book-reader/app/src/main/java/com/aryan/reader/epub/EpubParser.kt

/*
 * Episteme Reader - A native Android document reader.
 * Copyright (C) 2026 Episteme
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Affero General Public License as
 * published by the Free Software Foundation, either version 3 of the
 * License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU Affero General Public License for more details.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program.  If not, see <https://www.gnu.org/licenses/>.
 *
 * mail: epistemereader@gmail.com
 */
package com.aryan.reader.epub

import android.content.Context
import android.graphics.Bitmap
import android.graphics.BitmapFactory
import timber.log.Timber
import kotlinx.coroutines.Dispatchers
import kotlinx.coroutines.withContext
import kotlinx.serialization.Serializable
import kotlinx.serialization.decodeFromString
import kotlinx.serialization.encodeToString
import kotlinx.serialization.json.Json
import org.jsoup.Jsoup
import org.w3c.dom.Element
import org.w3c.dom.Node
import java.io.File
import java.io.FileOutputStream
import java.io.InputStream
import java.io.ByteArrayOutputStream
import java.net.URLDecoder
import java.nio.file.Paths
import java.util.UUID
import java.util.zip.ZipEntry
import java.util.zip.ZipFile
import kotlinx.coroutines.async
import kotlinx.coroutines.awaitAll
import kotlinx.coroutines.sync.Semaphore
import kotlinx.coroutines.sync.withPermit

class EpubParser(private val context: Context) {
    private val jsonSerializer = Json { ignoreUnknownKeys = true; encodeDefaults = true }

    data class EpubDocument(
        val metadata: Node, val manifest: Node, val spine: Node, val opfFilePath: String
    )

    data class EpubManifestItem(
        val id: String, val absPath: String, val mediaType: String, val properties: String
    )

    data class TempEpubChapter(
        val url: String,
        val title: String?,
        val htmlFilePath: String,
        val chapterIndex: Int,
        val plainTextContent: String,
        val htmlContent: String,
        val depth: Int,
        val isInToc: Boolean
    )

    // Helper class for NCX parsing results
    data class NcxMetadata(
        val title: String,
        val depth: Int
    )

    @Serializable
    private data class EpubExtractionCacheManifest(
        val bookId: String,
        val originalBookNameHint: String,
        val parserVersion: Int,
        val parseContent: Boolean,
        val shouldUseToc: Boolean,
        val sourceFingerprint: String? = null
    )

    // EpubFile can still represent in-memory file data during initial parsing before extraction
    data class EpubFile(val absPath: String, val data: ByteArray) {
        override fun equals(other: Any?): Boolean {
            if (this === other) return true
            if (javaClass != other?.javaClass) return false
            other as EpubFile
            if (absPath != other.absPath) return false
            return data.contentEquals(other.data)
        }
        override fun hashCode(): Int {
            var result = absPath.hashCode()
            result = 31 * result + data.contentHashCode()
            return result
        }
    }
    @Serializable
    data class EpubPageTarget(
        val id: String?,
        val value: String?,
        val label: String?,
        val contentSrc: String
    )

    companion object {
        const val TAG = "EpubParser"
        private const val BOOK_METADATA_FILE = "book_metadata.json"
        private const val CACHE_MANIFEST_FILE = "epub_cache_manifest.json"
        private const val EPUB_EXTRACTION_CACHE_VERSION = 1
        private const val MAX_METADATA_ENTRY_BYTES = 4 * 1024 * 1024
        private const val EPUB_COVER_MAX_DIMENSION = 1024
        private val EPUB_IMAGE_EXTENSIONS = setOf(".png", ".jpg", ".jpeg", ".gif", ".webp", ".svg")
    }

    internal val String.decodedURL: String
        get() = try {
            URLDecoder.decode(this, "UTF-8")
        } catch (e: Exception) {
            Timber.w(e, "Failed to decode URL: $this")
            this
        }

    private fun parsePageList(pageListElement: Element?, ncxFileParentDir: File): List<EpubPageTarget> {
        if (pageListElement == null) {
            Timber.d("No <pageList> element found in NCX.")
            return emptyList()
        }
        val pageTargets = mutableListOf<EpubPageTarget>()
        pageListElement.selectChildTag("pageTarget").forEach { ptElement ->
            val contentSrcRaw = ptElement.selectFirstChildTag("content")?.getAttributeValue("src")?.decodedURL
            if (contentSrcRaw != null) {
                val contentPathRelativeToEpubRoot = Paths.get(ncxFileParentDir.path, contentSrcRaw)
                    .normalize().toString().replace(File.separatorChar, '/')

