Add support for old EPUB 2 metadata (dc-metadata and x-metadata tags)

readium · Jul 31, 2020 · f744e6c · f744e6c
1 parent 4eec6ac
commit f744e6c
Show file tree

Hide file tree

Showing 4 changed files with 54 additions and 9 deletions.
diff --git a/r2-streamer-swift/Parser/EPUB/EPUBMetadataParser.swift b/r2-streamer-swift/Parser/EPUB/EPUBMetadataParser.swift
@@ -192,14 +192,14 @@ final class EPUBMetadataParser: Loggable {
 
     /// Parse and return the Epub unique identifier.
     /// https://github.com/readium/architecture/blob/master/streamer/parser/metadata.md#identifier
-    private lazy var uniqueIdentifier: String? = metadataElement?
-        .firstChild(xpath:"dc:identifier[@id=/opf:package/@unique-identifier]")?
-        .stringValue
+    private lazy var uniqueIdentifier: String? =
+        dcElement(tag: "identifier[@id=/opf:package/@unique-identifier]")?
+            .stringValue
 
     /// https://github.com/readium/architecture/blob/master/streamer/parser/metadata.md#publication-date
-    private lazy var publishedDate = metadataElement?
-        .firstChild(xpath: "dc:date[not(@opf:event) or @opf:event='publication']")?
-        .stringValue.dateFromISO8601
+    private lazy var publishedDate =
+        dcElement(tag: "date[not(@opf:event) or @opf:event='publication']")?
+            .stringValue.dateFromISO8601
 
     /// Parse the modifiedDate (date of last modification of the EPUB).
     /// https://github.com/readium/architecture/blob/master/streamer/parser/metadata.md#modification-date
@@ -210,7 +210,7 @@ final class EPUBMetadataParser: Loggable {
                 .first
         }
         let epub2Date = {
-            self.metadataElement?.firstChild(xpath: "dc:date[@opf:event='modification']")?
+            self.dcElement(tag: "date[@opf:event='modification']")?
                 .stringValue.dateFromISO8601
         }
         return epub3Date() ?? epub2Date()
@@ -479,4 +479,12 @@ final class EPUBMetadataParser: Loggable {
         }
     }
 
+    /// Returns the given `dc:` tag in the `metadata` element.
+    ///
+    /// This looks under `metadata/dc-metadata` as well, to be compatible with old EPUB 2 files.
+    private func dcElement(tag: String) -> XMLElement? {
+        return metadataElement?
+            .firstChild(xpath:"(.|opf:dc-metadata)/dc:\(tag)")
+    }
+
 }
diff --git a/r2-streamer-swift/Parser/EPUB/OPFMeta.swift b/r2-streamer-swift/Parser/EPUB/OPFMeta.swift
@@ -152,9 +152,10 @@ struct OPFMetaList {
         let prefixes = OPFVocabulary.prefixes(in: document)
         document.definePrefix("opf", forNamespace: "http://www.idpf.org/2007/opf")
         document.definePrefix("dc", forNamespace: "http://purl.org/dc/elements/1.1/")
-        
+
         // Parses `<meta>` and `<dc:x>` tags in order of appearance.
-        self.metas = document.xpath("/opf:package/opf:metadata/opf:meta|/opf:package/opf:metadata/dc:*")
+        let root = "/opf:package/opf:metadata"
+        self.metas = document.xpath("\(root)/opf:meta|\(root)/dc:*|\(root)/opf:dc-metadata/dc:*|\(root)/opf:x-metadata/opf:meta")
             .compactMap { meta in
                 if meta.tag == "meta" {
                     // EPUB 3

diff --git a/r2-streamer-swiftTests/Parser/EPUB/EPUBMetadataParserTests.swift b/r2-streamer-swiftTests/Parser/EPUB/EPUBMetadataParserTests.swift
@@ -99,6 +99,17 @@ class EPUBMetadataParserTests: XCTestCase {
         ))
     }
 
+    /// Old EPUB 2 files sometimes contain the `dc` tags under `dc-metadata` and `x-metadata`.
+    /// See http://idpf.org/epub/20/spec/OPF_2.0_final_spec.html#Section2.2
+    func testParseUnderDCMetadataElement() throws {
+        let sut = try parseMetadata("dc-metadata")
+
+        XCTAssertEqual(sut.identifier, "urn:uuid:1a16ce38-82bd-4e9b-861e-773c2e787a50")
+        XCTAssertEqual(sut.title, "Alice's Adventures in Wonderland")
+        XCTAssertEqual(sut.modified, "2012-04-02T12:47:00Z".dateFromISO8601)
+        XCTAssertEqual(sut.authors, [Contributor(name: "Lewis Carroll")])
+    }
+
     func testParseMainTitle() throws {
         let sut = try parseMetadata("title-main")
         XCTAssertEqual(sut.title, "Main title takes precedence")

diff --git a/r2-streamer-swiftTests/Samples/OPF/dc-metadata.opf b/r2-streamer-swiftTests/Samples/OPF/dc-metadata.opf
@@ -0,0 +1,25 @@
+<?xml version="1.0"?>
+<package xmlns="http://www.idpf.org/2007/opf" unique-identifier="bookid" version="2.0">
+ <!--
+    Old EPUB 2 files sometimes contain the `dc` tags under `dc-metadata` and `x-metadata`.
+    See http://idpf.org/epub/20/spec/OPF_2.0_final_spec.html#Section2.2
+ -->
+ <metadata>
+  <dc-metadata xmlns:dc="http://purl.org/dc/elements/1.1/"
+    xmlns:dcterms="http://purl.org/dc/terms/"
+    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+   <dc:title>Alice's Adventures in Wonderland</dc:title>
+   <dc:creator>Lewis Carroll</dc:creator>
+   <dc:identifier id="bookid">urn:uuid:1a16ce38-82bd-4e9b-861e-773c2e787a50</dc:identifier>
+  </dc-metadata>
+  <x-metadata>
+    <meta property="dcterms:modified">2012-04-02T12:47:00Z</meta>
+  </x-metadata>
+ </metadata>
+  <manifest>
+    <item id="titlepage" href="titlepage.xhtml" media-type="application/xhtml+xml" />
+  </manifest>
+  <spine>
+    <itemref idref="titlepage"/>
+  </spine>
+</package>