Read document XML path from _rels/.rels

2 files changed, 51 insertions, 6 deletions

diff --git a/lib/docx/docx-reader.js b/lib/docx/docx-reader.js
index 54f2a3b..67ff11d 100644
--- a/lib/docx/docx-reader.js
+++ b/lib/docx/docx-reader.js
@@ -24,6 +24,7 @@ function read(docxFile, input) {
         contentTypes: readContentTypesFromZipFile(docxFile),
         numbering: readNumberingFromZipFile(docxFile),
         styles: readStylesFromZipFile(docxFile),
+        fileRelationships: readFileRelationships(docxFile),
         docxFile: docxFile,
         files: new Files(input.path ? path.dirname(input.path) : null)
     }).also(function(result) {
@@ -59,8 +60,9 @@ function read(docxFile, input) {
             })
         };
     }).then(function(result) {
-        return readXmlFileWithBody("word/document.xml", result, function(bodyReader, xml) {
-            if (xml) {
+        var documentFilename = findDocumentFilename(docxFile, result.fileRelationships);
+        if (documentFilename) {
+            return readXmlFileWithBody(documentFilename, result, function(bodyReader, xml) {
                 return result.notes.flatMap(function(notes) {
                     return result.comments.flatMap(function(comments) {
                         var reader = new DocumentXmlReader({
@@ -71,13 +73,21 @@ function read(docxFile, input) {
                         return reader.convertXmlToDocument(xml);
                     });
                 });
-            } else {
-                throw new Error("Could not find word/document.xml in ZIP file. Are you sure this is a valid .docx file?");
-            }
-        });
+            });
+        } else {
+            throw new Error("Could not find word/document.xml in ZIP file. Are you sure this is a valid .docx file?");
+        }
     });
 }
 
+function findDocumentFilename(docxFile, relationships) {
+    var targets = ["word/document.xml"].concat(relationships.findTargetsByType("http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument"));
+    var validTargets = targets.filter(function(target) {
+        return docxFile.exists(target);
+    });
+    return validTargets[0];
+}
+
 function xmlFileReader(options) {
     return function(zipFile) {
         return readXmlFromZipFile(zipFile, options.filename)
@@ -133,3 +143,9 @@ var readStylesFromZipFile = xmlFileReader({
     readElement: stylesReader.readStylesXml,
     defaultValue: stylesReader.defaultStyles
 });
+
+var readFileRelationships = xmlFileReader({
+    filename: "_rels/.rels",
+    readElement: relationshipsReader.readRelationships,
+    defaultValue: relationshipsReader.defaultValue
+});
diff --git a/test/docx/docx-reader.tests.js b/test/docx/docx-reader.tests.js
index 283d03a..a25d76e 100644
--- a/test/docx/docx-reader.tests.js
+++ b/test/docx/docx-reader.tests.js
@@ -2,6 +2,7 @@ var assert = require("assert");
 
 var docxReader = require("../../lib/docx/docx-reader");
 var documents = require("../../lib/documents");
+var xml = require("../../lib/xml");
 
 var testing = require("../testing");
 var test = require("../test")(module);
@@ -38,3 +39,31 @@ test("hyperlink hrefs are read from relationships file", function() {
         assert.equal(hyperlink.children.length, 1);
     });
 });
+
+var relationshipNamespaces = {
+    "r": "http://schemas.openxmlformats.org/package/2006/relationships"
+};
+
+test("when document.xml is not present then document is found using _rels/.rels", function() {
+    var relationships = xml.element("r:Relationships", {}, [
+        xml.element("r:Relationship", {
+            "Type": "http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument",
+            "Target": "word/document2.xml"
+        })
+    ]);
+    
+    var docxFile = createFakeDocxFile({
+        "word/document2.xml": testData("simple/word/document.xml"),
+        "_rels/.rels": xml.writeString(relationships, relationshipNamespaces)
+    });
+    var expectedDocument = documents.Document([
+        documents.Paragraph([
+            documents.Run([
+                documents.Text("Hello.")
+            ])
+        ])
+    ]);
+    return docxReader.read(docxFile).then(function(result) {
+        assert.deepEqual(expectedDocument, result.value);
+    });
+});