                pageTargets.add(
                    EpubPageTarget(
                        id = ptElement.getAttributeValue("id"),
                        value = ptElement.getAttributeValue("value"),
                        label = ptElement.selectFirstChildTag("navLabel")?.selectFirstChildTag("text")?.textContent,
                        contentSrc = contentPathRelativeToEpubRoot
                    )
                )
            } else {
                Timber.w("PageTarget found with no content src: ${ptElement.getAttributeValue("id")}")
            }
        }
        Timber.d("Parsed ${pageTargets.size} page targets from NCX.")
        return pageTargets
    }

    private fun parseEpubCss(
        manifestItems: Map<String, EpubManifestItem>,
        filesContentMap: Map<String, EpubFile>,
        extractionRoot: File
    ): Map<String, String> {
        val listedCss = manifestItems.values
            .filter { it.mediaType == "text/css" }
            .mapNotNull { manifestItem ->
                val bytes = filesContentMap[manifestItem.absPath]?.data?.takeIf { it.isNotEmpty() }
                    ?: File(extractionRoot, manifestItem.absPath).takeIf { it.exists() }?.readBytes()

                bytes?.let { manifestItem.absPath to String(it, Charsets.UTF_8) }
            }

        val listedCssPaths = listedCss.map { it.first }.toSet()
        val unlistedCss = filesContentMap.asSequence()
            .filter { (path, _) -> path.endsWith(".css", ignoreCase = true) }
            .filterNot { (path, _) -> listedCssPaths.contains(path) }
            .mapNotNull { (path, file) ->
                val bytes = file.data.takeIf { it.isNotEmpty() }
                    ?: File(extractionRoot, path).takeIf { it.exists() }?.readBytes()

                bytes?.let { path to String(it, Charsets.UTF_8) }
            }
            .toList()

        val allCss = (listedCss + unlistedCss).toMap()

        Timber.d("Parsed ${allCss.size} CSS files (listed: ${listedCss.size}, unlisted: ${unlistedCss.size}): ${allCss.keys.joinToString()}")
        return allCss
    }

    suspend fun createEpubBook(
        inputStream: InputStream,
        bookId: String,
        shouldUseToc: Boolean = true,
        originalBookNameHint: String = "streamed_book",
        parseContent: Boolean = true,
        extractionDirOverride: File? = null,
        sourceFingerprint: String? = null
    ): EpubBook {
        return withContext(Dispatchers.IO) {
            Timber.d("Parsing EPUB input stream for bookId: $bookId")

            val shouldDeleteExtractionDir = !parseContent && extractionDirOverride == null
            val extractionDir = if (extractionDirOverride != null) {
                ImportedFileCache.prepareDirectory(extractionDirOverride)
            } else if (!parseContent) {
                ImportedFileCache.createTemporaryBookDir(context, bookId, "metadata")
            } else {
                val activeDir = ImportedFileCache.ensureActiveBookDir(context, bookId)
                readCachedEpubBook(
                    extractionDir = activeDir,
                    bookId = bookId,
                    originalBookNameHint = originalBookNameHint,
                    shouldUseToc = shouldUseToc,
                    sourceFingerprint = sourceFingerprint
                )?.let { cachedBook ->
                    Timber.tag("FileOpenPerf").d("[EPUB] Loaded extracted book from cache | bookId=$bookId")
                    return@withContext cachedBook
                }
                ImportedFileCache.resetActiveBookDir(context, bookId)
            }

            val tempFile = File.createTempFile("epub_stream", ".epub", context.cacheDir)
            val filesMap: Map<String, EpubFile>
            try {
                tempFile.outputStream().use { output ->
                    inputStream.copyTo(output)
                }
                filesMap = extractEpubContents(
                    zipFile = ZipFile(tempFile),
                    extractionDir = extractionDir,
                    parseContent = parseContent,
                    extractImagesForMetadata = shouldDeleteExtractionDir
                )
            } finally {
                tempFile.delete()
            }

            val document = createEpubDocument(filesMap)
            val book = parseAndCreateEbook(filesMap, document, shouldUseToc, extractionDir.absolutePath,
                originalBookNameHint, parseContent)
            if (parseContent && extractionDirOverride == null) {
                writeCachedEpubBook(
                    extractionDir = extractionDir,
                    bookId = bookId,
                    originalBookNameHint = originalBookNameHint,
                    shouldUseToc = shouldUseToc,
                    sourceFingerprint = sourceFingerprint,
                    book = book
                )
            }
            if (shouldDeleteExtractionDir) {
                extractionDir.deleteRecursively()
            }
            return@withContext book
        }
    }

