Merge pull request #4113 from advplyr/parsing-opf-v3

Update opf parser to support refines meta elements

server/scanner/OpfFileScanner.js

@@ -2,24 +2,26 @@ const { parseOpfMetadataXML } = require('../utils/parsers/parseOpfMetadata')
 const { readTextFile } = require('../utils/fileUtils')
 
 class OpfFileScanner {
-  constructor() { }
+  constructor() {}
 
   /**
    * Parse metadata from .opf file found in library scan and update bookMetadata
-   * 
-   * @param {import('../models/LibraryItem').LibraryFileObject} opfLibraryFileObj 
-   * @param {Object} bookMetadata 
+   *
+   * @param {import('../models/LibraryItem').LibraryFileObject} opfLibraryFileObj
+   * @param {Object} bookMetadata
    */
   async scanBookOpfFile(opfLibraryFileObj, bookMetadata) {
     const xmlText = await readTextFile(opfLibraryFileObj.metadata.path)
     const opfMetadata = xmlText ? await parseOpfMetadataXML(xmlText) : null
     if (opfMetadata) {
       for (const key in opfMetadata) {
-        if (key === 'tags') { // Add tags only if tags are empty
+        if (key === 'tags') {
+          // Add tags only if tags are empty
           if (opfMetadata.tags.length) {
             bookMetadata.tags = opfMetadata.tags
           }
-        } else if (key === 'genres') { // Add genres only if genres are empty
+        } else if (key === 'genres') {
+          // Add genres only if genres are empty
           if (opfMetadata.genres.length) {
             bookMetadata.genres = opfMetadata.genres
           }
@@ -42,4 +44,4 @@ class OpfFileScanner {
     }
   }
 }
-module.exports = new OpfFileScanner()
+module.exports = new OpfFileScanner()

server/utils/parsers/parseOpfMetadata.js

@@ -22,11 +22,22 @@ function parseCreators(metadata) {
       Object.keys(c['$'])
         .find((key) => key.startsWith('xmlns:'))
         ?.split(':')[1] || 'opf'
-    return {
+    const creator = {
       value: c['_'],
       role: c['$'][`${namespace}:role`] || null,
       fileAs: c['$'][`${namespace}:file-as`] || null
     }
+
+    const id = c['$']['id']
+    if (id && metadata.meta.refines?.some((r) => r.refines === `#${id}`)) {
+      const creatorMeta = metadata.meta.refines.filter((r) => r.refines === `#${id}`)
+      if (creatorMeta) {
+        creator.role = creatorMeta.find((r) => r.property === 'role')?.value || creator.role || null
+        creator.fileAs = creatorMeta.find((r) => r.property === 'file-as')?.value || creator.fileAs || null
+      }
+    }
+
+    return creator
   })
 }
 
@@ -187,7 +198,6 @@ module.exports.parseOpfMetadataJson = (json) => {
   const prefix = packageKey.split(':').shift()
   let metadata = prefix ? json[packageKey][`${prefix}:metadata`] || json[packageKey].metadata : json[packageKey].metadata
   if (!metadata) return null
-
   if (Array.isArray(metadata)) {
     if (!metadata.length) return null
     metadata = metadata[0]
@@ -198,12 +208,22 @@ module.exports.parseOpfMetadataJson = (json) => {
   metadata.meta = {}
   if (metadataMeta?.length) {
     metadataMeta.forEach((meta) => {
-      if (meta && meta['$'] && meta['$'].name) {
+      if (meta?.['$']?.name) {
         metadata.meta[meta['$'].name] = [meta['$'].content || '']
+      } else if (meta?.['$']?.refines) {
+        // https://www.w3.org/TR/epub-33/#sec-meta-elem
+
+        if (!metadata.meta.refines) {
+          metadata.meta.refines = []
+        }
+        metadata.meta.refines.push({
+          value: meta._,
+          refines: meta['$'].refines,
+          property: meta['$'].property
+        })
       }
     })
   }
-
   const creators = parseCreators(metadata)
   const authors = (fetchCreators(creators, 'aut') || []).map((au) => au?.trim()).filter((au) => au)
   const narrators = (fetchNarrators(creators, metadata) || []).map((nrt) => nrt?.trim()).filter((nrt) => nrt)
@@ -227,5 +247,6 @@ module.exports.parseOpfMetadataJson = (json) => {
 module.exports.parseOpfMetadataXML = async (xml) => {
   const json = await xmlToJSON(xml)
   if (!json) return null
+
   return this.parseOpfMetadataJson(json)
 }

