Epub tts fix (#103)

* build: bump version to 1.0.38 and improve CFI page lookup logic

* fix: resolve duplicate TTS audio and broken highlights in Vertical Mode

Updated the text extraction logic in epub_reader.js to pick only the most granular matching elements, preventing duplication when elements like <p> are nested inside <li>.

app/build.gradle.kts

@@ -26,8 +26,8 @@ android {
         applicationId = "com.aryan.reader"
         minSdk = 26
         targetSdk = 35
-        versionCode = 38
+        versionCode = 39
-        versionName = "1.0.37"
+        versionName = "1.0.38"
 
         testInstrumentationRunner = "androidx.test.runner.AndroidJUnitRunner"
         externalNativeBuild {

app/src/main/assets/epub_reader.js

@@ -889,6 +889,8 @@
         }
 
         , Text='${textToHighlight.substring(0, 50)}...' `);
+        console.log("TTS_LIST_DIAG: Attempting highlight. CFI: " + cfi + " Text: '" + textToHighlight.substring(0, 20) + "' Offset: " + startOffset);
+
         window.removeHighlight();
 
         if (!cfi || !textToHighlight) {
@@ -1098,12 +1100,15 @@
     window.extractTextWithCfiFromTop = function () {
         try {
             const ttsNodeSelector = "p, h1, h2, h3, h4, h5, h6, li, blockquote";
-            const allContentNodes = Array.from(document.body.querySelectorAll(ttsNodeSelector));
+            const allContentNodesRaw = Array.from(document.body.querySelectorAll(ttsNodeSelector));
+
+            // FIX: Filter out parents that contain matching children to prevent duplicates
+            const allContentNodes = allContentNodesRaw.filter(node => node.querySelector(ttsNodeSelector) === null);
 
             let startBlock = null;
             let startIndex = -1;
 
-            for (let i = 0; i < allContentNodes.size || i < allContentNodes.length; i++) {
+            for (let i = 0; i < allContentNodes.length; i++) {
                 const node = allContentNodes[i];
                 const rect = node.getBoundingClientRect();
 
@@ -1141,7 +1146,10 @@
 
     window.extractTextWithCfi = function () {
         const results =[];
-        const contentNodes = document.body.querySelectorAll("p, h1, h2, h3, h4, h5, h6, li, blockquote");
+        const ttsNodeSelector = "p, h1, h2, h3, h4, h5, h6, li, blockquote";
+        const contentNodesRaw = Array.from(document.body.querySelectorAll(ttsNodeSelector));
+
+        const contentNodes = contentNodesRaw.filter(node => node.querySelector(ttsNodeSelector) === null);
 
         contentNodes.forEach((node) => {
             const text = node.innerText ? node.innerText.trim() : "";
@@ -1156,8 +1164,7 @@
                 } catch (e) {}
             }
         });
-        const jsonResult = JSON.stringify(results);
+        return JSON.stringify(results);
-        return jsonResult;
     };
 
     window.extractTextWithCfiFromSelection = function() {
@@ -1186,8 +1193,14 @@
                 absoluteStartOffset += startOffset;
             }
 
-            const allContentNodes = Array.from(document.body.querySelectorAll(ttsNodeSelector));
+            const allContentNodesRaw = Array.from(document.body.querySelectorAll(ttsNodeSelector));
-            const startIndex = allContentNodes.findIndex(node => node === startBlock);
+            const allContentNodes = allContentNodesRaw.filter(node => node.querySelector(ttsNodeSelector) === null);
+
+            let startIndex = allContentNodes.findIndex(node => node === startBlock);
+
+            if (startIndex === -1) {
+                startIndex = allContentNodes.findIndex(node => startBlock.contains(node));
+            }
 
             if (startIndex === -1) return window.extractTextWithCfiFromTop();
 
