Backup_Repos/komga
1
0
Fork 0
mirror of https://github.com/gotson/komga.git synced 2025-12-22 00:13:30 +01:00
Code Issues Projects Releases Wiki Activity

fix(epub): some EPUB files would not be parsed correctly

Browse source
This commit is contained in:
Gauthier Roebroeck 2020-04-10 11:49:39 +08:00
parent a4f5015435
commit f0ff785d66
1 changed files with 11 additions and 11 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

22
komga/src/main/kotlin/org/gotson/komga/infrastructure/mediacontainer/EpubExtractor.kt
View file

@ -18,12 +18,10 @@ class EpubExtractor(contentDetector: ContentDetector) : ZipExtractor(contentDete
override fun getEntries(path: Path): List<MediaContainerEntry> {
ZipFile(path.toFile()).use { zip ->
try {
val packagePath = zip.getEntry("META-INF/container.xml").let { entry ->
val container = zip.getInputStream(entry).use { Jsoup.parse(it, null, "") }
container.getElementsByTag("rootfile").first().attr("full-path")
}
val packagePath = getPackagePath(zip)
val opf = zip.getInputStream(zip.getEntry(packagePath)).use { Jsoup.parse(it, null, "") }
val opfPath = Paths.get(packagePath).parentOrEmpty()
val manifest = opf.select("manifest > item")
.associate { it.attr("id") to ManifestItem(it.attr("id"), it.attr("href"), it.attr("media-type")) }
@ -32,15 +30,17 @@ class EpubExtractor(contentDetector: ContentDetector) : ZipExtractor(contentDete
.mapNotNull { manifest[it] }
.map { it.href }
val images = pages.flatMap { pagePath ->
val doc = zip.getInputStream(zip.getEntry(pagePath)).use { Jsoup.parse(it, null, "") }
doc.getElementsByTag("img")
.map { it.attr("src") }
.map { Paths.get(pagePath).parent?.resolve(it).toString() }
}
val images = pages
.map { opfPath.resolve(it).normalize() }
.flatMap { pagePath ->
val doc = zip.getInputStream(zip.getEntry(pagePath.toString())).use { Jsoup.parse(it, null, "") }
doc.getElementsByTag("img")
.map { it.attr("src") }
.map { pagePath.parentOrEmpty().resolve(it).normalize() }
}
return images.map { image ->
MediaContainerEntry(image, manifest.values.first { it.href == image }.mediaType)
MediaContainerEntry(image.toString(), manifest.values.first { it.href == opfPath.relativize(image).toString() }.mediaType)
}
} catch (e: Exception) {
logger.error(e) { "File is not a proper Epub, treating it as a zip file" }