test/server/utils/parsers/parseOpfMetadata.test.js

@@ -3,8 +3,8 @@ const expect = chai.expect
 const { parseOpfMetadataXML } = require('../../../../server/utils/parsers/parseOpfMetadata')
 
 describe('parseOpfMetadata - test series', async () => {
-    it('test one series', async () => {
-        const opf = `
+  it('test one series', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -13,12 +13,12 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([{ "name": "Serie", "sequence": "1" }])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([{ name: 'Serie', sequence: '1' }])
+  })
 
-    it('test more then 1 series - in correct order', async () => {
-        const opf = `
+  it('test more then 1 series - in correct order', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -31,16 +31,16 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([
-            { "name": "Serie 1", "sequence": "1" },
-            { "name": "Serie 2", "sequence": "2" },
-            { "name": "Serie 3", "sequence": "3" },
-        ])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([
+      { name: 'Serie 1', sequence: '1' },
+      { name: 'Serie 2', sequence: '2' },
+      { name: 'Serie 3', sequence: '3' }
+    ])
+  })
 
-    it('test messed order of series content and index', async () => {
-        const opf = `
+  it('test messed order of series content and index', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -52,15 +52,15 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([
-            { "name": "Serie 1", "sequence": "1" },
-            { "name": "Serie 3", "sequence": null },
-        ])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([
+      { name: 'Serie 1', sequence: '1' },
+      { name: 'Serie 3', sequence: null }
+    ])
+  })
 
-    it('test different values of series content and index', async () => {
-        const opf = `
+  it('test different values of series content and index', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -73,16 +73,16 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([
-            { "name": "Serie 1", "sequence": null },
-            { "name": "Serie 2", "sequence": "abc" },
-            { "name": "Serie 3", "sequence": null },
-        ])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([
+      { name: 'Serie 1', sequence: null },
+      { name: 'Serie 2', sequence: 'abc' },
+      { name: 'Serie 3', sequence: null }
+    ])
+  })
 
-    it('test empty series content', async () => {
-        const opf = `
+  it('test empty series content', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -91,12 +91,12 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([])
+  })
 
-    it('test series and index using an xml namespace', async () => {
-        const opf = `
+  it('test series and index using an xml namespace', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <ns0:package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <ns0:metadata>
@@ -105,14 +105,12 @@ describe('parseOpfMetadata - test series', async () => {
               </ns0:metadata>
             </ns0:package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([
-            { "name": "Serie 1", "sequence": null }
-        ])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([{ name: 'Serie 1', sequence: null }])
+  })
 
-    it('test series and series index not directly underneath', async () => {
-        const opf = `
+  it('test series and series index not directly underneath', async () => {
+    const opf = `
             <?xml version='1.0' encoding='UTF-8'?>
             <package xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf" xml:lang="en" version="3.0" unique-identifier="bookid">
               <metadata>
@@ -122,9 +120,21 @@ describe('parseOpfMetadata - test series', async () => {
               </metadata>
             </package>
         `
-        const parsedOpf = await parseOpfMetadataXML(opf)
-        expect(parsedOpf.series).to.deep.equal([
-            { "name": "Serie 1", "sequence": "1" }
-        ])
-    })
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.series).to.deep.equal([{ name: 'Serie 1', sequence: '1' }])
+  })
+
+  it('test author is parsed from refines meta', async () => {
+    const opf = `
+        <package version="3.0" unique-identifier="uuid_id" prefix="rendition: http://www.idpf.org/vocab/rendition/#" xmlns="http://www.idpf.org/2007/opf">
+          <metadata>
+            <dc:creator id="create1">Nevil Shute</dc:creator>
+            <meta refines="#create1" property="role" scheme="marc:relators">aut</meta>
+            <meta refines="#create1" property="file-as">Shute, Nevil</meta>
+          </metadata>
+        </package>
+      `
+    const parsedOpf = await parseOpfMetadataXML(opf)
+    expect(parsedOpf.authors).to.deep.equal(['Nevil Shute'])
+  })
 })