@@ -1197,11 +1210,13 @@
             nodesToProcess.forEach((node, index) => {
                 let fullText = node.textContent || "";
                 if (fullText.trim().length > 0 && node.offsetParent !== null) {
+                    if (node.tagName === 'LI' || node.closest('li')) {
+                        console.log("TTS_LIST_DIAG: Selection extracting node <" + node.tagName + "> inside LI. Text: '" + fullText.substring(0, 30) + "'");
+                    }
                     try {
                         const cfiObj = getCfiPathForElement(node, 0);
                         if (cfiObj && cfiObj.cfi) {
                             if (index === 0) {
-                                // Slice the very first block strictly from the selected character
                                 let sliced = fullText.substring(absoluteStartOffset);
                                 if (sliced.trim().length > 0) {
                                     results.push({

app/src/main/java/com/aryan/reader/epubreader/ChapterWebView.kt

@@ -131,6 +131,7 @@ class TtsJsBridge(
 ) {
     @JavascriptInterface
     fun onStructuredTextExtracted(json: String) {
+        Timber.tag("TTS_LIST_DIAG").d("Bridge received JSON: $json")
         if (json.isNotBlank() && json != "[]") {
             scope.launch {
                 ttsStructuredTextHandler(json)
@@ -545,6 +546,10 @@ fun ChapterWebView(
                                             .d("JS -> ${message.substringAfter("BookmarkDiagnosis: ")}")
                                     }
 
+                                    message.startsWith("TTS_LIST_DIAG:") -> {
+                                        Timber.tag("TTS_LIST_DIAG").d("JS -> ${message.substringAfter("TTS_LIST_DIAG: ")}")
+                                    }
+
                                     message.startsWith("CFI_DIAGNOSIS:") -> {
                                         Timber.d(
                                             "JS -> ${message.substringAfter("CFI_DIAGNOSIS: ")}"

app/src/main/java/com/aryan/reader/epubreader/EpubReaderScreen.kt

@@ -2447,21 +2447,17 @@ fun EpubReaderHost(
                                             ttsScope = scope,
                                             onTtsTextReady = { jsonString ->
                                                 scope.launch {
-                                                    Timber.d("Vertical: onTtsTextReady received JSON. Length: ${jsonString.length}")
+                                                    Timber.tag("TTS_LIST_DIAG").d("Vertical: Processing received JSON. Length: ${jsonString.length}") // Add this
-                                                    val ttsChunks =
+                                                    val ttsChunks = mutableListOf<TtsChunk>()
-                                                        mutableListOf<TtsChunk>()
                                                     try {
                                                         val jsonArray = JSONArray(jsonString)
-                                                        Timber.d("Vertical: Parsed JSON Array. Items: ${jsonArray.length()}")
                                                         for (i in 0 until jsonArray.length()) {
-                                                            val jsonObject =
+                                                            val jsonObject = jsonArray.getJSONObject(i)
-                                                                jsonArray.getJSONObject(i)
                                                             val text = jsonObject.getString("text")
-                                                            val cfiJsonString =
+                                                            val cfiJsonObject = JSONObject(jsonObject.getString("cfi"))
-                                                                jsonObject.getString("cfi")
-                                                            val cfiJsonObject =
-                                                                JSONObject(cfiJsonString)
                                                             val cfi = cfiJsonObject.getString("cfi")
+
+                                                            Timber.tag("TTS_LIST_DIAG").d("Processing Chunk[$i]: text='${text.take(40)}...' cfi='$cfi'")
                                                             val baseOffset = jsonObject.optInt("startOffset", 0)
 
                                                             val subChunks =

app/src/main/java/com/aryan/reader/paginatedreader/BookPaginator.kt

@@ -1041,17 +1041,21 @@ class BookPaginator(
     override fun findPageForCfiAndOffset(chapterIndex: Int, cfi: String, charOffset: Int): Int? {
         val index = chapterCharacterIndex[chapterIndex]
         if (index.isNullOrEmpty()) {
+            Timber.tag("TTS_PAGE_JUMP_DIAG").w("Lookup failed: No character index for chapter $chapterIndex")
             return null
         }
 
         val targetPath = CfiUtils.getPath(cfi)
-        val foundRange = index.find { range ->
+
+        val matches = index.filter { range ->
             val rangePath = CfiUtils.getPath(range.cfi)
-            val cfiMatches = targetPath == rangePath || targetPath.startsWith(rangePath) || rangePath.startsWith(targetPath)
+            val pathMatches = targetPath == rangePath || targetPath.startsWith(rangePath)
             val offsetMatches = charOffset >= range.startOffset && charOffset < range.endOffset
-            cfiMatches && offsetMatches
+            pathMatches && offsetMatches
         }
 
+        val foundRange = matches.maxByOrNull { CfiUtils.getPath(it.cfi).length }
+
         return if (foundRange != null) {
             val chapterStartPage = chapterStartPageIndices[chapterIndex]
             if (chapterStartPage == null) {