EPUB: add "identifiers" to doc props (koreader#560)

ie. "isbn:9783462312317", "asin:b0ckwvvv4z".
Frenzie · Apr 6, 2024 · 5d2724b · 5d2724b
1 parent e6a5cc9
commit 5d2724b
Show file tree

Hide file tree

Showing 3 changed files with 33 additions and 0 deletions.
diff --git a/crengine/include/lvdocview.h b/crengine/include/lvdocview.h
@@ -768,6 +768,8 @@ class LVDocView : public CacheLoadingCallback
     lString32 getDescription() { return m_doc_props->getStringDef(DOC_PROP_DESCRIPTION); }
     /// returns book keywords (separated by "; ")
     lString32 getKeywords() { return m_doc_props->getStringDef(DOC_PROP_KEYWORDS); }
+    /// returns book identifiers (scheme:identifier separated by "\n")
+    lString32 getIdentifiers() { return m_doc_props->getStringDef(DOC_PROP_IDENTIFIERS); }
     /// returns book series name and number (series name #1)
     lString32 getSeries()
     {

diff --git a/crengine/include/lvtinydom.h b/crengine/include/lvtinydom.h
@@ -94,6 +94,7 @@ extern const int gDOMVersionCurrent;
 #define DOC_PROP_LANGUAGE        "doc.language"
 #define DOC_PROP_DESCRIPTION     "doc.description"
 #define DOC_PROP_KEYWORDS        "doc.keywords"
+#define DOC_PROP_IDENTIFIERS     "doc.identifiers"
 #define DOC_PROP_SERIES_NAME     "doc.series.name"
 #define DOC_PROP_SERIES_NUMBER   "doc.series.number"
 #define DOC_PROP_ARC_NAME        "doc.archive.name"

diff --git a/crengine/src/epubfmt.cpp b/crengine/src/epubfmt.cpp
@@ -1384,6 +1384,36 @@ bool ImportEpubDocument( LVStreamRef stream, ldomDocument * m_doc, LVDocViewCall
         m_doc_props->setString(DOC_PROP_KEYWORDS, subjects);
         CRLog::info("Authors: %s Title: %s", LCSTR(authors), LCSTR(title));
 
+        // Return possibly multiple <dc:identifier> (identifiers)
+        // as a single doc_props string with values in a key-value format (scheme:identifier) separated by ;
+        bool identifiers_set = false;
+        lString32 identifiers;
+        // Iterate all package/metadata/identifier
+        lUInt16 identifier_id = doc->getElementNameIndex(U"identifier");
+        for (size_t i=0; i<nb_metadata_items; i++) {
+            ldomNode * item = metadata->getChildNode(i);
+            if ( item->getNodeId() != identifier_id )
+                continue;
+            lString32 scheme = item->getAttributeValue(U"scheme");
+            lString32 identifier;
+            // In version 3, scheme is not set but the type is rather included in the text itself
+            if (scheme.empty()) {
+                identifier = item->getText().trim();
+            }
+            else {
+                // In version 2, the scheme is only found as attribute
+                identifier << scheme << ":" << item->getText().trim();
+            }
+            if (identifiers_set) {
+                identifiers << "\n" << identifier;
+            }
+            else {
+                identifiers << identifier;
+                identifiers_set = true;
+            }
+        }
+        m_doc_props->setString(DOC_PROP_IDENTIFIERS, identifiers);
+
         bool hasSeriesMeta = false;
         bool hasSeriesIdMeta = false;
         // Iterate all package/metadata/meta