Move tests to test for easier Mocha usage

9 files changed, 1211 insertions, 0 deletions

diff --git a/test/docx/body-reader.tests.js b/test/docx/body-reader.tests.js
new file mode 100644
index 0000000..0a808b0
--- /dev/null
+++ b/test/docx/body-reader.tests.js
@@ -0,0 +1,724 @@
+var assert = require("assert");
+var path = require("path");
+
+var BodyReader = require("../../lib/docx/body-reader").BodyReader;
+var documents = require("../../lib/documents");
+var xml = require("../../lib/xml");
+var XmlElement = xml.Element;
+var Numbering = require("../../lib/docx/numbering-xml").Numbering;
+var Styles = require("../../lib/docx/styles-reader").Styles;
+var warning = require("../../lib/results").warning;
+
+var testing = require("../testing");
+var test = testing.test;
+var createFakeDocxFile = testing.createFakeDocxFile;
+
+function readXmlElement(element, options) {
+    options = Object.create(options || {});
+    options.styles = options.styles || new Styles({}, {});
+    return new BodyReader(options).readXmlElement(element);
+}
+
+function readXmlElementValue(element, options) {
+    var result = readXmlElement(element, options);
+    assert.deepEqual(result.messages, []);
+    return result.value;
+}
+
+var fakeContentTypes = {
+    findContentType: function(filePath) {
+        var extensionTypes = {
+            ".png": "image/png",
+            ".emf": "image/x-emf"
+        };
+        return extensionTypes[path.extname(filePath)];
+    }
+};
+
+describe("readXmlElement: ", function() {
+    test("paragraph has no style if it has no properties", function() {
+        var paragraphXml = new XmlElement("w:p", {}, []);
+        var paragraph = readXmlElementValue(paragraphXml);
+        assert.deepEqual(paragraph.styleId, null);
+    });
+    
+    test("paragraph has style ID and name read from paragraph properties if present", function() {
+        var styleXml = new XmlElement("w:pStyle", {"w:val": "Heading1"}, []);
+        var propertiesXml = new XmlElement("w:pPr", {}, [styleXml]);
+        var paragraphXml = new XmlElement("w:p", {}, [propertiesXml]);
+        
+        var styles = new Styles({"Heading1": {name: "Heading 1"}}, {});
+        
+        var paragraph = readXmlElementValue(paragraphXml, {styles: styles});
+        assert.deepEqual(paragraph.styleId, "Heading1");
+        assert.deepEqual(paragraph.styleName, "Heading 1");
+    });
+    
+    test("warning is emitted when paragraph style cannot be found", function() {
+        var styleXml = new XmlElement("w:pStyle", {"w:val": "Heading1"}, []);
+        var propertiesXml = new XmlElement("w:pPr", {}, [styleXml]);
+        var paragraphXml = new XmlElement("w:p", {}, [propertiesXml]);
+        
+        var styles = new Styles({}, {});
+        
+        var result = readXmlElement(paragraphXml, {styles: styles});
+        var paragraph = result.value;
+        assert.deepEqual(paragraph.styleId, "Heading1");
+        assert.deepEqual(paragraph.styleName, null);
+        assert.deepEqual(result.messages, [warning("Paragraph style with ID Heading1 was referenced but not defined in the document")]);
+    });
+    
+    test("paragraph has justification read from paragraph properties if present", function() {
+        var justificationXml = new XmlElement("w:jc", {"w:val": "center"}, []);
+        var propertiesXml = new XmlElement("w:pPr", {}, [justificationXml]);
+        var paragraphXml = new XmlElement("w:p", {}, [propertiesXml]);
+        var paragraph = readXmlElementValue(paragraphXml);
+        assert.deepEqual(paragraph.alignment, "center");
+    });
+    
+    test("paragraph has numbering properties from paragraph properties if present", function() {
+        var numberingPropertiesXml = new XmlElement("w:numPr", {}, [
+            new XmlElement("w:ilvl", {"w:val": "1"}),
+            new XmlElement("w:numId", {"w:val": "42"})
+        ]);
+        var propertiesXml = new XmlElement("w:pPr", {}, [numberingPropertiesXml]);
+        var paragraphXml = new XmlElement("w:p", {}, [propertiesXml]);
+        
+        var numbering = new Numbering({"42": {"1": {isOrdered: true, level: "1"}}});
+        
+        var paragraph = readXmlElementValue(paragraphXml, {numbering: numbering});
+        assert.deepEqual(paragraph.numbering, {level: "1", isOrdered: true});
+    });
+    
+    test("numbering properties are converted to numbering at specified level", function() {
+        var numberingPropertiesXml = new XmlElement("w:numPr", {}, [
+            new XmlElement("w:ilvl", {"w:val": "1"}),
+            new XmlElement("w:numId", {"w:val": "42"})
+        ]);
+        
+        var numbering = new Numbering({"42": {"1": {isOrdered: true, level: "1"}}});
+        
+        var reader = new BodyReader({
+            numbering: numbering
+        });
+        var numberingLevel = reader._readNumberingProperties(numberingPropertiesXml);
+        assert.deepEqual(numberingLevel, {level: "1", isOrdered: true});
+    });
+    
+    test("numbering properties are ignored if w:ilvl is missing", function() {
+        var numberingPropertiesXml = new XmlElement("w:numPr", {}, [
+            new XmlElement("w:numId", {"w:val": "42"})
+        ]);
+        
+        var numbering = new Numbering({"42": {"1": {isOrdered: true, level: "1"}}});
