Add new functions and test

.github/workflows/python-publish.yml

@@ -17,7 +17,7 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v4
       with:
-        python-version: '3.10'
+        python-version: "3.10"
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip

.github/workflows/tests_mr.yml

@@ -14,7 +14,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        python-version: ["3.10"]
+        python-version: ["3.9","3.12"]
 
     steps:
     - uses: actions/checkout@v4

.gitignore

@@ -164,6 +164,6 @@ tutorials/tutorial_data_processing/*
 tutorials/*PixelPandemonium/*
 tutorials/dataset-PixelPandemonium/*
 *.html
-_*.py
+#_*.py
 dicom_select
 examples

TPTBox/core/bids_constants.py

@@ -138,7 +138,15 @@
     "recon",
     "reformat",
     "subtraction",
+    "DSA",
+    "DSA3D",
+    "3DRA",
+    "XA",
     "RI",  # Raw input
+    "tmax",
+    "cbv",
+    "mtt",
+    "cbf",
     "stat",
     "snp",
     "log",
@@ -158,6 +166,88 @@
 formats_relaxed = [*formats, "t2", "t1", "t2c", "t1c", "cta", "mr", "snapshot", "t1dixon", "dwi"]
 # Recommended writing style: T1c, T2c; This list is not official and can be extended.
 
+modalities = {
+    "AR": "Autorefraction",
+    "AS": "Angioscopy (Retired)",
+    "ASMT": "Content Assessment Results",
+    "AU": "Audio",
+    "BDUS": "Bone Densitometry (ultrasound)",
+    "BI": "Biomagnetic imaging",
+    "BMD": "Bone Densitometry (X-Ray)",
+    "CD": "Color flow Doppler (Retired)",
+    "CF": "Cinefluorography (Retired)",
+    "CP": "Colposcopy (Retired)",
+    "CR": "Computed Radiography",
+    "CS": "Cystoscopy (Retired)",
+    "CT": "Computed Tomography",
+    "DD": "Duplex Doppler (Retired)",
+    "DF": "Digital fluoroscopy (Retired)",
+    "DG": "Diaphanography",
+    "DM": "Digital microscopy (Retired)",
+    "DOC": "Document",
+    "DS": "Digital Subtraction Angiography (Retired)",
+    "DX": "Digital Radiography",
+    "EC": "Echocardiography (Retired)",
+    "ECG": "Electrocardiography",
+    "EPS": "Cardiac Electrophysiology",
+    "ES": "Endoscopy",
+    "FA": "Fluorescein angiography (Retired)",
+    "FID": "Fiducials",
+    "FS": "Fundoscopy (Retired)",
+    "GM": "General Microscopy ",
+    "HC": "Hard Copy",
+    "HD": "Hemodynamic Waveform",
+    "IO": "Intra-Oral Radiography",
+    "IOL": "Intraocular Lens Data",
+    "IVOCT": "Intravascular Optical Coherence Tomography",
+    "IVUS": "Intravascular Ultrasound",
+    "KER": "Keratometry",
+    "KO": "Key Object Selection",
+    "LEN": "Lensometry",
+    "LP": "Laparoscopy (Retired)",
+    "LS": "Laser surface scan",
+    "MA": "Magnetic resonance angiography (Retired)",
+    "MG": "Mammography",
+    "MR": "Magnetic Resonance",
+    "MS": "Magnetic resonance spectroscopy (Retired)",
+    "NM": "Nuclear Medicine",
+    "OAM": "Ophthalmic Axial Measurements",
+    "OCT": "Optical Coherence Tomography (non-Ophthalmic)",
+    "OP": "Ophthalmic Photography",
+    "OPM": "Ophthalmic Mapping",
+    "OPR": "Ophthalmic Refraction (Retired)",
+    "OPT": "Ophthalmic Tomography",
+    "OPV": "Ophthalmic Visual Field",
+    "OSS": "Optical Surface Scan",
+    "OT": "Other ",
+    "PLAN": "Plan",
+    "PR": "Presentation State",
+    "PT": "Positron emission tomography (PET)",
+    "PX": "Panoramic X-Ray",
+    "REG": "Registration",
+    "RESP": "Respiratory Waveform",
+    "RF": "Radio Fluoroscopy",
+    "RG": "Radiographic imaging (conventional film/screen)",
+    "RTDOSE": "Radiotherapy Dose",
+    "RTIMAGE": "Radiotherapy Image",
+    "RTPLAN": "Radiotherapy Plan",
+    "RTRECORD": "RT Treatment Record",
+    "RTSTRUCT": "Radiotherapy Structure Set",
+    "RWV": "Real World Value Map",
+    "SEG": "Segmentation",
+    "SM": "Slide Microscopy",
+    "SMR": "Stereometric Relationship",
+    "SR": "SR Document",
+    "SRF": "Subjective Refraction",
+    "ST": "Single-photon emission computed tomography (SPECT) (Retired)",
+    "STAIN": "Automated Slide Stainer",
+    "TG": "Thermography",
+    "US": "Ultrasound",
+    "VA": "Visual Acuity",
+    "VF": "Videofluorography (Retired)",
+    "XA": "X-Ray Angiography",
+    "XC": "External-camera Photography",
+}
 
