Merge branch 'development_robert' of https://github.com/Hendrik-code/TPTBox into development_robert

Author: robert-graf

TPTBox/core/bids_constants.py

@@ -163,7 +163,8 @@
     "localizer",
     "difference",
     "labels",
-    "pet"
+    "report",
+    "pet",
 ]
 # https://bids-specification.readthedocs.io/en/stable/appendices/entity-table.html
 formats_relaxed = [*formats, "t2", "t1", "t2c", "t1c", "mr", "snapshot", "t1dixon", "dwi", "ctb"]
@@ -222,7 +223,7 @@
     "OPT": "Ophthalmic Tomography",
     "OPV": "Ophthalmic Visual Field",
     "OSS": "Optical Surface Scan",
-    "OT": "Other ",
+    "OT": "Other",
     "PLAN": "Plan",
     "PR": "Presentation State",
     "PT": "Positron emission tomography (PET)",

TPTBox/core/bids_files.py

@@ -687,7 +687,7 @@ def get_changed_bids(
         auto_add_run_id=False,
         additional_folder: str | None = None,
         dataset_path: str | None = None,
-        make_parent=True,
+        make_parent=False,
         non_strict_mode=False,
     ):
         ds = dataset_path if dataset_path is not None else self.get_path_decomposed()[0]

TPTBox/core/dicom/dicom_extract.py

@@ -190,7 +190,68 @@ def dicom_to_nifti_multiframe(ds, nii_path):
     return nii_path
 
 
-def _convert_to_nifti(dicom_out_path, nii_path):
+def _export_pdf_from_dicom(dcm_path, out_pdf):
+    assert len(dcm_path) == 1, dcm_path
+    ds = dcm_path[0]
+
+    # verify modality / SOP class
+    if ds.Modality.upper() != "PDF":
+        raise ValueError("Not a PDF DICOM")
+
+    if "EncapsulatedDocument" not in ds:
+        raise ValueError("No embedded PDF found")
+
+    pdf_bytes = ds.EncapsulatedDocument
+
+    out_pdf = Path(out_pdf)
+    out_pdf.write_bytes(pdf_bytes)
+
+
+def _collect_text(ds, txt_lines: list[str] | None = None):
+    if txt_lines is None:
+        txt_lines = []
+
+    def _help_collect_text(content_sequence, level: int = 0):
+        for item in content_sequence:
+            prefix = "  " * level
+
+            concept = ""
+
+            if hasattr(item, "ConceptNameCodeSequence"):
+                try:
+                    concept = item.ConceptNameCodeSequence[0].CodeMeaning
+                except Exception:
+                    pass
+
+            value = None
+
+            for attr in ["TextValue", "CodeMeaning", "NumericValue"]:
+                if hasattr(item, attr):
+                    value = getattr(item, attr)
+                    break
+
+            if concept or value is not None:
+                txt_lines.append(f"{prefix}{concept}: {value}")
+
+            if hasattr(item, "ContentSequence"):
+                _help_collect_text(
+                    item.ContentSequence,
+                    level + 1,
+                )
+
+    if hasattr(ds, "ContentSequence"):
+        _help_collect_text(ds.ContentSequence)
+    return txt_lines
+
+
+def _extract_txt_from_dicom(dcm_path, out_txt):
+    lines = []
+    for p in dcm_path:
+        lines = _collect_text(p, lines)
+    Path(out_txt).write_text("\n".join(lines))
+
+
+def _extract_nii_from_dicom(dicom_out_path, nii_path):
     """
     Convert DICOM files to NIfTI format and handle common conversion errors.
 
@@ -206,6 +267,7 @@ def _convert_to_nifti(dicom_out_path, nii_path):
         FunctionTimedOut: Raised if the DICOM-to-NIfTI conversion times out.
         ValueError: Raised for generic validation failures.
     """
+
     try:
         if isinstance(dicom_out_path, list):
             try:
@@ -217,6 +279,7 @@ def _convert_to_nifti(dicom_out_path, nii_path):
                     return True
             except Exception as e:
                 logger.on_debug("Multi-Frame DICOM did not work:", e)
+            ## The PDF dicom lands here
             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))
@@ -246,6 +309,9 @@ def _convert_to_nifti(dicom_out_path, nii_path):
         logger.print_error()
 
         return False
+    except Exception:
+        print(nii_path)
+
     return True
 
 
@@ -264,7 +330,7 @@ def _get_paths(
 ):
     if keys is None:
         keys = {}
-    (mri_format, keys) = extract_keys_from_json(
+    (mri_format, keys, ending) = extract_keys_from_json(
         simp_json,
         dcm_data_l,
         use_session,
@@ -277,7 +343,7 @@ def _get_paths(
     json_file_name, json_bids_name = _generate_bids_path(
         dataset_nifti_dir, keys, mri_format, simp_json, make_subject_chunks=make_subject_chunks, parent=parent
     )
-    nii_path = str(json_file_name).replace(".json", "") + ".nii.gz"
+    nii_path = str(json_file_name).replace(".json", "") + ending
     return json_file_name, json_bids_name, nii_path
 