    private fun readCachedEpubBook(
        extractionDir: File,
        bookId: String,
        originalBookNameHint: String,
        shouldUseToc: Boolean,
        sourceFingerprint: String?
    ): EpubBook? {
        val metadataFile = File(extractionDir, BOOK_METADATA_FILE)
        val manifestFile = File(extractionDir, CACHE_MANIFEST_FILE)
        if (!metadataFile.isFile || !manifestFile.isFile) return null

        return try {
            val manifest = jsonSerializer.decodeFromString<EpubExtractionCacheManifest>(manifestFile.readText())
            val isCompatible = manifest.bookId == bookId &&
                manifest.originalBookNameHint == originalBookNameHint &&
                manifest.parserVersion == EPUB_EXTRACTION_CACHE_VERSION &&
                manifest.parseContent &&
                manifest.shouldUseToc == shouldUseToc &&
                manifest.sourceFingerprint == sourceFingerprint

            if (!isCompatible) {
                Timber.d("EPUB extraction cache manifest is stale for bookId=$bookId")
                return null
            }

            val cachedBook = jsonSerializer.decodeFromString<EpubBook>(metadataFile.readText())
                .copy(extractionBasePath = extractionDir.absolutePath)

            cachedBook.takeIf { it.hasReadableExtractedContent() }
        } catch (e: Exception) {
            Timber.e(e, "Failed to read EPUB extraction cache for bookId=$bookId")
            null
        }
    }

    private fun writeCachedEpubBook(
        extractionDir: File,
        bookId: String,
        originalBookNameHint: String,
        shouldUseToc: Boolean,
        sourceFingerprint: String?,
        book: EpubBook
    ) {
        try {
            File(extractionDir, BOOK_METADATA_FILE).writeText(jsonSerializer.encodeToString(book))
            File(extractionDir, CACHE_MANIFEST_FILE).writeText(
                jsonSerializer.encodeToString(
                    EpubExtractionCacheManifest(
                        bookId = bookId,
                        originalBookNameHint = originalBookNameHint,
                        parserVersion = EPUB_EXTRACTION_CACHE_VERSION,
                        parseContent = true,
                        shouldUseToc = shouldUseToc,
                        sourceFingerprint = sourceFingerprint
                    )
                )
            )
        } catch (e: Exception) {
            Timber.e(e, "Failed to write EPUB extraction cache for bookId=$bookId")
        }
    }

    internal fun extractEpubContents(
        zipFile: ZipFile,
        extractionDir: File,
        parseContent: Boolean,
        extractImagesForMetadata: Boolean
    ): Map<String, EpubFile> {
        val filesMap = mutableMapOf<String, EpubFile>()
        zipFile.use { zf ->
            zf.entries().asSequence().filterNot { it.isDirectory }.forEach { entry ->
                val isEssential = isEssentialFile(entry.name, parseContent)
                val isImage = isEpubImageFile(entry.name)

                if (!parseContent) {
                    when {
                        isEssential -> {
                            val data = zf.readSmallEntryBytes(entry) ?: return@forEach
                            filesMap[entry.name] = EpubFile(absPath = entry.name, data = data)
                        }
                        isImage && extractImagesForMetadata -> {
                            val outputFile = safeExtractionFile(extractionDir, entry.name)
                                ?: return@forEach
                            outputFile.parentFile?.mkdirs()
                            zf.getInputStream(entry).use { input ->
                                FileOutputStream(outputFile).use { output ->
                                    input.copyTo(output)
                                }
                            }
                            filesMap[entry.name] = EpubFile(absPath = entry.name, data = ByteArray(0))
                        }
                    }
                    return@forEach
                }

                val outputFile = safeExtractionFile(extractionDir, entry.name) ?: return@forEach
                outputFile.parentFile?.mkdirs()
                zf.getInputStream(entry).use { input ->
                    FileOutputStream(outputFile).use { output ->
                        input.copyTo(output)
                    }
                }

                val lowerName = entry.name.lowercase()
                val isContainerOrOpf = lowerName.endsWith("container.xml") || lowerName.endsWith(".opf")
                val data = if (isContainerOrOpf) outputFile.readBytes() else ByteArray(0)
                filesMap[entry.name] = EpubFile(absPath = entry.name, data = data)
            }
        }
        return filesMap
    }

    private fun isEpubImageFile(fileName: String): Boolean {
        val lowerName = fileName.lowercase()
        return EPUB_IMAGE_EXTENSIONS.any { lowerName.endsWith(it) }
    }