 # Actual official final folder
 # func (task based and resting state functional MRI)
@@ -246,6 +336,7 @@
     # Others (never used)
     "Split": "split",
     "Density": "den",
+    "version": "version",
     "Description": "desc",
     "nameconflict": "nameconflict",
 }

TPTBox/core/bids_files.py

@@ -793,7 +793,8 @@ def get_changed_path(  # noqa: C901
                 if key in same_info:
                     continue
                 if value is not None:
-                    assert validate_entities(key, value, f"..._{key}-{value}_...", True)
+                    if not non_strict_mode:
+                        assert validate_entities(key, value, f"..._{key}-{value}_...", True), f"..._{key}-{value}_..."
                     final_info[key] = value
                 # file_name += f"{key}-{value}_"
             # sort by order

TPTBox/core/dicom/dicom2nii_utils.py

@@ -3,6 +3,7 @@
 import json
 import pickle
 from copy import deepcopy
+from datetime import date
 from pathlib import Path
 
 import numpy as np
@@ -152,9 +153,56 @@ def clean_dicom_data(dcm_data) -> dict:
     for tag in ["00291010", "00291020"]:
         if tag in py_dict and "InlineBinary" in py_dict[tag]:
             del py_dict[tag]["InlineBinary"]
+    py_dict = replace_birthdate_with_age(py_dict)
     return py_dict
 
 
+def replace_birthdate_with_age(d):
+    try:
+        # DICOM tags
+        BIRTH_TAG = "00100030"  # PatientBirthDate
+        STUDY_DATE_TAG = "00080020"  # StudyDate
+        AGE_TAG = "00101010"  # PatientAge
+
+        birth_str = d.get(BIRTH_TAG, {}).get("Value", [None])[0]
+        study_str = d.get(STUDY_DATE_TAG, {}).get("Value", [None])[0]
+
+        if not birth_str:
+            return d  # no birth date, nothing to do
+
+        # Parse birth date safely
+        try:
+            year = int(birth_str[:4])
+            month = int(birth_str[4:6]) if len(birth_str) >= 6 and birth_str[4:6] != "00" else 6
+            day = int(birth_str[6:8]) if len(birth_str) == 8 and birth_str[6:8] != "00" else 15
+            birth_date = date(year, month, day)
+        except Exception:
+            return d  # invalid date format, skip
+
+        # Reference date (study date or today)
+        try:
+            ref_date = date(
+                int(study_str[:4]),
+                int(study_str[4:6]) if study_str[4:6] != "00" else 6,
+                int(study_str[6:8]) if study_str[6:8] != "00" else 15,
+            )
+        except Exception:
+            ref_date = date.today()
+
+        # Compute integer age
+        age = ref_date.year - birth_date.year - ((ref_date.month, ref_date.day) < (birth_date.month, birth_date.day))
+
+        # Replace PatientBirthDate with PatientAge
+        d.pop(BIRTH_TAG, None)
+        d[AGE_TAG] = {
+            "vr": "AS",  # Age String
+            "Value": [f"{age:03d}Y"],  # DICOM age format (e.g. '034Y')
+        }
+    except Exception:
+        pass
+    return d
+
+
 def get_json_from_dicom(data: list[pydicom.FileDataset] | pydicom.FileDataset):
     if isinstance(data, list):
         data = data[0]