 
@@ -364,11 +430,17 @@ def _from_dicom_to_nii(
     if exist and Path(nii_path).exists():
         logger.print("already exists:", json_file_name, ltype=Log_Type.STRANGE, verbose=verbose)
         return nii_path
-    suc = _convert_to_nifti(dcm_data_l, nii_path)
+    add_grid = False
+    if nii_path.endswith(".pdf"):
+        _export_pdf_from_dicom(dcm_data_l, nii_path)
+    elif nii_path.endswith(".txt"):
+        _extract_txt_from_dicom(dcm_data_l, nii_path)
+    else:
+        add_grid = _extract_nii_from_dicom(dcm_data_l, nii_path)
 
-    if suc:
+    if add_grid:
         _add_grid_info_to_json(nii_path, json_file_name)
-    return nii_path if suc else None
+    return nii_path if add_grid else None
 
 
 def _add_grid_info_to_json(nii_path: Path | str, simp_json: Path | str, force_update=False, add=True):

TPTBox/core/dicom/dicom_header_to_keys.py

@@ -165,7 +165,15 @@ def extract_keys_from_json(  # noqa: C901
     def _get(key, default=None):
         if key not in simp_json:
             return keys.get(key, default)
-        return str(simp_json[key]).replace("_", "-").replace(" ", "-").replace(".", "-")
+        value = str(simp_json[key]).replace("_", "-").replace(" ", "-").replace(".", "-")
+        # remove invalid filename characters
+        value = re.sub(r'[<>:"/\\|?*\x00-\x1F]', "", value)
+        # collapse repeated dashes
+        value = re.sub(r"-+", "-", value)
+        # strip leading/trailing dots and dashes
+        value = value.strip(".-")
+
+        return value
 
     """Extract keys from JSON based on study and series descriptions."""
     #### NAKO FIXED ####
@@ -174,25 +182,28 @@ def _get(key, default=None):
         series_description = _get("SeriesDescription", "unnamed")
         """Determine the MRI format based on the series description."""
         if "T2_TSE" in series_description:
-            return "T2w", {"acq": "sag", "chunk": series_description.split("_")[-1], "sequ": simp_json["SeriesNumber"], **keys}
+            return "T2w", {"acq": "sag", "chunk": series_description.split("_")[-1], "sequ": simp_json["SeriesNumber"], **keys}, ".nii.gz"
         elif "3D_GRE_TRA" in series_description:
-            return "vibe", {
-                "acq": "ax",
-                "part": dixon_mapping[series_description.split("_")[-1].lower()],
-                "chunk": _get("ProtocolName", "unnamed").split("_")[-1],
-                **keys,
-            }
+            return (
+                "vibe",
+                {
+                    "acq": "ax",
+                    "part": dixon_mapping[series_description.split("_")[-1].lower()],
+                    "chunk": _get("ProtocolName", "unnamed").split("_")[-1],
+                    **keys,
+                },
+                ".nii.gz",
+            )
         elif "ME_vibe" in series_description:
-            return "mevibe", {
-                "acq": "ax",
-                "part": dixon_mapping[series_description.split("_")[-1].lower()],
-                "sequ": simp_json["SeriesNumber"],
-                **keys,
-            }
+            return (
+                "mevibe",
+                {"acq": "ax", "part": dixon_mapping[series_description.split("_")[-1].lower()], "sequ": simp_json["SeriesNumber"], **keys},
+                ".nii.gz",
+            )
         elif "PD" in series_description:
-            return "pd", {"acq": "iso", **keys}
+            return "pd", {"acq": "iso", **keys}, ".nii.gz"
         elif "T2_HASTE" in series_description:
-            return "T2haste", {"acq": "ax", **keys}
+            return "T2haste", {"acq": "ax", **keys}, ".nii.gz"
         else:
             raise NotImplementedError(series_description)
     # GENERAL
@@ -321,9 +332,14 @@ def _get(key, default=None):
                     " km " in series_description.lower() or series_description.startswith("km") or series_description.endswith("km")
                 ) and keys.get("ce") is None:
                     keys["ce"] = "ContrastAgent"
+        elif modality.lower() == "pdf":
+            return "report", keys, ".pdf"
+        elif modality.lower() == "sr":
+            keys["desc"] = _get("SeriesDescription", None)
+            return "report", keys, ".txt"
         else:
-            raise NotImplementedError(f"modality='{modality.upper()}', ({modalities.get(modality.upper())})")
+            raise NotImplementedError(f"modality='{modality}', ({modalities.get(modality.upper(), 'Non Standard Modality key')})")
 