+        
+        var reader = new BodyReader({
+            numbering: numbering
+        });
+        var numberingLevel = reader._readNumberingProperties(numberingPropertiesXml);
+        assert.equal(numberingLevel, null);
+    });
+    
+    test("numbering properties are ignored if w:numId is missing", function() {
+        var numberingPropertiesXml = new XmlElement("w:numPr", {}, [
+            new XmlElement("w:ilvl", {"w:val": "1"})
+        ]);
+        
+        var numbering = new Numbering({"42": {"1": {isOrdered: true, level: "1"}}});
+        
+        var reader = new BodyReader({
+            numbering: numbering
+        });
+        var numberingLevel = reader._readNumberingProperties(numberingPropertiesXml);
+        assert.equal(numberingLevel, null);
+    });
+    
+    test("run has no style if it has no properties", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.styleId, null);
+    });
+    
+    test("run has style ID and name read from run properties if present", function() {
+        var runStyleXml = new XmlElement("w:rStyle", {"w:val": "Heading1Char"});
+        var runXml = runWithProperties([runStyleXml]);
+        
+        var styles = new Styles({}, {"Heading1Char": {name: "Heading 1 Char"}});
+        
+        var run = readXmlElementValue(runXml, {styles: styles});
+        assert.deepEqual(run.styleId, "Heading1Char");
+        assert.deepEqual(run.styleName, "Heading 1 Char");
+    });
+    
+    test("warning is emitted when run style cannot be found", function() {
+        var runStyleXml = new XmlElement("w:rStyle", {"w:val": "Heading1Char"});
+        var runXml = runWithProperties([runStyleXml]);
+        
+        var styles = new Styles({}, {});
+        
+        var result = readXmlElement(runXml, {styles: styles});
+        var run = result.value;
+        assert.deepEqual(run.styleId, "Heading1Char");
+        assert.deepEqual(run.styleName, null);
+        assert.deepEqual(result.messages, [warning("Run style with ID Heading1Char was referenced but not defined in the document")]);
+    });
+    
+    test("isBold is false if bold element is not present", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.isBold, false);
+    });
+    
+    test("isBold is true if bold element is present", function() {
+        var boldXml = new XmlElement("w:b");
+        var runXml = runWithProperties([boldXml]);
+        var run = readXmlElementValue(runXml);
+        assert.equal(run.isBold, true);
+    });
+
+    test("isUnderline is false if underline element is not present", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.isUnderline, false);
+    });
+
+    test("isUnderline is true if underline element is present", function() {
+        var underlineXml = new XmlElement("w:u");
+        var runXml = runWithProperties([underlineXml]);
+        var run = readXmlElementValue(runXml);
+        assert.equal(run.isUnderline, true);
+    });
+
+    test("isStrikethrough is false if strikethrough element is not present", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.isStrikethrough, false);
+    });
+
+    test("isStrikethrough is true if strikethrough element is present", function() {
+        var strikethroughXml = new XmlElement("w:strike");
+        var runXml = runWithProperties([strikethroughXml]);
+        var run = readXmlElementValue(runXml);
+        assert.equal(run.isStrikethrough, true);
+    });
+
+    test("isItalic is false if bold element is not present", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.isItalic, false);
+    });
+    
+    test("isItalic is true if bold element is present", function() {
+        var italicXml = new XmlElement("w:i");
+        var runXml = runWithProperties([italicXml]);
+        var run = readXmlElementValue(runXml);
+        assert.equal(run.isItalic, true);
+    });
+    
+    test("run has baseline vertical alignment by default", function() {
+        var runXml = runWithProperties([]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.verticalAlignment, documents.verticalAlignment.baseline);
+    });
+    
+    test("run has vertical alignment read from properties", function() {
+        var verticalAlignmentXml = new XmlElement("w:vertAlign", {"w:val": "superscript"});
+        var runXml = runWithProperties([verticalAlignmentXml]);
+        
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run.verticalAlignment, documents.verticalAlignment.superscript);
+    });
+    
+    test("run properties not included as child of run", function() {
+        var runStyleXml = new XmlElement("w:rStyle");
+        var runPropertiesXml = new XmlElement("w:rPr", {}, [runStyleXml]);
+        var runXml = new XmlElement("w:r", {}, [runPropertiesXml]);
+        var result = readXmlElement(runXml);
+        assert.deepEqual(result.value.children, []);
+    });
+    
+    test("w:tab is read as document tab element", function() {
+        var tabXml = new XmlElement("w:tab");
+        var result = readXmlElement(tabXml);
+        assert.deepEqual(result.value, new documents.Tab());
+    });
+    
+    test("w:tbl is read as document table element", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            new XmlElement("w:tr", {}, [