TPTBox/core/dicom/dicom_extract.py

@@ -11,6 +11,7 @@
 
 import dicom2nifti
 import dicom2nifti.exceptions
+import nibabel as nib
 import numpy as np
 import pydicom
 from dicom2nifti import convert_dicom
@@ -87,6 +88,74 @@ def _generate_bids_path(
     return fname.file["json"], fname
 
 
+def dicom_to_nifti_multiframe(ds, nii_path):
+    pixel_array = ds.pixel_array
+    if len(pixel_array.shape) != 3 and len(pixel_array.shape) != 4:
+        raise ValueError(f"Expected a shape with 3 colums not {len(pixel_array.shape)}; {pixel_array.shape=}")
+    n_frames = pixel_array.shape[0]
+
+    # Pixel spacing (mm)
+    if hasattr(ds, "PixelSpacing"):
+        dy, dx = (float(v) for v in ds.PixelSpacing)
+        # Image orientation (row and column direction cosines)
+        orientation = [float(v) for v in ds.ImageOrientationPatient]
+        row_cosines = np.array(orientation[0:3])
+        col_cosines = np.array(orientation[3:6])
+        # Normal vector (slice direction)
+        slice_cosines = np.cross(row_cosines, col_cosines)
+
+        # Image position (origin of first slice)
+        origin = np.array([float(v) for v in ds.ImagePositionPatient])
+        # Slice spacing - robust: Abstand zwischen Slice 0 und 1
+        if n_frames > 1:
+            pos1 = np.array([float(v) for v in ds.PerFrameFunctionalGroupsSequence[0].PlanePositionSequence[0].ImagePositionPatient])
+            pos2 = np.array([float(v) for v in ds.PerFrameFunctionalGroupsSequence[1].PlanePositionSequence[0].ImagePositionPatient])
+            dz = np.linalg.norm(pos2 - pos1)
+        else:
+            dz = float(getattr(ds, "SpacingBetweenSlices", ds.SliceThickness))
+
+        # Affine bauen
+        affine = np.eye(4)
+        affine[0:3, 0] = row_cosines * dx
+        affine[0:3, 1] = col_cosines * dy
+        affine[0:3, 2] = slice_cosines * dz
+        affine[0:3, 3] = origin
+        nii = nib.Nifti1Image(np.transpose(pixel_array, (2, 1, 0)), affine)
+
+    elif hasattr(ds, "ImagerPixelSpacing"):
+        dy, dx = (float(v) for v in ds.ImagerPixelSpacing)
+        # Einfaches affine (nur 2D + Zeit, keine Lage im Patientenraum)
+        affine = np.eye(4)
+        affine[0, 0] = -dx
+        affine[1, 1] = -dy
+        nii = nib.Nifti1Image(np.transpose(pixel_array, (2, 1, 0)), affine)
+
+    else:
+        if hasattr(ds, "RelatedSeriesSequence"):
+            raise NotImplementedError("RelatedSeriesSequence Affine lookup not implemented")
+        raise NotImplementedError("No spatial metadata found")
+        ### Some could be solved by looking up the "RelatedSeriesSequence"
+        # "RelatedSeriesSequence": [
+        # {
+        #    "StudyInstanceUID": "1.2.276.0.38.1.1.1.7712.20250929100319.54200288",
+        #    "SeriesInstanceUID": "1.3.46.670589.7.8.1.6.1403526999.1.9608.1759142950287.2",
+        #    "PurposeOfReferenceCodeSequence": []
+        # }
+        # ],
+        # --- No geometry info (e.g. RGB screen captures or video frames) ---
+        print("⚠️ No spatial metadata found — assuming pixel size = 1mm and identity orientation.")
+        affine = np.eye(4)
+        affine[0, 0] = 1.0
+        affine[1, 1] = 1.0
+        affine[2, 2] = 1.0
+        nii = nib.Nifti1Image(pixel_array, affine)
+
+    # Reihenfolge anpassen: Nibabel erwartet (X,Y,Z)
+    nib.save(nii, nii_path)
+
+    return nii_path
+
+
 def _convert_to_nifti(dicom_out_path, nii_path):
     """
     Convert DICOM files to NIfTI format and handle common conversion errors.
@@ -105,15 +174,25 @@ def _convert_to_nifti(dicom_out_path, nii_path):
     """
     try:
         if isinstance(dicom_out_path, list):
+            try:
+                if len(dicom_out_path) == 1:
+                    ds = dicom_out_path[0]
+                    if hasattr(ds, "pixel_array") and len(ds.pixel_array.shape) >= 2:
