TEI: also write pos in PC

INL · Oct 17, 2024 · 3ad1ace · 3ad1ace
1 parent d65a8c4
commit 3ad1ace
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 8 deletions.
diff --git a/server/src/main/kotlin/org/ivdnt/galahad/port/tei/export/LayerToTEIConverter.kt b/server/src/main/kotlin/org/ivdnt/galahad/port/tei/export/LayerToTEIConverter.kt
@@ -111,7 +111,8 @@ class LayerToTEIConverter(
             val alphaNumeric = Regex("""[a-zA-Z0-9]""")
             if (!term.literals.contains(alphaNumeric)) {
                 // Interpret as punctuation only if it doesn't contain any alphanumeric characters
-                writer.writeRaw("<pc xml:id=\"${term.targets[0].id}\">${getLiteral()}</pc>")
+                val pos = term.posOrEmpty.escapeXML()
+                writer.writeRaw("<pc pos=\"$pos\" xml:id=\"${term.targets[0].id}\">${getLiteral()}</pc>")
             } else {
                 // Clear the pos and interpret as <w>
                 val lemma = term.lemmaOrEmpty.escapeXML()

diff --git a/server/src/main/kotlin/org/ivdnt/galahad/port/tei/export/TEITextMerger.kt b/server/src/main/kotlin/org/ivdnt/galahad/port/tei/export/TEITextMerger.kt
@@ -7,8 +7,6 @@ import org.ivdnt.galahad.data.layer.WordForm
 import org.ivdnt.galahad.evaluation.comparison.LayerComparison.Companion.truncatedPcMatch
 import org.ivdnt.galahad.port.folia.export.deepcopy
 import org.ivdnt.galahad.port.xml.getPlainTextContent
-import org.ivdnt.galahad.taggers.TaggerStore
-import org.ivdnt.galahad.tagset.TagsetStore
 import org.ivdnt.galahad.util.*
 import org.w3c.dom.Document
 import org.w3c.dom.Element
@@ -235,12 +233,12 @@ open class TEITextMerger(
         val termToAdd = layer.termForWordForm(wf)
 
         val wTag = if (layer.tagset.punctuationTags.contains(termToAdd.pos) && !termToAdd.literals.contains(alphaNumeric)) {
-            val n = document.createElement("pc")
-            n
+            val element = document.createElement("pc")
+            element
         } else {
-            val n = document.createElement("w")
-            n.setAttribute("lemma", termToAdd.lemmaOrEmpty)
-            n
+            val element = document.createElement("w")
+            element.setAttribute("lemma", termToAdd.lemmaOrEmpty)
+            element
         }
 
         // Empty pos if it is a PC and it contains alphanumeric characters (so it can't be PC anyway).