+                new XmlElement("w:tc", {}, [
+                    new XmlElement("w:p", {}, [])
+                ])
+            ])
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            new documents.TableRow([
+                new documents.TableCell([
+                    new documents.Paragraph([])
+                ])
+            ])
+        ]));
+    });
+    
+    test("w:gridSpan is read as colSpan for table cell", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            new XmlElement("w:tr", {}, [
+                new XmlElement("w:tc", {}, [
+                    new XmlElement("w:tcPr", {}, [
+                        new XmlElement("w:gridSpan", {"w:val": "2"})
+                    ]),
+                    new XmlElement("w:p", {}, [])
+                ])
+            ])
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            new documents.TableRow([
+                new documents.TableCell([
+                    new documents.Paragraph([])
+                ], {colSpan: 2})
+            ])
+        ]));
+    });
+    
+    test("w:vMerge is read as rowSpan for table cell", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            row(emptyCell()),
+            row(emptyCell(vMerge("restart"))),
+            row(emptyCell(vMerge("continue"))),
+            row(emptyCell(vMerge("continue"))),
+            row(emptyCell())
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            docRow([docEmptyCell()]),
+            docRow([docEmptyCell({rowSpan: 3})]),
+            docRow([]),
+            docRow([]),
+            docRow([docEmptyCell()])
+        ]));
+    });
+    
+    test("w:vMerge without val is treated as continue", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            row(emptyCell(vMerge("restart"))),
+            row(emptyCell(vMerge()))
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            docRow([docEmptyCell({rowSpan: 2})]),
+            docRow([])
+        ]));
+    });
+    
+    test("w:vMerge accounts for cells spanning columns", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            row(emptyCell(), emptyCell(), emptyCell(vMerge("restart"))),
+            row(emptyCell(gridSpan("2")), emptyCell(vMerge("continue"))),
+            row(emptyCell(), emptyCell(), emptyCell(vMerge("continue"))),
+            row(emptyCell(), emptyCell(), emptyCell())
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            docRow([docEmptyCell(), docEmptyCell(), docEmptyCell({rowSpan: 3})]),
+            docRow([docEmptyCell({colSpan: 2})]),
+            docRow([docEmptyCell(), docEmptyCell()]),
+            docRow([docEmptyCell(), docEmptyCell(), docEmptyCell()])
+        ]));
+    });
+    
+    test("no vertical cell merging if merged cells do not line up", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            row(emptyCell(gridSpan("2"), vMerge("restart"))),
+            row(emptyCell(), emptyCell(vMerge("continue")))
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.value, new documents.Table([
+            docRow([docEmptyCell({colSpan: 2})]),
+            docRow([docEmptyCell(), docEmptyCell()])
+        ]));
+    });
+    
+    test("warning if non-row in table", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            new XmlElement("w:p")
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.messages, [warning("unexpected non-row element in table, cell merging may be incorrect")]);
+    });
+    
+    test("warning if non-cell in table row", function() {
+        var tableXml = new XmlElement("w:tbl", {}, [
+            row(new XmlElement("w:p"))
+        ]);
+        var result = readXmlElement(tableXml);
+        assert.deepEqual(result.messages, [warning("unexpected non-cell element in table row, cell merging may be incorrect")]);
+    });
+    
+    function row() {
+        return new XmlElement("w:tr", {}, Array.prototype.slice.call(arguments));
+    }
+    
+    function emptyCell() {
+        return new XmlElement("w:tc", {}, [
+            new XmlElement("w:tcPr", {}, Array.prototype.slice.call(arguments))
+        ]);
+    }
+    
+    function vMerge(val) {
+        return new XmlElement("w:vMerge", {"w:val": val}, []);
+    }
+    
+    function gridSpan(val) {
+        return new XmlElement("w:gridSpan", {"w:val": val});
+    }
+    
+    function docRow(children) {
+        return new documents.TableRow(children);
+    }
+    
+    function docEmptyCell(properties) {
+        return new documents.TableCell([], properties);
+    }
+    
+    test("emits warning on unrecognised element", function() {
+        var unrecognisedElement = new XmlElement("w:not-an-element");
+        var result = readXmlElement(unrecognisedElement);
+        assert.deepEqual(
+            result.messages,
+            [{
+                type: "warning",
+                message: "An unrecognised element was ignored: w:not-an-element"
+            }]
+        );
+    });
+
+    test("w:bookmarkStart is read as a bookmarkStart", function() {
+        var bookmarkStart = new XmlElement("w:bookmarkStart", {"w:name": "_Peter", "w:id": "42"});
+        var result = readXmlElement(bookmarkStart);
+        assert.deepEqual(result.value.name, "_Peter");
+        assert.deepEqual(result.value.type, "bookmarkStart");