    private fun safeExtractionFile(extractionDir: File, entryName: String): File? {
        val outputFile = File(extractionDir, entryName)
        val root = extractionDir.canonicalFile
        val target = outputFile.canonicalFile
        val rootPath = root.path
        val targetPath = target.path
        val isInsideRoot = targetPath == rootPath || targetPath.startsWith(rootPath + File.separator)

        if (!isInsideRoot) {
            Timber.w("Skipping unsafe EPUB entry outside extraction root: $entryName")
            return null
        }

        return outputFile
    }

    private fun ZipFile.readSmallEntryBytes(entry: ZipEntry): ByteArray? {
        if (entry.size > MAX_METADATA_ENTRY_BYTES.toLong()) {
            Timber.w("Skipping oversized EPUB metadata entry: ${entry.name} (${entry.size} bytes)")
            return null
        }

        val initialSize = entry.size
            .takeIf { it in 0..MAX_METADATA_ENTRY_BYTES.toLong() }
            ?.toInt()
            ?: DEFAULT_BUFFER_SIZE
        val buffer = ByteArray(DEFAULT_BUFFER_SIZE)
        var totalBytes = 0

        return getInputStream(entry).use { input ->
            ByteArrayOutputStream(initialSize).use { output ->
                while (true) {
                    val read = input.read(buffer)
                    if (read == -1) break

                    totalBytes += read
                    if (totalBytes > MAX_METADATA_ENTRY_BYTES) {
                        Timber.w("Skipping oversized EPUB metadata entry while reading: ${entry.name}")
                        return null
                    }

                    output.write(buffer, 0, read)
                }
                output.toByteArray()
            }
        }
    }

    private suspend fun parseAndCreateEbook(
        filesContentMap: Map<String, EpubFile>,
        document: EpubDocument,
        shouldUseToc: Boolean,
        extractionBasePath: String,
        originalFilePathOrKey: String,
        parseContent: Boolean = true
    ): EpubBook = withContext(Dispatchers.IO) {
        val metadataTitle =
            document.metadata.selectFirstChildTag("dc:title")?.textContent ?: File(originalFilePathOrKey).nameWithoutExtension
        val metadataAuthor =
            document.metadata.selectFirstChildTag("dc:creator")?.textContent ?: "Unknown Author"
        val metadataLanguage =
            document.metadata.selectFirstChildTag("dc:language")?.textContent ?: "en"
        val metadataCoverId = getMetadataCoverId(document.metadata)
        val metadataDescription =
            document.metadata.selectFirstChildTag("dc:description")?.textContent

        Timber.d("EpubParser: Extracted OPF metadata: title='$metadataTitle', author='$metadataAuthor'")

        var metadataSeriesName: String? = null
        var metadataSeriesIndex: Double? = null

        document.metadata.selectChildTag("meta")
            .ifEmpty { document.metadata.selectChildTag("opf:meta") }
            .forEach { meta ->
                val nameAttr = meta.getAttributeValue("name")
                val contentAttr = meta.getAttributeValue("content")

                if (nameAttr == "calibre:series") metadataSeriesName = contentAttr
                if (nameAttr == "calibre:series_index") metadataSeriesIndex = contentAttr?.toDoubleOrNull()
            }

        val opfRelativePath = document.opfFilePath
        val opfParentDir = File(opfRelativePath).parentFile ?: File("")
        val manifestItems = getManifestItems(document.manifest, opfParentDir)
        var pageTargets: List<EpubPageTarget> = emptyList()
        val ncxMetadataMap = mutableMapOf<String, NcxMetadata>()
        val extractionRoot = File(extractionBasePath)

        if (shouldUseToc) {
            Timber.d("shouldUseToc is true. Attempting to parse NCX.")
            val tocFileItem = manifestItems.values.firstOrNull {
                it.absPath.endsWith(".ncx", ignoreCase = true)
            }
            if (tocFileItem != null) {
                val ncxParentDir = File(tocFileItem.absPath).parentFile ?: File("")
                val ncxData = filesContentMap[tocFileItem.absPath]?.data?.takeIf { it.isNotEmpty() }
                    ?: File(extractionRoot, tocFileItem.absPath).takeIf { it.exists() }?.readBytes()

                val tocDocumentNode = ncxData?.let { parseXMLFile(it) }

                if (tocDocumentNode != null) {
                    Timber.d("Successfully parsed NCX file: ${tocFileItem.absPath}")
                    val pageListElement = tocDocumentNode.selectFirstTag("pageList") as Element?
                    pageTargets = parsePageList(pageListElement, ncxParentDir)

