Test versification mismatch

sillsdev · Aug 30, 2024 · d8be703 · d8be703
1 parent 7a819ed
commit d8be703
Showing 1 changed file with 73 additions and 0 deletions.
diff --git a/tests/SIL.Machine.Tests/Corpora/UsfmManualTests.cs b/tests/SIL.Machine.Tests/Corpora/UsfmManualTests.cs
@@ -1,4 +1,5 @@
 using System.IO.Compression;
+using System.Text;
 using System.Text.Json;
 using NUnit.Framework;
 
@@ -169,4 +170,76 @@ async Task GetUsfmAsync(string projectPath)
             await GetUsfmAsync(ParatextProjectPath);
         }
     }
+
+    [Test]
+    public void Test()
+    {
+        var sourceCorpus = new ParatextTextCorpus(CorporaTestHelpers.UsfmSourceProjectPath);
+        var targetCorpus = new ParatextTextCorpus(CorporaTestHelpers.UsfmTargetProjectPath);
+
+        var rows = AlignPretranslateCorpus(sourceCorpus.FilterTexts(["SUS"]), targetCorpus.FilterTexts(["SUS"]))
+            .Select(p =>
+                (
+                    Refs: (IReadOnlyList<ScriptureRef>)
+                        p.Refs.Select(r => ScriptureRef.Parse(r, targetCorpus.Versification)).ToArray(),
+                    p.Translation
+                )
+            )
+            .OrderBy(p => p.Refs[0]);
+
+        var updater = new FileParatextProjectTextUpdater(CorporaTestHelpers.UsfmSourceProjectPath);
+        string newUsfm = updater.UpdateUsfm("SUS", rows.ToArray(), stripAllText: true, preferExistingText: true);
+        Assert.That(
+            newUsfm,
+            Contains.Substring(
+                "\\v 65 et rex Astyages adpositus est ad patres suos et suscepit Cyrus Perses regnum eius"
+            )
+        );
+    }
+
+    private static IEnumerable<(IReadOnlyList<string> Refs, string Translation)> AlignPretranslateCorpus(
+        ITextCorpus srcCorpus,
+        ITextCorpus trgCorpus
+    )
+    {
+        int rowCount = 0;
+        StringBuilder srcSegBuffer = new();
+        StringBuilder trgSegBuffer = new();
+        List<ScriptureRef> refs = [];
+        foreach (ParallelTextRow row in srcCorpus.AlignRows(trgCorpus, allSourceRows: true))
+        {
+            if (!row.IsTargetRangeStart && row.IsTargetInRange)
+            {
+                refs.AddRange(row.TargetRefs.Cast<ScriptureRef>());
+                if (row.SourceText.Length > 0)
+                {
+                    if (srcSegBuffer.Length > 0)
+                        srcSegBuffer.Append(' ');
+                    srcSegBuffer.Append(row.SourceText);
+                }
+                rowCount++;
+            }
+            else
+            {
+                if (rowCount > 0)
+                {
+                    yield return (refs.Select(r => r.ToString()).ToArray(), srcSegBuffer.ToString());
+                    srcSegBuffer.Clear();
+                    trgSegBuffer.Clear();
+                    refs.Clear();
+                    rowCount = 0;
+                }
+
+                refs.AddRange(row.TargetRefs.Cast<ScriptureRef>());
+                srcSegBuffer.Append(row.SourceText);
+                trgSegBuffer.Append(row.TargetText);
+                rowCount++;
+            }
+        }
+
+        if (rowCount > 0)
+        {
+            yield return (refs.Select(r => r.ToString()).ToArray(), srcSegBuffer.ToString());
+        }
+    }
 }