epub-parser: find siblings from parent's sibling node aswell (#151)

Currently, the EPUB parser only finds siblings located inside its own parent, which may result in an empty or incomplete chapter body if some of the sibling nodes are located within the siblings of its parent node instead of being direct siblings. --------------------------- Signed-off-by: starry-shivam <[email protected]>
Pool-Of-Tears · Apr 16, 2024 · 51a4475 · 51a4475
1 parent 4b7d0f1
commit 51a4475
Show file tree

Hide file tree

Showing 3 changed files with 52 additions and 8 deletions.
diff --git a/.idea/misc.xml b/.idea/misc.xml
diff --git a/app/src/main/java/com/starry/myne/epub/EpubUtils.kt b/app/src/main/java/com/starry/myne/epub/EpubUtils.kt
@@ -20,6 +20,7 @@ package com.starry.myne.epub
 import org.w3c.dom.Document
 import org.w3c.dom.Element
 import org.w3c.dom.Node
+import org.jsoup.nodes.Node as JsoupNode
 import org.w3c.dom.NodeList
 import org.xml.sax.InputSource
 import java.io.File
@@ -58,3 +59,13 @@ fun Node.getAttributeValue(attribute: String): String? =
 
 val NodeList.elements get() = (0..length).asSequence().mapNotNull { item(it) as? Element }
 val Node.childElements get() = childNodes.elements
+
+fun JsoupNode.nextSiblingNodes(): List<org.jsoup.nodes.Node> {
+    val siblings = mutableListOf<org.jsoup.nodes.Node>()
+    var nextSibling = nextSibling()
+    while (nextSibling != null) {
+        siblings.add(nextSibling)
+        nextSibling = nextSibling.nextSibling()
+    }
+    return siblings
+}
diff --git a/app/src/main/java/com/starry/myne/epub/EpubXMLFileParser.kt b/app/src/main/java/com/starry/myne/epub/EpubXMLFileParser.kt
@@ -22,6 +22,7 @@ import android.util.Log
 import org.jsoup.Jsoup
 import org.jsoup.nodes.TextNode
 import java.io.File
+import org.jsoup.nodes.Node
 import kotlin.io.path.invariantSeparatorsPathString
 
 /**
@@ -86,7 +87,7 @@ class EpubXMLFileParser(
                 val fragmentElement = document.selectFirst("#$fragmentId")
                 title = fragmentElement?.selectFirst("h1, h2, h3, h4, h5, h6")?.text() ?: ""
                 val bodyBuilder = StringBuilder()
-                var currentNode: org.jsoup.nodes.Node? = fragmentElement?.nextSibling()
+                var currentNode: Node? = fragmentElement?.nextSibling()
                 val nextFragmentIdElement = if (nextFragmentId != null) {
                     document.selectFirst("#$nextFragmentId")
                 } else {
@@ -96,7 +97,7 @@ class EpubXMLFileParser(
 
                 while (currentNode != null && currentNode != nextFragmentIdElement) {
                     bodyBuilder.append(getNodeStructuredText(currentNode) + "\n\n")
-                    currentNode = currentNode.nextSibling()
+                    currentNode = getNextSibling(currentNode)
                 }
                 bodyContent = bodyBuilder.toString()
             }
@@ -115,6 +116,39 @@ class EpubXMLFileParser(
         )
     }
 
+    private fun getNextSibling(currentNode: Node?): Node? {
+        var nextSibling: Node? = currentNode?.nextSibling()
+
+        if (nextSibling == null) {
+            var parentNode = currentNode?.parent()
+            while (parentNode != null) {
+                nextSibling = parentNode.nextSibling()
+                if (nextSibling != null) {
+                    // If the parent's next sibling is not null, traverse its descendants
+                    // to find the next node
+                    return traverseDescendants(nextSibling)
+                }
+                parentNode = parentNode.parent()
+            }
+        }
+
+        return nextSibling
+    }
+
+    private fun traverseDescendants(node: Node): Node? {
+        val children = node.childNodes()
+        if (children.isNotEmpty()) {
+            return children.first()
+        }
+
+        val siblings = node.nextSiblingNodes()
+        if (siblings.isNotEmpty()) {
+            return traverseDescendants(siblings.first())
+        }
+
+        return null
+    }
+
     fun parseAsImage(absolutePathImage: String): String {
         // Use run catching so it can be run locally without crash
         val bitmap = zipFile[absolutePathImage]?.data?.runCatching {
@@ -130,7 +164,7 @@ class EpubXMLFileParser(
     }
 
     // Rewrites the image node to xml for the next stage.
-    private fun declareImgEntry(node: org.jsoup.nodes.Node): String {
+    private fun declareImgEntry(node: Node): String {
         val attrs = node.attributes().associate { it.key to it.value }
         val relPathEncoded = attrs["src"] ?: attrs["xlink:href"] ?: ""
 
@@ -143,8 +177,8 @@ class EpubXMLFileParser(
         return parseAsImage(absolutePathImage)
     }
 
-    private fun getPTraverse(node: org.jsoup.nodes.Node): String {
-        fun innerTraverse(node: org.jsoup.nodes.Node): String =
+    private fun getPTraverse(node: Node): String {
+        fun innerTraverse(node: Node): String =
             node.childNodes().joinToString("") { child ->
                 when {
                     child.nodeName() == "br" -> "\n"
@@ -159,7 +193,7 @@ class EpubXMLFileParser(
         return if (paragraph.isNotEmpty()) "$paragraph\n\n" else ""
     }
 
-    private fun getNodeTextTraverse(node: org.jsoup.nodes.Node): String {
+    private fun getNodeTextTraverse(node: Node): String {
         val children = node.childNodes()
         if (children.isEmpty())
             return ""
@@ -181,7 +215,7 @@ class EpubXMLFileParser(
         }
     }
 
-    private fun getNodeStructuredText(node: org.jsoup.nodes.Node): String {
+    private fun getNodeStructuredText(node: Node): String {
         val children = node.childNodes()
         if (children.isEmpty())
             return ""