                    val navMapElement = tocDocumentNode.selectFirstTag("navMap") as Element?
                    if (navMapElement != null) {
                        // Recursively parse navMap
                        ncxMetadataMap.putAll(parseNavMapRecursive(navMapElement, ncxParentDir))
                    } else {
                        Timber.d("No <navMap> element found in NCX.")
                    }
                } else {
                    Timber.w("NCX file item '${tocFileItem.absPath}' found in manifest but could not be parsed.")
                }
            } else {
                Timber.d("No NCX file found in manifest. Skipping NCX-based PageList/NavMap.")
            }
        } else {
            Timber.d("shouldUseToc is false. Skipping NCX parsing for PageList/NavMap.")
        }

        Timber.d("Parsing chapters based on OPF spine for rendering order. NCX titles/depth will be used if available.")

        val tableOfContents = if (shouldUseToc) {
            val tocFileItem = manifestItems.values.firstOrNull {
                it.absPath.endsWith(".ncx", ignoreCase = true)
            }
            if (tocFileItem != null) {
                val ncxParentDir = File(tocFileItem.absPath).parentFile ?: File("")
                val ncxData = filesContentMap[tocFileItem.absPath]?.data?.takeIf { it.isNotEmpty() }
                    ?: File(extractionRoot, tocFileItem.absPath).takeIf { it.exists() }?.readBytes()
                val tocDocumentNode = ncxData?.let { parseXMLFile(it) }
                val navMapElement = tocDocumentNode?.selectFirstTag("navMap") as Element?

                if (navMapElement != null) {
                    parseTableOfContents(navMapElement, ncxParentDir)
                } else {
                    emptyList()
                }
            } else {
                emptyList()
            }
        } else {
            emptyList()
        }

        val chaptersFromSpine = if (parseContent) {
            parseUsingSpine(document.spine, manifestItems, filesContentMap, ncxMetadataMap, extractionRoot)
        } else {
            emptyList()
        }

        Timber.d("Parsing images (for cover and general access)")
        val images = if (parseContent) {
            parseEpubImages(manifestItems, filesContentMap, extractionRoot)
        } else {
            emptyList()
        }

        Timber.d("Parsing cover image")
        val coverImage = parseCoverImage(metadataCoverId, manifestItems, filesContentMap, extractionRoot)

        val cssContent = if (parseContent) {
            parseEpubCss(manifestItems, filesContentMap, extractionRoot)
        } else {
            emptyMap()
        }

        Timber.d("EpubBook created with ${chaptersFromSpine.size} spine chapters.")
        return@withContext EpubBook(
            fileName = metadataTitle.asFileName(),
            title = metadataTitle,
            author = metadataAuthor,
            language = metadataLanguage,
            coverImage = coverImage,
            chapters = chaptersFromSpine, chaptersForPagination = chaptersFromSpine,
            images = images,
            pageList = pageTargets,
            tableOfContents = tableOfContents,
            extractionBasePath = extractionBasePath,
            css = cssContent,
            seriesName = metadataSeriesName,
            seriesIndex = metadataSeriesIndex,
            description = metadataDescription
        )
    }

    private fun parseTableOfContents(
        navMapElement: Element,
        ncxParentDir: File
    ): List<EpubTocEntry> {
        val result = mutableListOf<EpubTocEntry>()

        fun recurse(element: Element, currentDepth: Int) {
            val navPoints = element.childElements.filter { it.tagName == "navPoint" }
            for (navPoint in navPoints) {
                val label = navPoint.selectFirstChildTag("navLabel")
                    ?.selectFirstChildTag("text")?.textContent?.trim() ?: "Untitled"

                val contentSrc = navPoint.selectFirstChildTag("content")
                    ?.getAttributeValue("src")?.decodedURL

                if (contentSrc != null) {
                    val fullPathRaw = Paths.get(ncxParentDir.path, contentSrc)
                        .normalize().toString().replace(File.separatorChar, '/')

                    val parts = fullPathRaw.split("#", limit = 2)
                    val absolutePath = parts[0]
                    val fragmentId = if (parts.size > 1) parts[1] else null

                    result.add(EpubTocEntry(label, absolutePath, fragmentId, currentDepth))

                    recurse(navPoint, currentDepth + 1)
                }
            }
        }

        recurse(navMapElement, 0)
        return result
    }


    @Throws(EpubParserException::class)
    private fun createEpubDocument(files: Map<String, EpubFile>): EpubDocument {
        val containerFile = files["META-INF/container.xml"]
            ?: throw EpubParserException("META-INF/container.xml file missing")

        val rawOpfPath = parseXMLFile(containerFile.data)?.selectFirstTag("rootfile")
            ?.getAttributeValue("full-path")?.decodedURL
            ?: throw EpubParserException("Invalid container.xml: Could not find rootfile full-path.")

