLog,
-    archive_location: str,
+    archive_path: Path,
 ):
     """
     Populate a DICOM archive database object with information from its DICOM summary and DICOM
@@ -83,8 +84,8 @@ def populate_dicom_archive(
     dicom_archive.creating_user            = dicom_import_log.creator_name
     dicom_archive.sum_type_version         = dicom_import_log.summary_version
     dicom_archive.tar_type_version         = dicom_import_log.archive_version
-    dicom_archive.source_location          = dicom_import_log.source_path
-    dicom_archive.archive_location         = archive_location
+    dicom_archive.source_location          = str(dicom_import_log.source_path)
+    dicom_archive.archive_location         = str(archive_path)
     dicom_archive.scanner_manufacturer     = dicom_summary.info.scanner.manufacturer or ''
     dicom_archive.scanner_model            = dicom_summary.info.scanner.model or ''
     dicom_archive.scanner_serial_number    = dicom_summary.info.scanner.serial_number or ''

python/lib/import_dicom_study/import_log.py

@@ -3,6 +3,7 @@
 import socket
 from dataclasses import dataclass
 from datetime import datetime
+from pathlib import Path
 
 from lib.import_dicom_study.text_dict import DictWriter
 
@@ -13,8 +14,8 @@ class DicomStudyImportLog:
     Information about the past import of a DICOM study.
     """
 
-    source_path: str
-    target_path: str
+    source_path: Path
+    target_path: Path
     creator_host: str
     creator_os: str
     creator_name: str
@@ -32,8 +33,8 @@ def write_dicom_study_import_log_to_string(import_log: DicomStudyImportLog):
     """
 
     return DictWriter([
-        ("Taken from dir",                   import_log.source_path),
-        ("Archive target location",          import_log.target_path),
+        ("Taken from dir",                   str(import_log.source_path)),
+        ("Archive target location",          str(import_log.target_path)),
         ("Name of creating host",            import_log.creator_host),
         ("Name of host OS",                  import_log.creator_os),
         ("Created by user",                  import_log.creator_name),
@@ -46,7 +47,7 @@ def write_dicom_study_import_log_to_string(import_log: DicomStudyImportLog):
     ]).write()
 
 
-def write_dicom_study_import_log_to_file(import_log: DicomStudyImportLog, file_path: str):
+def write_dicom_study_import_log_to_file(import_log: DicomStudyImportLog, file_path: Path):
     """
     Serialize a DICOM study import log into a text file.
     """
@@ -56,7 +57,7 @@ def write_dicom_study_import_log_to_file(import_log: DicomStudyImportLog, file_p
         file.write(string)
 
 
-def make_dicom_study_import_log(source: str, target: str, tarball_md5_sum: str, zipball_md5_sum: str):
+def make_dicom_study_import_log(source: Path, target: Path, tarball_md5_sum: str, zipball_md5_sum: str):
     """
     Create a DICOM study import log from the provided arguments about a DICOM study, as well as the
     current execution environment.

python/lib/import_dicom_study/summary_get.py

@@ -1,4 +1,5 @@
 import os
+from pathlib import Path
 
 import pydicom
 import pydicom.errors
@@ -17,7 +18,7 @@
 from lib.util.fs import iter_all_dir_files
 
 
-def get_dicom_study_summary(dicom_study_dir_path: str, verbose: bool):
+def get_dicom_study_summary(dicom_study_dir_path: Path, verbose: bool):
     """
     Get information about a DICOM study by reading the files in the DICOM study directory.
     """
@@ -31,7 +32,7 @@ def get_dicom_study_summary(dicom_study_dir_path: str, verbose: bool):
         if verbose:
             print(f"Processing file '{file_rel_path}' ({i}/{len(file_rel_paths)})")
 
-        file_path = os.path.join(dicom_study_dir_path, file_rel_path)
+        file_path = dicom_study_dir_path / file_rel_path
 
         try:
             dicom = pydicom.dcmread(file_path)  # type: ignore
@@ -112,13 +113,13 @@ def get_dicom_file_info(dicom: pydicom.Dataset) -> DicomStudyDicomFile:
     )
 