+    });
+    
+    test('_GoBack bookmark is ignored', function() {
+        var bookmarkStart = new XmlElement("w:bookmarkStart", {"w:name": "_GoBack"});
+        var result = readXmlElement(bookmarkStart);
+        assert.deepEqual(result.value, []);
+    });
+    
+    test("can read imagedata elements with r:id attribute", function() {
+        var imagedataElement = new XmlElement("v:imagedata", {"r:id": "rId5", "o:title": "It's a hat"});
+        
+        var imageBuffer = new Buffer("Not an image at all!");
+        var reader = new BodyReader({
+            relationships: {
+                "rId5": {target: "media/hat.png"}
+            },
+            contentTypes: fakeContentTypes,
+            docxFile: createFakeDocxFile({
+                "word/media/hat.png": imageBuffer
+            })
+        });
+        var result = reader.readXmlElement(imagedataElement);
+        assert.deepEqual(result.messages, []);
+        var element = result.value;
+        assert.equal("image", element.type);
+        assert.equal(element.altText, "It's a hat");
+        assert.equal(element.contentType, "image/png");
+        return element.read()
+            .then(function(readValue) {
+                assert.equal(readValue, imageBuffer);
+            });
+    });
+    
+    test("can read inline pictures", function() {
+        var drawing = createInlineImage({
+            blip: createEmbeddedBlip("rId5"),
+            description: "It's a hat"
+        });
+        
+        var imageBuffer = new Buffer("Not an image at all!");
+        var reader = new BodyReader({
+            relationships: {
+                "rId5": {target: "media/hat.png"}
+            },
+            contentTypes: fakeContentTypes,
+            docxFile: createFakeDocxFile({
+                "word/media/hat.png": imageBuffer
+            })
+        });
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, []);
+        var element = single(result.value);
+        assert.equal("image", element.type);
+        assert.equal(element.altText, "It's a hat");
+        assert.equal(element.contentType, "image/png");
+        return element.read()
+            .then(function(readValue) {
+                assert.equal(readValue, imageBuffer);
+            });
+    });
+    
+    test("can read anchored pictures", function() {
+        var drawing = new XmlElement("w:drawing", {}, [
+            new XmlElement("wp:anchor", {}, [
+                new XmlElement("wp:docPr", {descr: "It's a hat"}),
+                new XmlElement("a:graphic", {}, [
+                    new XmlElement("a:graphicData", {}, [
+                        new XmlElement("pic:pic", {}, [
+                            new XmlElement("pic:blipFill", {}, [
+                                new XmlElement("a:blip", {"r:embed": "rId5"})
+                            ])
+                        ])
+                    ])
+                ])
+            ])
+        ]);
+        
+        var imageBuffer = new Buffer("Not an image at all!");
+        var reader = new BodyReader({
+            relationships: {
+                "rId5": {target: "media/hat.png"}
+            },
+            contentTypes: fakeContentTypes,
+            docxFile: createFakeDocxFile({
+                "word/media/hat.png": imageBuffer
+            })
+        });
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, []);
+        var element = single(result.value);
+        assert.equal("image", element.type);
+        assert.equal(element.altText, "It's a hat");
+        return element.read()
+            .then(function(readValue) {
+                assert.equal(readValue, imageBuffer);
+            });
+    });
+    
+    test("can read linked pictures", function() {
+        var drawing = createInlineImage({
+            blip: createLinkedBlip("rId5"),
+            description: "It's a hat"
+        });
+        
+        var imageBuffer = new Buffer("Not an image at all!");
+        var reader = new BodyReader({
+            relationships: {
+                "rId5": {target: "file:///media/hat.png"}
+            },
+            contentTypes: fakeContentTypes,
+            files: testing.createFakeFiles({
+                "file:///media/hat.png": imageBuffer
+            })
+        });
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, []);
+        var element = single(result.value);
+        assert.equal("image", element.type);
+        assert.equal(element.altText, "It's a hat");
+        assert.equal(element.contentType, "image/png");
+        return element.read()
+            .then(function(readValue) {
+                assert.equal(readValue, imageBuffer);
+            });
+    });
+    
+    test("warning if unsupported image type", function() {
+        var drawing = createInlineImage({
+            blip: createEmbeddedBlip("rId5"),
+            description: "It's a hat"
+        });
+        
+        var imageBuffer = new Buffer("Not an image at all!");
+        var reader = new BodyReader({
+            relationships: {
+                "rId5": {target: "media/hat.emf"}
+            },
+            contentTypes: fakeContentTypes,
+            docxFile: createFakeDocxFile({
+                "word/media/hat.emf": imageBuffer
+            })