        val opfFilePath = rawOpfPath.trimStart('/')

        val opfFile = files[opfFilePath]
            ?: throw EpubParserException(".opf file missing at normalized path '$opfFilePath'.")

        val document = parseXMLFile(opfFile.data)
            ?: throw EpubParserException(".opf file failed to parse data from '$opfFilePath'")
        val metadata = document.selectFirstTag("metadata")
            ?: document.selectFirstTag("opf:metadata")
            ?: throw EpubParserException(".opf file metadata section missing in '$opfFilePath'")
        val manifest = document.selectFirstTag("manifest")
            ?: document.selectFirstTag("opf:manifest")
            ?: throw EpubParserException(".opf file manifest section missing in '$opfFilePath'")
        val spine = document.selectFirstTag("spine")
            ?: document.selectFirstTag("opf:spine")
            ?: throw EpubParserException(".opf file spine section missing in '$opfFilePath'")

        return EpubDocument(metadata, manifest, spine, opfFilePath)
    }


    private fun getMetadataCoverId(metadata: Node): String? {
        return metadata.selectChildTag("meta")
            .ifEmpty { metadata.selectChildTag("opf:meta") }
            .find { it.getAttributeValue("name") == "cover" }?.getAttributeValue("content")
    }

    private fun getManifestItems(
        manifest: Node,
        opfParentDir: File
    ): Map<String, EpubManifestItem> {
        return manifest.selectChildTag("item")
            .ifEmpty { manifest.selectChildTag("opf:item") }
            .mapNotNull { itemElement ->
                val href = itemElement.getAttribute("href")?.decodedURL ?: return@mapNotNull null
                val pathRelativeToEpubRoot = Paths.get(opfParentDir.path, href)
                    .normalize().toString().replace(File.separatorChar, '/')

                EpubManifestItem(
                    id = itemElement.getAttribute("id"),
                    absPath = pathRelativeToEpubRoot,
                    mediaType = itemElement.getAttribute("media-type"),
                    properties = itemElement.getAttribute("properties")
                )
            }.associateBy { it.id }
    }

    private fun parseNavMapRecursive(
        element: Element,
        ncxFileParentDir: File,
        depth: Int = 0
    ): Map<String, NcxMetadata> {
        val result = mutableMapOf<String, NcxMetadata>()

        val navPoints = element.childElements.filter { it.tagName == "navPoint" }

        for (navPoint in navPoints) {
            val navLabelText = navPoint.selectFirstChildTag("navLabel")
                ?.selectFirstChildTag("text")?.textContent?.trim()
            val contentSrcRaw = navPoint.selectFirstChildTag("content")
                ?.getAttributeValue("src")?.decodedURL

            if (navLabelText != null && contentSrcRaw != null && navLabelText.isNotEmpty()) {
                val contentPathRelativeToEpubRoot = Paths.get(ncxFileParentDir.path, contentSrcRaw)
                    .normalize().toString().replace(File.separatorChar, '/')
                    .substringBefore('#')

                if (!result.containsKey(contentPathRelativeToEpubRoot)) {
                    result[contentPathRelativeToEpubRoot] = NcxMetadata(navLabelText, depth)
                    Timber.d("NCX Map: '$contentPathRelativeToEpubRoot' -> '$navLabelText' (Depth $depth)")
                }
            }
            result.putAll(parseNavMapRecursive(navPoint, ncxFileParentDir, depth + 1))
        }

        return result
    }

    private fun generateId(): String {
        return UUID.randomUUID().toString()
    }

    private suspend fun parseUsingSpine(
        spine: Node,
        manifestItems: Map<String, EpubManifestItem>,
        filesContentMap: Map<String, EpubFile>,
        ncxMetadataMap: Map<String, NcxMetadata>,
        extractionRoot: File
    ): List<EpubChapter> = withContext(Dispatchers.Default) {
        val parsingSemaphore = Semaphore(6)

        val spineItems = spine.selectChildTag("itemref")
            .ifEmpty { spine.selectChildTag("opf:itemref") }

        val deferredChapters = spineItems.mapIndexed { index, itemRef ->
            async {
                parsingSemaphore.withPermit {
                    val idRef = itemRef.getAttribute("idref")
                    val item = manifestItems[idRef] ?: return@withPermit null

