From c6ec14da3ea90349144fd9d65f06a50ac16b8215 Mon Sep 17 00:00:00 2001
From: OneWay <info@oneway.asia>
Date: Sun, 11 Oct 2026 22:25:23 +0800
Subject: [PATCH] fix(indexing): keep parsing progress moving

The parse phase estimated progress from a line-count formula that hit
its ceiling early, so the percentage froze on large dictionaries.
Estimate it from the position in the decompressed stream against the
uncompressed size instead.
---
 .../research/data/parser/DslIndexer.kt        | 12 ++++++---
 .../data/repository/KotlinDictionaryEngine.kt | 27 +++++++++++++++++++
 2 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/app/src/main/java/com/example/research/data/parser/DslIndexer.kt b/app/src/main/java/com/example/research/data/parser/DslIndexer.kt
index d93f318..ea4de13 100644
--- a/app/src/main/java/com/example/research/data/parser/DslIndexer.kt
+++ b/app/src/main/java/com/example/research/data/parser/DslIndexer.kt
@@ -49,6 +49,7 @@ class DslIndexer(context: Context? = null) {
     suspend fun createIndex(
         dslInputStream: java.io.InputStream,
         indexFile: File,
+        totalInputBytes: Long = -1L,
         onProgress: (Int) -> Unit = {},
         onOperationChange: (String) -> Unit = {},
         onFileProgress: (Float) -> Unit = {},
@@ -65,7 +66,7 @@ class DslIndexer(context: Context? = null) {
             try { foldedIndexFile(indexFile).delete() } catch (_: Throwable) { /* Legacy v5 leftover */ }
             onOperationChange("Parsing file")
             onFileProgress(0.0f)
-            counts = parseDslToTempFile(bufferedStream, tempFile,
+            counts = parseDslToTempFile(bufferedStream, tempFile, totalInputBytes,
                 onProgress, onFileProgress, onCharsetDetected)
             onFileProgress(0.5f)
             onOperationChange("Sorting index")
@@ -88,6 +89,7 @@ class DslIndexer(context: Context? = null) {
     private suspend fun parseDslToTempFile(
         inputStream: java.io.InputStream,
         tempFile: File,
+        totalInputBytes: Long,
         onProgress: (Int) -> Unit,
         onFileProgress: (Float) -> Unit = {},
         onCharsetDetected: (DslCharsetDetector.DetectedCharset) -> Unit = {}
@@ -230,8 +232,12 @@ class DslIndexer(context: Context? = null) {
                     val yieldThreshold = adaptiveConfig.yieldInterval
                     if (yieldCounter >= yieldThreshold) {
                         onProgress(entryCount)
-                        val estimatedProgress = (lineCount * 10).toFloat() / (Int.MAX_VALUE / 1000)
-                        onFileProgress(estimatedProgress.coerceIn(0f, 0.4f))
+                        val parseProgress = if (totalInputBytes > 0L) {
+                            (reader.position.toFloat() / totalInputBytes.toFloat())
+                        } else {
+                            (lineCount * 10).toFloat() / (Int.MAX_VALUE / 1000)
+                        }
+                        onFileProgress(parseProgress.coerceIn(0f, 0.4f))
 
                         if (resourceManager != null) {
                             if (resourceManager.shouldYield(adaptiveConfig)) {
diff --git a/app/src/main/java/com/example/research/data/repository/KotlinDictionaryEngine.kt b/app/src/main/java/com/example/research/data/repository/KotlinDictionaryEngine.kt
index c8b1a25..c4e8b8c 100644
--- a/app/src/main/java/com/example/research/data/repository/KotlinDictionaryEngine.kt
+++ b/app/src/main/java/com/example/research/data/repository/KotlinDictionaryEngine.kt
@@ -22,6 +22,7 @@ import java.io.BufferedInputStream
 import java.io.File
 import java.io.FileInputStream
 import java.io.InputStream
+import java.io.RandomAccessFile
 import java.util.concurrent.CancellationException
 import java.util.zip.GZIPInputStream
 
@@ -51,12 +52,14 @@ class KotlinDictionaryEngine(private val context: Context) : ProgressIndexingEng
 
             val indexFile = File(indexPath)
             val metadata = readMetadata(dictPath)
+            val totalInputBytes = uncompressedInputLength(dictFile)
 
             val articleCount = withContext(Dispatchers.IO) {
                 openDslInputStream(dictPath).use { inputStream ->
                     dslIndexer.createIndex(
                         dslInputStream = inputStream,
                         indexFile = indexFile,
+                        totalInputBytes = totalInputBytes,
                         onProgress = { count ->
                             onArticlesIndexed(count)
                         },
@@ -207,6 +210,30 @@ class KotlinDictionaryEngine(private val context: Context) : ProgressIndexingEng
         }
     }
 
+    private fun uncompressedInputLength(file: File): Long {
+        val name = file.name.lowercase()
+        if (!name.endsWith(".dz") && !name.endsWith(".gz")) return file.length()
+        return readGzipUncompressedLength(file).takeIf { it > 0L } ?: file.length()
+    }
+
+    private fun readGzipUncompressedLength(file: File): Long = try {
+        RandomAccessFile(file, "r").use { raf ->
+            val fileLength = raf.length()
+            if (fileLength < 18L) {
+                -1L
+            } else {
+                raf.seek(fileLength - 4L)
+                val b0 = raf.readUnsignedByte()
+                val b1 = raf.readUnsignedByte()
+                val b2 = raf.readUnsignedByte()
+                val b3 = raf.readUnsignedByte()
+                (b0 or (b1 shl 8) or (b2 shl 16) or (b3 shl 24)).toLong() and 0xFFFFFFFFL
+            }
+        }
+    } catch (_: Exception) {
+        -1L
+    }
+
     private fun resolveIndexPath(indexArg: String): String {
         val candidate = File(indexArg)
         if (candidate.exists()) return candidate.absolutePath