+        });
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, [warning("Image of type image/x-emf is unlikely to display in web browsers")]);
+        var element = single(result.value);
+        assert.equal(element.contentType, "image/x-emf");
+    });
+    
+    test("no elements created if image cannot be found in w:drawing", function() {
+        var drawing = new XmlElement("w:drawing", {}, []);
+        
+        var reader = new BodyReader({});
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, []);
+        assert.deepEqual(result.value, []);
+    });
+    
+    test("no elements created if image cannot be found in wp:inline", function() {
+        var drawing = new XmlElement("wp:inline", {}, []);
+        
+        var reader = new BodyReader({});
+        var result = reader.readXmlElement(drawing);
+        assert.deepEqual(result.messages, []);
+        assert.deepEqual(result.value, []);
+    });
+    
+    test("children of w:ins are converted normally", function() {
+        var runXml = new XmlElement("w:r", {}, []);
+        var insXml = new XmlElement("w:ins", {}, [runXml]);
+        var result = readXmlElement(insXml);
+        assert.deepEqual(result.value[0].type, "run");
+    });
+    
+    test("children of w:smartTag are converted normally", function() {
+        var runXml = new XmlElement("w:r", {}, []);
+        var smartTagXml = new XmlElement("w:smartTag", {}, [runXml]);
+        var result = readXmlElement(smartTagXml);
+        assert.deepEqual(result.value[0].type, "run");
+    });
+    
+    test("w:hyperlink is read as document hyperlink if it has a relationship ID", function() {
+        var runXml = new XmlElement("w:r", {}, []);
+        var hyperlinkXml = new XmlElement("w:hyperlink", {"r:id": "r42"}, [runXml]);
+        var relationships = {
+            "r42": {target: "http://example.com"}
+        };
+        var result = readXmlElement(hyperlinkXml, {relationships: relationships});
+        assert.deepEqual(result.value.href, "http://example.com");
+        assert.deepEqual(result.value.children[0].type, "run");
+    });
+
+    test("w:hyperlink is read as document hyperlink if it has an anchor", function() {
+        var runXml = new XmlElement("w:r", {}, []);
+        var hyperlinkXml = new XmlElement("w:hyperlink", {"w:anchor": "_Peter"}, [runXml]);
+        var result = readXmlElement(hyperlinkXml);
+        assert.deepEqual(result.value.anchor, "_Peter");
+        assert.deepEqual(result.value.children[0].type, "run");
+    });
+
+    test("w:hyperlink is ignored if it does not have a relationship ID nor anchor", function() {
+        var runXml = new XmlElement("w:r", {}, []);
+        var hyperlinkXml = new XmlElement("w:hyperlink", {}, [runXml]);
+        var result = readXmlElement(hyperlinkXml);
+        assert.deepEqual(result.value[0].type, "run");
+    });
+    
+    test("w:br is read as line break", function() {
+        var breakXml = new XmlElement("w:br", {}, []);
+        var result = readXmlElement(breakXml);
+        assert.deepEqual(result.value.type, "lineBreak");
+        assert.deepEqual(result.messages, []);
+    });
+    
+    test("w:bookmarkEnd is ignored without warning", function() {
+        var ignoredElement = new XmlElement("w:bookmarkEnd");
+        var result = readXmlElement(ignoredElement);
+        assert.deepEqual(result.messages, []);
+        assert.deepEqual([], result.value);
+    });
+    
+    test("warning on breaks that aren't line breaks", function() {
+        var breakXml = new XmlElement("w:br", {"w:type": "page"}, []);
+        var result = readXmlElement(breakXml);
+        assert.deepEqual(result.value, []);
+        assert.deepEqual(result.messages, [warning("Unsupported break type: page")]);
+    });
+
+    test("text boxes have content appended after containing paragraph", function() {
+        var textbox = new XmlElement("w:pict", {}, [
+            new XmlElement("v:shape", {}, [
+                new XmlElement("v:textbox", {}, [
+                    new XmlElement("w:txbxContent", {}, [
+                        paragraphWithStyleId("textbox-content")
+                    ])
+                ])
+            ])
+        ]);
+        var paragraph = new XmlElement("w:p", {}, [
+            new XmlElement("w:r", {}, [textbox])
+        ]);
+        var result = readXmlElement(paragraph);
+        assert.deepEqual(result.value[1].styleId, "textbox-content");
+    });
+
+    test("mc:Fallback is used when mc:AlternateContent is read", function() {
+        var styles = new Styles({"first": {name: "First"}, "second": {name: "Second"}}, {});
+        var textbox = new XmlElement("mc:AlternateContent", {}, [
+            new XmlElement("mc:Choice", {"Requires": "wps"}, [
+                paragraphWithStyleId("first")
+            ]),
+            new XmlElement("mc:Fallback", {}, [
+                paragraphWithStyleId("second")
+            ])
+        ]);
+        var result = readXmlElement(textbox, {styles: styles});
+        assert.deepEqual(result.value[0].styleId, "second");
+    });
+
+    test("w:sdtContent is used when w:sdt is read", function() {
+        var element = xml.element("w:sdt", {}, [
+            xml.element("w:sdtContent", {}, [