                    val fileBytes = filesContentMap[item.absPath]?.data?.takeIf { it.isNotEmpty() }
                        ?: File(extractionRoot, item.absPath).takeIf { it.exists() }?.readBytes()
                        ?: return@withPermit null

                    val mediaType = item.mediaType
                    val absPath = item.absPath

                    if (mediaType.startsWith("application/xhtml+xml") ||
                        mediaType.startsWith("text/html") ||
                        absPath.endsWith(".html", ignoreCase = true) ||
                        absPath.endsWith(".xhtml", ignoreCase = true) ||
                        absPath.endsWith(".xml", ignoreCase = true)
                    ) {
                        val rawHtml = String(fileBytes, Charsets.UTF_8)
                        val document = Jsoup.parse(rawHtml)
                        val plainText = document.text()

                        val parser = EpubXMLFileParser(
                            fileRelativePath = absPath,
                            data = fileBytes,
                            fragmentId = null
                        )
                        val res = parser.parseForTitleAndPath(document)

                        val chapterTitleFromHtml = res.title
                        val ncxKey = absPath.substringBefore('#')
                        val ncxData = ncxMetadataMap[ncxKey]

                        val isEffectiveInToc = if (ncxMetadataMap.isNotEmpty()) {
                            ncxData != null
                        } else {
                            true
                        }

                        val finalChapterTitle = if (ncxData != null && ncxData.title.isNotBlank()) {
                            ncxData.title
                        } else {
                            chapterTitleFromHtml
                        }
                        val finalDepth = ncxData?.depth ?: 0

                        TempEpubChapter(
                            url = absPath,
                            title = finalChapterTitle,
                            htmlFilePath = res.effectiveHtmlPath,
                            chapterIndex = index + 1,
                            plainTextContent = plainText,
                            htmlContent = "", // OPTIMIZATION: Don't store HTML in memory, it's on disk
                            depth = finalDepth,
                            isInToc = isEffectiveInToc
                        )
                    } else if (mediaType.startsWith("image/")) {
                        // Image handling remains similar, but usually small enough
                        val htmlContent = """
                            <!DOCTYPE html><html style="margin:0;padding:0;height:100%;"><head><title>Image</title></head><body style="margin:0;padding:0;height:100%;text-align:center;"><img src="$absPath" alt="Image from spine" style="object-fit:contain;width:100%;height:100%;"/></body></html>
                        """.trimIndent()

                        val ncxKey = absPath.substringBefore('#')
                        val ncxData = ncxMetadataMap[ncxKey]
                        val isEffectiveInToc = if (ncxMetadataMap.isNotEmpty()) ncxData != null else true

                        TempEpubChapter(
                            url = absPath,
                            title = ncxData?.title ?: "Image",
                            htmlFilePath = absPath,
                            chapterIndex = index + 1,
                            plainTextContent = "[Image]",
                            htmlContent = htmlContent,
                            depth = ncxData?.depth ?: 0,
                            isInToc = isEffectiveInToc
                        )
                    } else {
                        null
                    }
                }
            }
        }

        val tempChapters = deferredChapters.toList().awaitAll().filterNotNull()

        return@withContext tempChapters.map { tempChapter ->
            EpubChapter(
                chapterId = generateId(),
                absPath = tempChapter.url,
                title = tempChapter.title?.takeIf { it.isNotBlank() } ?: "Chapter ${tempChapter.chapterIndex}",
                htmlFilePath = tempChapter.htmlFilePath,
                plainTextContent = tempChapter.plainTextContent,
                htmlContent = tempChapter.htmlContent,
                depth = tempChapter.depth,
                isInToc = tempChapter.isInToc
            )
        }.filter { it.htmlFilePath.isNotBlank() }
    }

    private fun parseEpubImages(
        manifestItems: Map<String, EpubManifestItem>,
        filesContentMap: Map<String, EpubFile>,
        @Suppress("UNUSED_PARAMETER") extractionRoot: File
    ): List<EpubImage> {
        val listedImages = manifestItems.values
            .filter { it.mediaType.startsWith("image/") }
            .map { manifestItem ->
                EpubImage(absPath = manifestItem.absPath)
            }

        val listedPaths = listedImages.map { it.absPath }.toSet()

        val unlistedImages = filesContentMap.keys
            .filter { path ->
                val lowerPath = path.lowercase()
                EPUB_IMAGE_EXTENSIONS.any { lowerPath.endsWith(it) } && !listedPaths.contains(path)
            }
            .map { path ->
                EpubImage(absPath = path)
            }