             # ".*sub.*t1.*": "subtraktion",
         # "subtraktion.*t1.*": "subtraktion",
-        return mri_format, keys
+        return mri_format, keys, ".nii.gz"

TPTBox/core/nii_wrapper.py

@@ -985,7 +985,7 @@ def resample_from_to(self, to_vox_map:Image_Reference|Has_Grid|tuple[SHAPE,AFFIN
                     # padding after = remaining dst size after src
                     pad_after = dst_shape-shift-src_shape
                     pad = tuple((int(b), int(a)) for b, a in zip(pad_before, pad_after))
-                    ret = s.apply_pad(pad, mode=mode,)
+                    ret = s.apply_pad(pad, mode=mode,inplace=inplace,verbose=verbose)
 
                     #TODO SET raise_error=False before committing
                     valid = ret.assert_affine(mapping,raise_error=True,origin_tolerance=0.0001,error_tolerance=0.0001,shape_tolerance=0)
@@ -1801,7 +1801,7 @@ def infect(self: NII, reference_mask: NII, inplace=False,verbose=True,axis:int|s
         """
         self.assert_affine(reference_mask)
         if _do_crop:
-            crop = reference_mask.compute_crop(0,5)
+            crop = reference_mask.compute_crop(0,5,raise_error=False)
             s = self.apply_crop(crop)
             reference_mask = reference_mask.apply_crop(crop)
         else:

TPTBox/segmentation/VibeSeg/inference_nnunet.py

@@ -227,7 +227,7 @@ def to_int(a: str, k: None | int = None):
                 try:
                     return enum_[a].value
                 except Exception:
-                    print("no ", enum_)
+                    pass
             if k is not None and k not in unknown_strings.values():
                 return k
             unknown_strings[a] = unknown_strings["max"]
@@ -247,7 +247,7 @@ def to_int(a: str, k: None | int = None):
         logger.print(f"{mapping=}")
         seg_nii.map_labels_(mapping)
     if out_file is not None and (not Path(out_file).exists() or override):
-        seg_nii.save(out_file)
+        seg_nii.set_dtype("smallest_uint").save(out_file)
     del nnunet
 
     torch.cuda.empty_cache()

TPTBox/stitching/stitching.py

@@ -324,6 +324,7 @@ def main(  # noqa: C901
     is_segmentation: bool = False,
     dtype: type | str = float,
     save=True,
+    ramp_path=None,
 ):
     np.set_printoptions(precision=2, floatmode="fixed")
     if is_segmentation:
@@ -529,7 +530,8 @@ def main(  # noqa: C901
             occupancy_arr = occupancy_arr[ex_slice]
         assert output is not None
         nii_occ = set_array(nii_out, occupancy_arr)
-        output = output.replace(".nii.gz", "_ramps.nii.gz")
+        nii_occ.set_data_dtype(np.int8)
+        output = output.replace(".nii.gz", "_ramps.nii.gz").replace("_msk_", "_") if ramp_path is None else ramp_path
         if save:
             nib.save(nii_occ, output)  # type: ignore
             print("Saved ", output) if verbose else None

TPTBox/stitching/stitching_tools.py

@@ -22,6 +22,7 @@ def stitching(
     dtype: type = float,
     match_histogram=False,
     store_ramp=False,
+    ramp_path=None,
 ):
     out = str(out.file["nii.gz"]) if isinstance(out, BIDS_FILE) else str(out)
     files = [to_nii(bf).nii for bf in bids_files]
@@ -37,6 +38,7 @@ def stitching(
         kick_out_fully_integrated_images=kick_out_fully_integrated_images,
         is_segmentation=is_seg,
         dtype=dtype,
+        ramp_path=ramp_path,
     )
 
 

unit_tests/test_stiching.py

@@ -78,7 +78,7 @@ def test_stitching(
         store_ramp=False,
         verbose=False,
         min_value=0,
-        bias_field=True,
+        bias_field=False,
         crop_to_bias_field=False,
         crop_empty=False,
         histogram=None,