Support dicomweb fields without Value/uri/binary

Author: realcundo

Cargo.toml

@@ -2,7 +2,7 @@
 name = "nu_plugin_dcm"
 version = "0.3.1"
 license = "MIT"
-description = "A nushell plugin to parse Dicom files"
+description = "A nushell plugin to parse Dicom files and DICOMweb records"
 homepage = "https://github.com/realcundo/nu_plugin_dcm"
 repository = "https://github.com/realcundo/nu_plugin_dcm"
 readme = "README.md"

README.md

@@ -8,14 +8,14 @@ Use `ls *.dcm | dcm`, `ls *.dcm | get name` or `ls *.dcm | select name type | dc
 A [nushell](https://www.nushell.sh/) plugin to parse [DICOM](https://en.wikipedia.org/wiki/DICOM) objects.
 
 This plugin is in the early stage of the development. It is usable but it might not be able to cope
-with all DICOM objects or DicomWeb inputs. See [Known limitations for details](#known-limitations).
+with all DICOM objects or DICOMweb records. See [Known limitations for details](#known-limitations).
 
 I'm still trying to figure out what is the most useful way of using this plugin. Please feel free to try it out,
 send feedback in [Discussions](https://github.com/realcundo/nu_plugin_dcm/discussions) or report problems in [Issues](https://github.com/realcundo/nu_plugin_dcm/issues).
 