        Timber.d("Identified ${listedImages.size + unlistedImages.size} images (content not loaded into memory).")
        return (listedImages + unlistedImages).distinctBy { it.absPath }
    }

    private fun parseCoverImage(
        metadataCoverId: String?,
        manifestItems: Map<String, EpubManifestItem>,
        filesContentMap: Map<String, EpubFile>,
        extractionRoot: File
    ): Bitmap? {
        val coverManifestItem = manifestItems[metadataCoverId]
        if (coverManifestItem != null) {
            val coverImage = decodeEpubImage(coverManifestItem.absPath, filesContentMap, extractionRoot)
            if (coverImage != null) {
                return coverImage
            } else {
                Timber.e("Cover image file content not found for path: ${coverManifestItem.absPath}")
            }
        } else {
            if (metadataCoverId != null) {
                Timber.w("Cover image ID '$metadataCoverId' not found in manifest.")
            } else {
                Timber.d("No cover image ID specified in metadata.")
            }
        }

        val commonCoverNames = listOf("cover.jpg", "cover.jpeg", "cover.png")
        for (name in commonCoverNames) {
            val possiblePaths = listOf(
                name, "images/$name", "Images/$name", "image/$name", "Image/$name",
                "OEBPS/images/$name", "OEBPS/Images/$name", "OEBPS/image/$name", "OEBPS/Image/$name",
                "OPS/images/$name", "OPS/Images/$name", "OPS/image/$name", "OPS/Image/$name"
            )
            for (path in possiblePaths) {
                if (filesContentMap.containsKey(path)) {
                    decodeEpubImage(path, filesContentMap, extractionRoot)?.let {
                        Timber.d("Found fallback cover image at $path")
                        return it
                    }
                }
                manifestItems.values.find { item -> item.absPath.equals(path, ignoreCase = true) && item.mediaType.startsWith("image/") }?.let { manifestItem ->
                    decodeEpubImage(manifestItem.absPath, filesContentMap, extractionRoot)?.let {
                        Timber.d("Found fallback cover image via manifest item (case-insensitive) at ${manifestItem.absPath}")
                        return it
                    }
                }
            }
        }
        Timber.d("Cover image could not be loaded from metadata or common fallbacks.")
        return null
    }

    private fun decodeEpubImage(
        path: String,
        filesContentMap: Map<String, EpubFile>,
        extractionRoot: File
    ): Bitmap? {
        filesContentMap[path]?.data?.takeIf { it.isNotEmpty() }?.let { bytes ->
            return decodeSampledByteArray(bytes)
        }

        val imageFile = File(extractionRoot, path).takeIf { it.exists() && it.isFile } ?: return null
        return decodeSampledFile(imageFile)
    }

    private fun decodeSampledByteArray(bytes: ByteArray): Bitmap? {
        val bounds = BitmapFactory.Options().apply {
            inJustDecodeBounds = true
        }
        BitmapFactory.decodeByteArray(bytes, 0, bytes.size, bounds)
        val options = BitmapFactory.Options().apply {
            inSampleSize = calculateBitmapSampleSize(bounds)
        }
        return BitmapFactory.decodeByteArray(bytes, 0, bytes.size, options)
    }

    private fun decodeSampledFile(file: File): Bitmap? {
        val bounds = BitmapFactory.Options().apply {
            inJustDecodeBounds = true
        }
        BitmapFactory.decodeFile(file.absolutePath, bounds)
        val options = BitmapFactory.Options().apply {
            inSampleSize = calculateBitmapSampleSize(bounds)
        }
        return BitmapFactory.decodeFile(file.absolutePath, options)
    }

    private fun calculateBitmapSampleSize(options: BitmapFactory.Options): Int {
        val width = options.outWidth
        val height = options.outHeight
        if (width <= 0 || height <= 0) return 1

        var sampleSize = 1
        while ((width / sampleSize) > EPUB_COVER_MAX_DIMENSION || (height / sampleSize) > EPUB_COVER_MAX_DIMENSION) {
            sampleSize *= 2
        }
        return sampleSize
    }

    private fun isEssentialFile(fileName: String, parseContent: Boolean): Boolean {
        val lowerName = fileName.lowercase()

        if (lowerName.endsWith("container.xml") || lowerName.endsWith(".opf")) {
            return true
        }

        if (!parseContent) {
            return false
        }

        return lowerName.endsWith(".xml") ||
                lowerName.endsWith(".ncx") ||
                lowerName.endsWith(".html") ||
                lowerName.endsWith(".xhtml") ||
                lowerName.endsWith(".htm") ||
                lowerName.endsWith(".css")
    }
}