Merge pull request #147 from sustrev/main

E2E UID and birthdate error handling

oct_converter/readers/e2e.py

@@ -8,6 +8,7 @@
 from pathlib import Path
 
 import numpy as np
+from construct.core import StreamError
 
 from oct_converter.image_types import FundusImageWithMetaData, OCTVolumeWithMetaData
 from oct_converter.readers.binary_structs import e2e_binary
@@ -137,14 +138,26 @@ def _make_lut():
                         self.sex = patient_data.sex
                         self.first_name = patient_data.first_name
                         self.surname = patient_data.surname
-                        julian_birthdate = (patient_data.birthdate / 64) - 14558805
-                        self.birthdate = self.julian_to_ymd(julian_birthdate)
-                        # TODO: There are conflicting ideas of how to parse E2E's birthdate
-                        # https://bitbucket.org/uocte/uocte/wiki/Heidelberg%20File%20Format suggests the above,
-                        # whereas https://github.com/neurodial/LibE2E/blob/master/E2E/dataelements/patientdataelement.cpp
-                        # suggests that DOB is given as a Windows date. Neither option seems accurate to
-                        # test files with known-correct birthdates. More investigation is needed.
                         self.patient_id = patient_data.patient_id
+                        if len(str(patient_data.birthdate)) == 8:
+                            # Encountered a file where birthdate had been stored as YYYYMMDD,
+                            # this is an attempt to catch that.
+                            self.birthdate = str(patient_data.birthdate)
+                        else:
+                            try:
+                                julian_birthdate = (
+                                    patient_data.birthdate / 64
+                                ) - 14558805
+                                self.birthdate = self.julian_to_ymd(julian_birthdate)
+                                # TODO: There are conflicting ideas of how to parse E2E's birthdate
+                                # https://bitbucket.org/uocte/uocte/wiki/Heidelberg%20File%20Format suggests the above,
+                                # whereas https://github.com/neurodial/LibE2E/blob/master/E2E/dataelements/patientdataelement.cpp
+                                # suggests that DOB is given as a Windows date. Neither option seems accurate to
+                                # test files with known-correct birthdates. More investigation is needed.
+                            except ValueError:
+                                # If the julian_to_ymd function cannot parse it into a date obj,
+                                # it throws a ValueError
+                                self.birthdate = None
                     except Exception:
                         pass
 
@@ -564,11 +577,14 @@ def _convert_to_dict(container):
                     metadata["time_data"].append(_convert_to_dict(time_data))
 
                 elif chunk.type in [52, 54, 1000, 1001]:  # various UIDs
-                    raw = f.read(chunk.size)
-                    uid_data = e2e_binary.uid_data.parse(raw)
-                    metadata["uid_data"].append(
-                        {chunk.type: _convert_to_dict(uid_data)}
-                    )
+                    try:
+                        raw = f.read(chunk.size)
+                        uid_data = e2e_binary.uid_data.parse(raw)
+                        metadata["uid_data"].append(
+                            {chunk.type: _convert_to_dict(uid_data)}
+                        )
+                    except StreamError:
+                        pass
 
                 # Chunks 1005, 1006, and 1007 seem to contain strings of device data,
                 # including some servicers and distributors and other entities,