+                xml.element("w:t", {}, [xml.text("Blackdown")])
+            ])
+        ]);
+        var result = readXmlElement(element);
+        assert.deepEqual(result.value, [new documents.Text("Blackdown")]);
+    });
+    
+    test("text nodes are ignored when reading children", function() {
+        var runXml = new XmlElement("w:r", {}, [xml.text("[text]")]);
+        var run = readXmlElementValue(runXml);
+        assert.deepEqual(run, new documents.Run([]));
+    });
+});
+
+function paragraphWithStyleId(styleId) {
+    return new XmlElement("w:p", {}, [
+        new XmlElement("w:pPr", {}, [
+            new XmlElement("w:pStyle", {"w:val": styleId}, [])
+        ])
+    ]);
+}
+
+function runWithProperties(children) {
+    return new XmlElement("w:r", {}, [createRunPropertiesXml(children)]);
+}
+
+function createRunPropertiesXml(children) {
+    return new XmlElement("w:rPr", {}, children);
+}
+
+function single(array) {
+    if (array.length === 1) {
+        return array[0];
+    } else {
+        throw new Error("Array has " + array.length + " elements");
+    }
+}
+
+function createInlineImage(options) {
+    return new XmlElement("w:drawing", {}, [
+        new XmlElement("wp:inline", {}, [
+            new XmlElement("wp:docPr", {descr: options.description}),
+            new XmlElement("a:graphic", {}, [
+                new XmlElement("a:graphicData", {}, [
+                    new XmlElement("pic:pic", {}, [
+                        new XmlElement("pic:blipFill", {}, [
+                            options.blip
+                        ])
+                    ])
+                ])
+            ])
+        ])
+    ]);
+}
+
+function createEmbeddedBlip(relationshipId) {
+    return new XmlElement("a:blip", {"r:embed": relationshipId});
+}
+
+function createLinkedBlip(relationshipId) {
+    return new XmlElement("a:blip", {"r:link": relationshipId});
+}
diff --git a/test/docx/content-types-reader.tests.js b/test/docx/content-types-reader.tests.js
new file mode 100644
index 0000000..a223d09
--- /dev/null
+++ b/test/docx/content-types-reader.tests.js
@@ -0,0 +1,47 @@
+var assert = require("assert");
+
+var readContentTypesFromXml = require("../../lib/docx/content-types-reader").readContentTypesFromXml;
+var XmlElement = require("../../lib/xml").Element;
+var test = require("../testing").test;
+
+
+describe('readContentTypesFromXml', function() {
+    test('reads default-per-extension from XML', function() {
+        var contentTypes = readContentTypesFromXml({
+            root: new XmlElement("content-types:Types", {}, [
+                new XmlElement("content-types:Default", {Extension: "png", ContentType: "image/png"})
+            ])
+        });
+        assert.equal(contentTypes.findContentType("word/media/hat.png"), "image/png");
+    });
+    
+    test('reads overrides in preference to defaults', function() {
+        var contentTypes = readContentTypesFromXml({
+            root: new XmlElement("content-types:Types", {}, [
+                new XmlElement("content-types:Default", {Extension: "png", ContentType: "image/png"}),
+                new XmlElement("content-types:Override", {PartName: "/word/media/hat.png", ContentType: "image/hat"})
+            ])
+        });
+        assert.equal(contentTypes.findContentType("word/media/hat.png"), "image/hat");
+    });
+    
+    test('fallback content types have common image types', function() {
+        var contentTypes = readContentTypesFromXml({
+            root: new XmlElement("content-types:Types", {}, [])
+        });
+        assert.equal(contentTypes.findContentType("word/media/hat.png"), "image/png");
+        assert.equal(contentTypes.findContentType("word/media/hat.gif"), "image/gif");
+        assert.equal(contentTypes.findContentType("word/media/hat.jpg"), "image/jpeg");
+        assert.equal(contentTypes.findContentType("word/media/hat.jpeg"), "image/jpeg");
+        assert.equal(contentTypes.findContentType("word/media/hat.bmp"), "image/bmp");
+        assert.equal(contentTypes.findContentType("word/media/hat.tif"), "image/tiff");
+        assert.equal(contentTypes.findContentType("word/media/hat.tiff"), "image/tiff");
+    });
+    
+    test('fallback content types are case insensitive on extension', function() {
+        var contentTypes = readContentTypesFromXml({
+            root: new XmlElement("content-types:Types", {}, [])
+        });
+        assert.equal(contentTypes.findContentType("word/media/hat.PnG"), "image/png");
+    });
+});
diff --git a/test/docx/docx-reader.tests.js b/test/docx/docx-reader.tests.js
new file mode 100644
index 0000000..69ade7e
--- /dev/null
+++ b/test/docx/docx-reader.tests.js
@@ -0,0 +1,42 @@
+var assert = require("assert");
+
+var docxReader = require("../../lib/docx/docx-reader");
+var documents = require("../../lib/documents");
+
+var testing = require("../testing");
+var test = testing.test;
+var testData = testing.testData;
+var createFakeDocxFile = testing.createFakeDocxFile;
+
+
+describe("docx-reader", function() {
+    test("can read document with single paragraph with single run of text", function() {
+        var expectedDocument = documents.Document([
+            documents.Paragraph([
+                documents.Run([
+                    documents.Text("Hello.")