+                        dicom_to_nifti_multiframe(ds, nii_path)
+
+                    return True
+            except Exception as e:
+                logger.on_debug("Multi-Frame DICOM did not work:", e)
             convert_dicom.dicom_array_to_nifti(dicom_out_path, nii_path, True)
         else:
             # func_timeout(10, dicom2nifti.dicom_series_to_nifti, (dicom_out_path, nii_path, True))
             dicom2nifti.dicom_series_to_nifti(dicom_out_path, nii_path, True)
         logger.print("Save ", nii_path, Log_Type.SAVE)
     except dicom2nifti.exceptions.ConversionValidationError as e:
         if e.args[0] in ["NON_IMAGING_DICOM_FILES"]:
+            s = f"dicom_array_to_nifti len={len(dicom_out_path)}" if isinstance(dicom_out_path, list) else "dicom_series_to_nifti"
             Path(str(nii_path).replace(".nii.gz", ".json")).unlink(missing_ok=True)
-            logger.on_debug(f"Not exportable '{Path(nii_path).name}':", e.args[0])
+            logger.on_debug(f"Not exportable '{Path(nii_path).name}':", e.args[0], s)
             return False
         for key, reason in [
             ("validate_orthogonal", "NON_CUBICAL_IMAGE/GANTRY_TILT"),
@@ -311,7 +390,7 @@ def _read_dicom_files(dicom_out_path: Path) -> tuple[dict[str, list[FileDataset]
         path = Path(_paths)
         if path.is_file():
             try:
-                dcm_data = pydicom.dcmread(path, defer_size="1 KB", force=True)
+                dcm_data = pydicom.dcmread(path, defer_size="1 KB", force=True)  # , stop_before_pixels=True
                 try:
                     typ = (
                         str(dcm_data.get_item((0x0008, 0x0008)).value)
@@ -324,7 +403,6 @@ def _read_dicom_files(dicom_out_path: Path) -> tuple[dict[str, list[FileDataset]
                 except Exception:
                     typ = ""
                 key1 = str(dcm_data.SeriesInstanceUID)
-
                 key = f"{key1}_{typ}"
                 if not hasattr(dcm_data, "ImageOrientationPatient"):
                     key += "_" + dcm_data.get("SOPInstanceUID", 0)
@@ -437,6 +515,7 @@ def extract_dicom_folder(
     validate_slicecount=True,
     validate_orientation=True,
     validate_orthogonal=False,
+    validate_slice_increment=True,
     n_cpu: int | None = 1,
     override_subject_name: Callable[[dict, Path], str] | None = None,
     skip_localizer=True,
@@ -463,7 +542,8 @@ def extract_dicom_folder(
         convert_dicom.settings.disable_validate_orientation()
     if not validate_orthogonal:
         convert_dicom.settings.disable_validate_orthogonal()
-
+    if not validate_slice_increment:
+        convert_dicom.settings.disable_validate_slice_increment()
     outs = {}
 
     for p in _find_all_files(dicom_folder):
@@ -512,6 +592,8 @@ def process_series(key, files, parts):
                     try:
                         key2, out = process_series(key, files, parts)
                         outs[key2] = out
+                    except NotImplementedError as e:
+                        logger.on_warning("NotImplementedError:", e)
                     except Exception:
                         logger.print_error()
 
@@ -523,6 +605,10 @@ def process_series(key, files, parts):
 
 
 if __name__ == "__main__":
+    for p in Path("/DATA/NAS/datasets_source/brain/dsa").iterdir():
+        extract_dicom_folder(p, Path("/DATA/NAS/datasets_source/brain/", "dataset-DSA"), False, False, validate_slice_increment=False)
+
+    sys.exit()
     # s = "/home/robert/Downloads/bein/dataset-oberschenkel/rawdata/sub-1-3-46-670589-11-2889201787-2305829596-303261238-2367429497/mr/sub-1-3-46-670589-11-2889201787-2305829596-303261238-2367429497_sequ-406_mr.nii.gz"
     # nii2 = NII.load(s, False)
     # print(nii2.affine, nii2.orientation)