Merge pull request #843 from bids-standard/revise-subjectmetadata

change subjectMetadata to array
bids-standard · Sep 24, 2019 · 2c0a610 · 2c0a610
2 parents f552be1 + 4ca3000
commit 2c0a610
Show file tree

Hide file tree

Showing 3 changed files with 53 additions and 37 deletions.
diff --git a/bids-validator/tests/bids.spec.js b/bids-validator/tests/bids.spec.js
@@ -117,12 +117,13 @@ describe('BIDS example datasets ', function() {
       var warnings = issues.warnings
       assert(summary.sessions.length === 0)
       assert(summary.subjects.length === 1)
-      assert.deepEqual(summary.subjectMetadata, {
-        '01': {
-          sex: 'M',
+      assert.deepEqual(summary.subjectMetadata, [
+        {
+          participantId: '01',
           age: 25,
+          sex: 'M',
         },
-      })
+      ])
       assert.deepEqual(summary.tasks, ['rhyme judgment'])
       assert.isFalse(summary.dataProcessed)
       assert(summary.modalities.includes('T1w'))

diff --git a/bids-validator/utils/__tests__/collectSubjectMetadata.spec.js b/bids-validator/utils/__tests__/collectSubjectMetadata.spec.js
@@ -8,15 +8,11 @@ sub-01	34	F
 sub-02	38	M
 `
     const subjectMetadata = collectSubjectMetadata(tsvFile)
-    assert.deepEqual(subjectMetadata, {
-      '01': {
-        age: 34,
-        sex: 'F',
-      },
-      '02': {
-        age: 38,
-        sex: 'M',
-      },
+    assert.lengthOf(subjectMetadata, 2)
+    assert.deepEqual(subjectMetadata[0], {
+      participantId: '01',
+      age: 34,
+      sex: 'F',
     })
   })
   it('extracts tsv string to subjectMetadata object', () => {

diff --git a/bids-validator/utils/summary/collectSubjectMetadata.js b/bids-validator/utils/summary/collectSubjectMetadata.js
@@ -1,16 +1,20 @@
+const PARTICIPANT_ID = 'participantId'
+const AGE = 'age'
 /**
- * Go from tsv format string with participant_id as a required header to object of form
- * {
- *   participant_id_1: {
+ * Go from tsv format string with participant_id as a required header to array of form
+ * [
+ *   {
+ *     participantId: 'participant_id_1'
  *     foo: 'x',
  *     ...
  *   },
- *   participant_id_2: {
+ *   {
+ *     participantId: 'participant_id_2'
  *     foo: 'y',
  *     ...
  *   }
  *   ...
- * }
+ * ]
  *
  * returns null if participant_id is not a header or file contents do not exist
  * @param {string} participantsTsvContent
@@ -21,29 +25,44 @@ const collectSubjectMetadata = participantsTsvContent => {
       .split('\n')
       .filter(row => row !== '')
       .map(row => row.split('\t'))
-    const [headers, ...subjectData] = contentTable
-    const participant_idIndex = headers.findIndex(
-      header => header === 'participant_id',
+    const [snakeCaseHeaders, ...subjectData] = contentTable
+    const headers = snakeCaseHeaders.map(header =>
+      header === 'participant_id' ? PARTICIPANT_ID : header,
     )
-    if (participant_idIndex === -1) return null
+    const targetKeys = [PARTICIPANT_ID, 'age', 'sex', 'group']
+      .map(key => ({
+        key,
+        index: headers.findIndex(targetKey => targetKey === key),
+      }))
+      .filter(({ index }) => index !== -1)
+    const participantIdKey = targetKeys.find(
+      ({ key }) => key === PARTICIPANT_ID,
+    )
+    const ageKey = targetKeys.find(({ key }) => key === AGE)
+    if (participantIdKey === undefined) return null
     else
-      return subjectData.reduce(
-        (subjectMetadata, data) => ({
-          ...subjectMetadata,
-          [data[participant_idIndex].replace(/^sub-/, '')]: data.reduce(
-            (subjectMetadata, datum, i) =>
-              i === participant_idIndex
-                ? subjectMetadata
-                : {
-                    ...subjectMetadata,
-                    [headers[i]]:
-                      headers[i] === 'age' ? parseInt(datum) : datum,
-                  },
+      return subjectData
+        .map(data => {
+          // this first map is for transforming any data coming out of participants.tsv:
+          // strip subject ids to match metadata.subjects: 'sub-01' -> '01'
+          data[participantIdKey.index] = data[participantIdKey.index].replace(
+            /^sub-/,
+            '',
+          )
+          // make age an integer
+          if (ageKey) data[ageKey.index] = parseInt(data[ageKey.index])
+          return data
+        })
+        .map(data =>
+          //extract all target metadata for each subject
+          targetKeys.reduce(
+            (subject, { key, index }) => ({
+              ...subject,
+              [key]: data[index],
+            }),
             {},
           ),
-        }),
-        {},
-      )
+        )
   }
 }