+                ])
+            ])
+        ]);
+        var docxFile = createFakeDocxFile({
+            "word/document.xml": testData("simple/word/document.xml")
+        });
+        return docxReader.read(docxFile).then(function(result) {
+            assert.deepEqual(expectedDocument, result.value);
+        });
+    });
+    
+    test("hyperlink hrefs are read from relationships file", function() {
+        var docxFile = createFakeDocxFile({
+            "word/document.xml": testData("hyperlinks/word/document.xml"),
+            "word/_rels/document.xml.rels": testData("hyperlinks/word/_rels/document.xml.rels")
+        });
+        return docxReader.read(docxFile).then(function(result) {
+            var paragraph = result.value.children[0];
+            assert.equal(1, paragraph.children.length);
+            var hyperlink = paragraph.children[0];
+            assert.equal(hyperlink.href, "http://www.example.com");
+            assert.equal(hyperlink.children.length, 1);
+        });
+    });
+});
diff --git a/test/docx/files.tests.js b/test/docx/files.tests.js
new file mode 100644
index 0000000..2a0919c
--- /dev/null
+++ b/test/docx/files.tests.js
@@ -0,0 +1,94 @@
+var path = require("path");
+var fs = require("fs");
+var assert = require("assert");
+
+var promises = require("../../lib/promises");
+var Files = require("../../lib/docx/files").Files;
+var uriToPath = require("../../lib/docx/files").uriToPath;
+
+var testing = require("../testing");
+var test = testing.test;
+
+var readFile = promises.promisify(fs.readFile.bind(fs));
+
+
+describe("Files", function() {
+    test("can open files with file URI", function() {
+        var filePath = path.resolve(testing.testPath("tiny-picture.png"));
+        var files = new Files(null);
+        return files.read("file:///" + filePath.replace(/^\//, ""), "base64").then(function(contents) {
+            return readFile(filePath, "base64").then(function(expectedContents) {
+                assert.deepEqual(contents, expectedContents);
+            });
+        });
+    });
+    
+    test("can open files with relative URI", function() {
+        var filePath = path.resolve(testing.testPath("tiny-picture.png"));
+        var files = new Files(testing.testPath("."));
+        return files.read("tiny-picture.png", "base64").then(function(contents) {
+            return readFile(filePath, "base64").then(function(expectedContents) {
+                assert.deepEqual(contents, expectedContents);
+            });
+        });
+    });
+    
+    test("given base is not set when opening relative uri then error is raised", function() {
+        var files = new Files(null);
+        return assertError(files.read("not-a-real-file.png", "base64"), function(err) {
+            assert.equal(err.message, "could not find external image 'not-a-real-file.png', path of input document is unknown");
+        });
+    });
+    
+    test("error if relative uri cannot be opened", function() {
+        var files = new Files("/tmp");
+        return assertError(files.read("not-a-real-file.png", "base64"), function(err) {
+            assertRegex(err.message, /could not open external image: 'not-a-real-file.png' \(document directory: '\/tmp'\)\nENOENT.*\/tmp\/not-a-real-file.png.*/);
+        });
+    });
+});
+
+function assertError(promise, func) {
+    return promise.then(function() {
+        assert(false, "Expected error");
+    }, func);
+}
+
+function assertRegex(actual, expected) {
+    assert.ok(expected.test(actual), "Expected regex: " + expected + "\nbut was: " + actual);
+}
+
+
+describe("uriToPath", function() {
+    test("leading slash is retained on non-Windows file URIs", function() {
+        assert.equal(uriToPath("file:///a/b/c", "linux"), "/a/b/c");
+        assert.equal(uriToPath("file:///a/b/c", "win32"), "/a/b/c");
+    });
+    
+    test("URI is unquoted", function() {
+        assert.equal(uriToPath("file:///a%20b"), "/a b");
+    });
+    
+    test("when host is set to localhost then path can be found", function() {
+        assert.equal(uriToPath("file://localhost/a/b/c"), "/a/b/c");
+    });
+    
+    test("when host is set but not localhost then path cannot be found", function() {
+        assert.throws(function() {
+            uriToPath("file://example/a/b/c");
+        }, /Could not convert URI to path: file:\/\/example\/a\/b\/c/);
+    });
+    
+    test("leading slash is not dropped on Windows file URIs when platform is not Windows", function() {
+        assert.equal(uriToPath("file:///c:/a", "linux"), "/c:/a");
+    });
+    
+    test("leading slash is dropped on Windows file URIs when platform is Windows", function() {
+        assert.equal(uriToPath("file:///c:/a", "win32"), "c:/a");
+        assert.equal(uriToPath("file:///C:/a", "win32"), "C:/a");
+    });
+    
+    test("relative URI is unquoted", function() {
+        assert.equal(uriToPath("a%20b/c"), "a b/c");
+    });
+});
diff --git a/test/docx/notes-reader.tests.js b/test/docx/notes-reader.tests.js
new file mode 100644
index 0000000..79ffed3
--- /dev/null
+++ b/test/docx/notes-reader.tests.js
@@ -0,0 +1,39 @@
+var assert = require("assert");
+
+var createFootnotesReader = require("../../lib/docx/notes-reader").createFootnotesReader;
+var BodyReader = require("../../lib/docx/body-reader").BodyReader;
+var documents = require("../../lib/documents");
+var XmlElement = require("../../lib/xml").Element;
+var test = require("../testing").test;
+
+
+describe('readFootnotesXml', function() {
+    test('ID and body of footnote are read', function() {
+        var bodyReader = new BodyReader({});
+        var footnoteBody = [new XmlElement("w:p", {}, [])];
+        var footnotes = createFootnotesReader(bodyReader)({
+            root: new XmlElement("w:footnotes", {}, [
+                new XmlElement("w:footnote", {"w:id": "1"}, footnoteBody)
+            ])
+        });
+        assert.equal(footnotes.value.length, 1);
+        assert.deepEqual(footnotes.value[0].body, [new documents.Paragraph([])]);
+        assert.deepEqual(footnotes.value[0].noteId, "1");
+    });
+});
