Pull basic manga/chapter metadata from EPUB files. (#2962)

* Fill manga and chapter metadata with EPUB metadata where appropriate.

* Use fuzzy-match to remove manga titles from chapter names.

This allows removing manga titles from the metadata title of local EPUB chapters with characters that can't be in the manga directory's name due to filesystem limitations.

* Use more standard way of parsing EPUB dates.

* Use date format compatible with lower Android versions for EPUBs.

app/src/main/java/eu/kanade/tachiyomi/source/LocalSource.kt

@@ -112,12 +112,20 @@ class LocalSource(private val context: Context) : CatalogueSource {
                     }
                 }
 
-                // Copy the cover from the first chapter found.
+                val chapters = fetchChapterList(this).toBlocking().first()
-                if (thumbnail_url == null) {
+                if (chapters.isNotEmpty()) {
-                    val chapters = fetchChapterList(this).toBlocking().first()
+                    val chapter = chapters.last()
-                    if (chapters.isNotEmpty()) {
+                    val format = getFormat(chapter)
+                    if (format is Format.Epub) {
+                        EpubFile(format.file).use { epub ->
+                            epub.fillMangaMetadata(this)
+                        }
+                    }
+
+                    // Copy the cover from the first chapter found.
+                    if (thumbnail_url == null) {
                         try {
-                            val dest = updateCover(chapters.last(), this)
+                            val dest = updateCover(chapter, this)
                             thumbnail_url = dest?.absolutePath
                         } catch (e: Exception) {
                             Timber.e(e)
@@ -158,14 +166,22 @@ class LocalSource(private val context: Context) : CatalogueSource {
             .map { chapterFile ->
                 SChapter.create().apply {
                     url = "${manga.url}/${chapterFile.name}"
-                    val chapName = if (chapterFile.isDirectory) {
+                    name = if (chapterFile.isDirectory) {
                         chapterFile.name
                     } else {
                         chapterFile.nameWithoutExtension
                     }
-                    val chapNameCut = chapName.replace(manga.title, "", true).trim(' ', '-', '_')
-                    name = if (chapNameCut.isEmpty()) chapName else chapNameCut
                     date_upload = chapterFile.lastModified()
+
+                    val format = getFormat(this)
+                    if (format is Format.Epub) {
+                        EpubFile(format.file).use { epub ->
+                            epub.fillChapterMetadata(this)
+                        }
+                    }
+
+                    val chapNameCut = stripMangaTitle(name, manga.title)
+                    if (chapNameCut.isNotEmpty()) name = chapNameCut
                     ChapterRecognition.parseChapterNumber(this, manga)
                 }
             }
@@ -180,6 +196,40 @@ class LocalSource(private val context: Context) : CatalogueSource {
         return Observable.just(chapters)
     }
 
+    /**
+     * Strips the manga title from a chapter name, matching only based on alphanumeric and whitespace
+     * characters.
+     */
+    private fun stripMangaTitle(chapterName: String, mangaTitle: String): String {
+        var chapterNameIndex = 0
+        var mangaTitleIndex = 0
+        while (chapterNameIndex < chapterName.length && mangaTitleIndex < mangaTitle.length) {
+            val chapterChar = chapterName.get(chapterNameIndex)
+            val mangaChar = mangaTitle.get(mangaTitleIndex)
+            if (!chapterChar.equals(mangaChar, true)) {
+                val invalidChapterChar = !chapterChar.isLetterOrDigit() && !chapterChar.isWhitespace()
+                val invalidMangaChar = !mangaChar.isLetterOrDigit() && !mangaChar.isWhitespace()
+
+                if (!invalidChapterChar && !invalidMangaChar) {
+                    return chapterName
+                }
+
+                if (invalidChapterChar) {
+                    chapterNameIndex++
+                }
+
+                if (invalidMangaChar) {
+                    mangaTitleIndex++
+                }
+            } else {
+                chapterNameIndex++
+                mangaTitleIndex++
+            }
+        }
+
+        return chapterName.substring(chapterNameIndex).trimStart(' ', '-', '_', ',', ':')
+    }
+
     override fun fetchPageList(chapter: SChapter): Observable<List<Page>> {
         return Observable.error(Exception("Unused"))
     }

app/src/main/java/eu/kanade/tachiyomi/util/storage/EpubFile.kt

@@ -1,8 +1,13 @@
 package eu.kanade.tachiyomi.util.storage
 
+import eu.kanade.tachiyomi.source.model.SChapter
+import eu.kanade.tachiyomi.source.model.SManga
 import java.io.Closeable
 import java.io.File
 import java.io.InputStream
+import java.text.ParseException
+import java.text.SimpleDateFormat
+import java.util.Locale
 import java.util.zip.ZipEntry
 import java.util.zip.ZipFile
 import org.jsoup.Jsoup
@@ -44,6 +49,58 @@ class EpubFile(file: File) : Closeable {
         return zip.getEntry(name)
     }
 
+    /**
+     * Fills manga metadata using this epub file's metadata.
+     */
+    fun fillMangaMetadata(manga: SManga) {
+        val ref = getPackageHref()
+        val doc = getPackageDocument(ref)
+
+        val creator = doc.getElementsByTag("dc:creator").first()
+        val description = doc.getElementsByTag("dc:description").first()
+
+        manga.author = creator?.text()
+        manga.description = description?.text()
+    }
+
+    /**
+     * Fills chapter metadata using this epub file's metadata.
+     */
+    fun fillChapterMetadata(chapter: SChapter) {
+        val ref = getPackageHref()
+        val doc = getPackageDocument(ref)
+
+        val title = doc.getElementsByTag("dc:title").first()
+        val publisher = doc.getElementsByTag("dc:publisher").first()
+        val creator = doc.getElementsByTag("dc:creator").first()
+        var date = doc.getElementsByTag("dc:date").first()
+        if (date == null) {
+            date = doc.select("meta[property=dcterms:modified]").first()
+        }
+
+        if (title != null) {
+            chapter.name = title.text()
+        }
+
+        if (publisher != null) {
+            chapter.scanlator = publisher.text()
+        } else if (creator != null) {
+            chapter.scanlator = creator.text()
+        }
+
+        if (date != null) {
+            val dateFormat = SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ssZ", Locale.getDefault())
+            try {
+                val parsedDate = dateFormat.parse(date.text())
+                if (parsedDate != null) {
+                    chapter.date_upload = parsedDate.time
+                }
+            } catch (e: ParseException) {
+                // Empty
+            }
+        }
+    }
+
     /**
      * Returns the path of all the images found in the epub file.
      */