 
-def get_other_file_info(file_path: str) -> DicomStudyOtherFile:
+def get_other_file_info(file_path: Path) -> DicomStudyOtherFile:
     """
     Get information about a non-DICOM file within a DICOM study.
     """
 
     return DicomStudyOtherFile(
-        os.path.basename(file_path),
+        file_path.name,
         compute_file_md5_hash(file_path),
     )
 

python/lib/import_dicom_study/summary_write.py

@@ -1,5 +1,6 @@
 import xml.etree.ElementTree as ET
 from functools import cmp_to_key
+from pathlib import Path
 
 from lib.import_dicom_study.summary_type import (
     DicomStudyDicomFile,
@@ -14,14 +15,14 @@
 from lib.util.iter import count, flatten
 
 
-def write_dicom_study_summary_to_file(dicom_summary: DicomStudySummary, filename: str):
+def write_dicom_study_summary_to_file(dicom_summary: DicomStudySummary, file_path: Path):
     """
     Serialize a DICOM study summary object into a text file.
     """
 
-    string = write_dicom_study_summary(dicom_summary)
-    with open(filename, 'w') as file:
-        file.write(string)
+    summary = write_dicom_study_summary(dicom_summary)
+    with open(file_path, 'w') as file:
+        file.write(summary)
 
 
 def write_dicom_study_summary(dicom_summary: DicomStudySummary) -> str:

python/lib/import_dicom_study/text.py

@@ -3,8 +3,8 @@
 different types of values.
 """
 
-import os
 from datetime import date, datetime
+from pathlib import Path
 
 from lib.util.crypto import compute_file_md5_hash
 
@@ -66,9 +66,9 @@ def read_float_none(string: str | None):
     return float(string)
 
 
-def compute_md5_hash_with_name(path: str):
+def compute_md5_hash_with_name(path: Path):
     """
     Get the MD5 sum hash of a file with the filename appended.
     """
 
-    return f'{compute_file_md5_hash(path)}   {os.path.basename(path)}'
+    return f'{compute_file_md5_hash(path)}   {path.name}'

python/lib/util/crypto.py

@@ -1,7 +1,8 @@
 import hashlib
+from pathlib import Path
 
 
-def compute_file_blake2b_hash(file_path: str) -> str:
+def compute_file_blake2b_hash(file_path: Path | str) -> str:
     """
     Compute the BLAKE2b hash of a file.
     """
@@ -15,7 +16,7 @@ def compute_file_blake2b_hash(file_path: str) -> str:
     return hash.hexdigest()
 
 
-def compute_file_md5_hash(file_path: str) -> str:
+def compute_file_md5_hash(file_path: Path | str) -> str:
     """
     Compute the MD5 hash of a file.
     """

python/lib/util/fs.py

@@ -5,6 +5,7 @@
 import tempfile
 from collections.abc import Iterator
 from datetime import datetime
+from pathlib import Path
 
 import lib.exitcode
 from lib.env import Env
@@ -26,16 +27,17 @@ def extract_archive(env: Env, tar_path: str, prefix: str, dir_path: str) -> str:
     return extract_path
 
 
-def iter_all_dir_files(dir_path: str) -> Iterator[str]:
+def iter_all_dir_files(dir_path: Path) -> Iterator[Path]:
     """
     Iterate through all the files in a directory recursively, and yield the path of each file
     relative to that directory.
     """
 
-    for sub_dir_path, _, file_names in os.walk(dir_path):
-        for file_name in file_names:
-            file_path = os.path.join(sub_dir_path, file_name)
-            yield os.path.relpath(file_path, start=dir_path)
+    for item_path in dir_path.iterdir():
+        if item_path.is_dir():
+            yield from iter_all_dir_files(item_path)
+        elif item_path.is_file():
+            yield item_path.relative_to(dir_path)
 
 
 def remove_directory(env: Env, path: str):