DicomExtract (#8)

* attempted to add DICOM Extraction * changed file name to match class * adding function for mCT metadata extraction * Add parameter extraction code * Delete MetadataExtractor/Extractors/Data/ParameterExtract.py removed superfluous file * Delete MetadataExtractor/Extractors/Data/mCTExtract.py removed superfluous file * Update IDataExtract.py Removed , "dcm" from default * Updated DicomExtract.py to use dicom_metadata Removed the undefined attribute parameter. * adding pydicom, adding DicomExtract to config new file: Examples/2022-03-07_18h21_M3_Cy7.mCT new file: Examples/2022-03-07_18h21_M3_Cy7.parameters new file: Examples/EnIm1.dcm modified: MetadataExtractor/Extractors/Data/DicomExtract.py modified: MetadataExtractor/Extractors/Data/IDataExtract.py deleted: MetadataExtractor/Extractors/Data/ParameterExtract.py deleted: MetadataExtractor/Extractors/Data/mCTExtract.py modified: defaultConfigs.py modified: requirements.txt * dicom extractor appears to work now
BenediktHeinrichs · Apr 24, 2024 · d1fde15 · d1fde15
1 parent 6873e08
commit d1fde15
Show file tree

Hide file tree

Showing 6 changed files with 117 additions and 2 deletions.
diff --git a/Examples/2022-03-07_18h21_M3_Cy7.mCT b/Examples/2022-03-07_18h21_M3_Cy7.mCT
@@ -0,0 +1,31 @@
+# mCT settings file v1.0
+
+[mCT]
+name               = FLT totalbody normal
+magnification      = total_body
+scan_angle         = 360.000000
+scan_mode          = normal
+lin_stage_position = -1.000000
+mode               = step_and_shoot
+angle_start        = 0.000000
+angle_speed_deg_s  = 0.000000
+angle_step_deg     = 0.750000
+binning            = x11
+exposure_ms        = 75.000000
+frame_averaging    = 1
+xray_voltage_kV    = 55.000000
+xray_current_mA    = 0.170000
+exposure2_ms       = 0.000000
+xray_voltage2_kV   = 0.000000
+xray_current2_mA   = 0.000000
+origin_offset_z    = 236.200000
+HU_calibration     = 20342.444444
+BeamHardening      = 0.245556
+HU_calibration2    = 0.000000
+BeamHardening2     = 0.000000
+NrScans            = 1
+gating_proj_per_angle = 1
+filter_1           = Al400
+filter_2           = 
+fixed_filter       = Al 100�m
+Pitch (if helical) = 0.000000
diff --git a/Examples/2022-03-07_18h21_M3_Cy7.parameters b/Examples/2022-03-07_18h21_M3_Cy7.parameters
@@ -0,0 +1,23 @@
+# U-SPECT parameter file v4.0
+
+[System]
+ScannerID       = "80481"
+SoftwareVersion = "12.27-st"
+ScannerType     = "U-CT"
+
+[Animal]
+ScanID          = "2022-03-07_18h21_M3_Cy7"
+AnimalID        = "SS_FluoInserts"
+
+[Images]
+FovXY_mm        = 88.000000    
+FovZ_mm         = 230.000000    
+
+[Acquisition]
+AcqStartTime    = "2022-03-07 18:50:00"
+SPECT           = 0    
+CT              = 1    
+OI              = 0    
+FLT             = 1    
+BLT             = 0    
+Fluoroscopy     = 0    
diff --git a/Examples/EnIm1.dcm b/Examples/EnIm1.dcm
diff --git a/MetadataExtractor/Extractors/Data/DicomExtract.py b/MetadataExtractor/Extractors/Data/DicomExtract.py
@@ -0,0 +1,59 @@
+from .IDataExtract import IDataExtract
+from MetadataExtractor.Util import metadataCreation, metadataFormatter
+import pydicom
+from pathlib import Path
+from pydicom._version import __version_info__
+import logging
+
+log = logging.getLogger(__name__)
+
+
+class DicomExtract(IDataExtract):
+    def extract(self, fileInfo):
+        log.info('Extracting metadata for Dicom file "' + fileInfo["file"] + '".')
+        text = ""
+
+        dicom_metadata = pydicom.dcmread(fileInfo["file"])
+
+        for elem in dicom_metadata.iterall():
+            if elem.tag.is_private:
+                continue  # Skip private tags
+            tag_str = elem.tag
+            value = elem.value
+            if isinstance(value, pydicom.valuerep.DSfloat):
+                value = float(value)
+            logging.info(dicom_metadata.keys)
+                # Construct a DataElement instance
+            data_element = pydicom.DataElement(tag_str, elem.VR, value)
+            dicom_metadata[tag_str] = data_element
+
+        values = []
+        identifier = fileInfo["identifier"]
+
+        for attribute in dicom_metadata.keys():
+            objectValue = dicom_metadata[attribute]
+            values.append(
+                {
+                    "predicate": "mexattr:" + metadataFormatter.replaceForbiddenValues(str(attribute)),
+                    "object": objectValue,
+                }
+            )
+
+        metadata = metadataCreation.addEntryToFileGraph(
+            fileInfo,
+            self._IExtract__config,
+            {
+                "additionalPrefixes": [
+                    "@prefix dcat: <http://www.w3.org/ns/dcat#>",
+                    "@prefix dcterms: <http://purl.org/dc/terms/>",
+                ],
+                "identifier": identifier,
+                "ontology": "mex",
+                "values": values,
+            },
+        )
+
+        return (text, metadata)
+
+    def registerMimeTypes(self):
+        self.mimeTypes["concrete"] = ["application/dcm", "application/dicom"]
diff --git a/defaultConfigs.py b/defaultConfigs.py
@@ -73,7 +73,8 @@ def getDefaultConfig():
             "Data": [
                 "ExeExtract",
                 "FcsExtract",
-                "Hdf5Extract"
+                "Hdf5Extract",
+                "DicomExtract"
             ],
             "Image": [
                 "DescriptiveImageExtract",

diff --git a/requirements.txt b/requirements.txt
@@ -52,4 +52,5 @@ pandas==1.5.1
 flowkit==1.0.0
 openai-whisper
 soundfile
-pefile==2023.2.7
+pefile==2023.2.7
+pydicom==2.4.4