 ## Usage
 `dcm` plugin reads its input from single values, or from list of values:
-- `dcm`: expects a string/filename, file record (must contain `name` and `type`), dicomweb record, or binary DICOM data
+- `dcm`: expects a string/filename, file record (must contain `name` and `type`), DICOMweb record, or binary DICOM data
   - `ls *.dcm | dcm`: process a list of files, resulting in a list of dicom records
   - `ls *.dcm | select name type | dcm`: process a list of files specified by their filename, resulting in a list of dicom records
   - `open --raw file.dcm | into binary | dcm`: process a binary stream, resulting in a dicom record
@@ -40,8 +40,8 @@ See Examples for more details.
   ```
 
   Without `into binary`, `dcm` would see a list of strings, assuming it's a list of filenames.
-- For dicomweb inputs, only the first of "Alphabetic", "Ideographic", "Phonetic" Patient Names is extracted.
-- For dicomweb inputs, `BulkDataURI` and `InlineBinary` are not extracted and `nothing` is returned as their values.
+- For DICOMweb inputs, only the first of "Alphabetic", "Ideographic", "Phonetic" Patient Names is extracted.
+- For DICOMweb inputs, `BulkDataURI` and `InlineBinary` are not extracted and `nothing` is returned as their values.
 
 
 ## Examples

src/dicomweb.rs

@@ -33,11 +33,16 @@ pub fn is_dicom_record(record: &Record) -> bool {
             // Check https://dicom.nema.org/medical/dicom/current/output/chtml/part18/sect_F.2.2.html
             if let Ok(potential_record) = potential_value.as_record() {
                 potential_tag.len() == 8
+                    // the tag must be a valid hex string
                     && potential_tag
                         .chars()
                         .all(|c| c.is_ascii_hexdigit())
+                    // must contain "vr"
                     && potential_record.contains("vr")
-                    && (potential_record.contains("Value") || potential_record.contains("BulkDataURI") || potential_record.contains("InlineBinary"))
+                    // tags must be one of the following
+                    && (potential_record
+                        .iter()
+                        .all(|(key, _value)| key == "Value" || key == "BulkDataURI" || key == "InlineBinary" || key == "vr"))
             } else {
                 false
             }
@@ -143,8 +148,8 @@ impl DicomWebDump<'_, '_> {
         let value = match record.get("Value") {
             Some(value) => value,
             None => {
-                // don't return error if BulkDataURI/InlineBinary exist
-                if record.contains("BulkDataURI") || record.contains("InlineBinary") {
+                // don't return error if BulkDataURI/InlineBinary exist or if no other fields are present (just having "vr" with nothing else is valid)
+                if record.len() == 1 || record.contains("BulkDataURI") || record.contains("InlineBinary") {
                     return Ok(Value::nothing(record_span));
                 } else {
                     return Err(DicomWebError::MissingRequiredColumn { column: "Value|BulkDataURI|InlineBinary", span: record_span });

src/plugin.rs

@@ -258,8 +258,8 @@ impl DcmPluginCommand {
                 }
 
                 // Output generic error
-                Err(LabeledError::new("Cannot process records directly, unless they are DicomWeb records")
-                    .with_label("For files, select file name or binary data from the record before passing it to dcm", *internal_span))
+                Err(LabeledError::new("Cannot process records directly, unless they are Fila or DicomWeb records")
+                    .with_label("For files, select file name, binary data, or use records with `name` and `type`", *internal_span))
             }
             Value::Binary { val, internal_span, .. } => {
                 let cursor = Cursor::new(val);

tests/assets/dicomweb-example.json

@@ -119,8 +119,7 @@
       "Value": ["ISO_IR 192"]
     },
     "00080020": {
-      "vr": "DT",
-      "Value": ["20130309"]
+      "vr": "DT"
     },
     "00080030": {
       "vr": "TM",
@@ -136,7 +135,7 @@
     },
     "00080061": {
       "vr": "CS",
-      "Value": ["CT", "PET"]
+      "Value": ["CT", "MG"]
     },
     "00080090": {
       "vr": "PN",

tests/dicomweb.rs

@@ -1,5 +1,5 @@
 use nu_protocol::Span;
-use test_utils::{get_string_by_cell_path, setup_plugin_for_test};
+use test_utils::{assert_nothing_by_cell_path, get_string_by_cell_path, get_string_list_by_cell_path, setup_plugin_for_test};
 
 mod test_utils;
 
@@ -12,9 +12,19 @@ fn read_dicomweb_list() -> Result<(), nu_protocol::ShellError> {
     let result = plugin_test.eval("open dicomweb-example.json | dcm")?;
     let result = result.into_value(TEST_SPAN)?;
 
+    assert_eq!(
+        result
+            .as_list()
+            .unwrap()
+            .len(),
+        2
+    );
+
     {
         assert_eq!(get_string_by_cell_path(&result, "0.StudyInstanceUID"), "1.2.392.200036.9116.2.2.2.1762893313.1029997326.945873");
+        assert_eq!(get_string_list_by_cell_path(&result, "0.ModalitiesInStudy"), vec!["CT".to_string(), "PET".to_string()]);
         assert_eq!(get_string_by_cell_path(&result, "0.PatientName"), "Wang^XiaoDong");
+        assert_eq!(get_string_by_cell_path(&result, "0.StudyDate"), "20130409");
         assert_eq!(get_string_by_cell_path(&result, "0.OtherPatientIDsSequence.0.PatientID"), "54321");
         assert_eq!(get_string_by_cell_path(&result, "0.OtherPatientIDsSequence.0.IssuerOfPatientID"), "Hospital B");
         assert_eq!(get_string_by_cell_path(&result, "0.OtherPatientIDsSequence.1.PatientID"), "24680");
@@ -23,7 +33,9 @@ fn read_dicomweb_list() -> Result<(), nu_protocol::ShellError> {
 
     {
         assert_eq!(get_string_by_cell_path(&result, "1.StudyInstanceUID"), "1.2.392.200036.9116.2.2.2.2162893313.1029997326.945876");
+        assert_eq!(get_string_list_by_cell_path(&result, "1.ModalitiesInStudy"), vec!["CT".to_string(), "MG".to_string()]);
         assert_eq!(get_string_by_cell_path(&result, "1.PatientName"), "Wang^XiaoDong");
+        assert_nothing_by_cell_path(&result, "1.StudyDate");
         assert_eq!(get_string_by_cell_path(&result, "1.OtherPatientIDsSequence.0.PatientID"), "54321");
         assert_eq!(get_string_by_cell_path(&result, "1.OtherPatientIDsSequence.0.IssuerOfPatientID"), "Hospital B2");
         assert_eq!(get_string_by_cell_path(&result, "1.OtherPatientIDsSequence.1.PatientID"), "24680");

tests/test_utils.rs

@@ -88,6 +88,26 @@ pub fn get_string_by_cell_path(
         .to_owned()
 }
 
+/// Asserts that the value at `path` is a list of strings and returns it. Panics on failure.
+#[allow(dead_code)]
+pub fn get_string_list_by_cell_path(
+    value: &Value,
+    path: &str,
+) -> Vec<String> {
+    let result_value = get_value_by_cell_path(value, path);
+    result_value
+        .as_list()
+        .and_then(|list| {
+            list.iter()
+                .map(|v| {
+                    v.as_str()
+                        .map(|s| s.to_string())
+                })
+                .collect::<Result<Vec<_>, _>>()
+        })
+        .unwrap_or_else(|e| panic!("Expected list<string> at path '{}', but found '{}'. Error: {}", path, result_value.get_type(), e))
+}
+
 /// Asserts that the value at `path` is an int and returns it. Panics on failure.
 #[allow(dead_code)]
 pub fn get_int_by_cell_path(