+
+footnoteTypeIsIgnored('continuationSeparator');
+footnoteTypeIsIgnored('separator');
+
+function footnoteTypeIsIgnored(type) {
+    describe('footnotes of type ' + type + ' are ignored', function() {
+        test('ID and body of footnote are read', function() {
+            var footnotes = createFootnotesReader()({
+                root: new XmlElement("w:footnotes", {}, [
+                    new XmlElement("w:footnote", {"w:id": "1", "w:type": type}, [])
+                ])
+            });
+            assert.equal(footnotes.value.length, 0);
+        });
+    });
+}
diff --git a/test/docx/numbering-xml.tests.js b/test/docx/numbering-xml.tests.js
new file mode 100644
index 0000000..2fea5fb
--- /dev/null
+++ b/test/docx/numbering-xml.tests.js
@@ -0,0 +1,32 @@
+var duck = require("duck");
+
+var readNumberingXml = require("../../lib/docx/numbering-xml").readNumberingXml;
+var XmlElement = require("../../lib/xml").Element;
+var test = require("../testing").test;
+
+
+describe('readNumberingXml', function() {
+    test('w:num element inherits levels from w:abstractNum', function() {
+        var numbering = readNumberingXml({
+            root: new XmlElement("w:numbering", {}, [
+                new XmlElement("w:abstractNum", {"w:abstractNumId": "42"}, [
+                    new XmlElement("w:lvl", {"w:ilvl": "0"}, [
+                        new XmlElement("w:numFmt", {"w:val": "bullet"})
+                    ]),
+                    new XmlElement("w:lvl", {"w:ilvl": "1"}, [
+                        new XmlElement("w:numFmt", {"w:val": "decimal"})
+                    ])
+                ]),
+                new XmlElement("w:num", {"w:numId": "47"}, [
+                    new XmlElement("w:abstractNumId", {"w:val": "42"})
+                ])
+            ])
+        });
+        duck.assertThat(numbering.findLevel("47", "0"), duck.hasProperties({
+            isOrdered: false
+        }));
+        duck.assertThat(numbering.findLevel("47", "1"), duck.hasProperties({
+            isOrdered: true
+        }));
+    });
+});
diff --git a/test/docx/office-xml-reader.tests.js b/test/docx/office-xml-reader.tests.js
new file mode 100644
index 0000000..9d202ba
--- /dev/null
+++ b/test/docx/office-xml-reader.tests.js
@@ -0,0 +1,26 @@
+var assert = require("assert");
+
+var xml = require("../../lib/xml");
+var officeXmlReader = require("../../lib/docx/office-xml-reader");
+var test = require("../testing").test;
+
+
+describe("office-xml-reader", function() {
+    test("mc:AlternateContent is replaced by contents of mc:Fallback", function() {
+        var xmlString =
+            '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>' +
+            '<numbering xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006">' +
+            '<mc:AlternateContent>' +
+            '<mc:Choice Requires="w14">' +
+            '<choice/>' +
+            '</mc:Choice>' +
+            '<mc:Fallback>' +
+            '<fallback/>' +
+            '</mc:Fallback>' +
+            '</mc:AlternateContent>' +
+            '</numbering>';
+        return officeXmlReader.read(xmlString).then(function(element) {
+            assert.deepEqual(element.root.children, [xml.element("fallback")]);
+        });
+    });
+});
diff --git a/test/docx/style-map.tests.js b/test/docx/style-map.tests.js
new file mode 100644
index 0000000..55171f8
--- /dev/null
+++ b/test/docx/style-map.tests.js
@@ -0,0 +1,115 @@
+var assert = require("assert");
+
+var JSZip = require("jszip");
+
+var zipfile = require("../../lib/zipfile");
+var styleMap = require("../../lib/docx/style-map");
+var test = require("../testing").test;
+
+describe("style-map", function() {
+    test('reading embedded style map on document without embedded style map returns null', function() {
+        var zip = normalDocx();
+        
+        return styleMap.readStyleMap(zip).then(function(contents) {
+            assert.equal(contents, null);
+        });
+    });
+    
+    test('embedded style map can be read after being written', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return styleMap.readStyleMap(zip).then(function(contents) {
+                assert.equal(contents, "p => h1");
+            });
+        });
+    });
+    
+    test('embedded style map is written to separate file', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return zip.read("mammoth/style-map", "utf8").then(function(contents) {
+                assert.equal(contents, "p => h1");
+            });
+        });
+    });
+    
+    test('embedded style map is referenced in relationships', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return zip.read("word/_rels/document.xml.rels", "utf8").then(function(contents) {
+                assert.equal(contents, expectedRelationshipsXml);
+            });
+        });
+    });
+    
+    test('re-embedding style map replaces original', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return styleMap.writeStyleMap(zip, "p => h2");
+        }).then(function() {
+            return zip.read("word/_rels/document.xml.rels", "utf8").then(function(contents) {
+                assert.equal(contents, expectedRelationshipsXml);
+            });
+        }).then(function() {
+            return styleMap.readStyleMap(zip).then(function(contents) {
+                assert.equal(contents, "p => h2");
+            });
+        });
+    });
+    
+    test('embedded style map has override content type in [Content_Types].xml', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return zip.read("[Content_Types].xml", "utf8").then(function(contents) {
+                assert.equal(contents, expectedContentTypesXml);
+            });
+        });
+    });
+    
+    test('replacing style map keeps content type', function() {
+        var zip = normalDocx();
+        
+        return styleMap.writeStyleMap(zip, "p => h1").then(function() {
+            return styleMap.writeStyleMap(zip, "p => h2");
+        }).then(function() {
+            return zip.read("[Content_Types].xml", "utf8").then(function(contents) {
+                assert.equal(contents, expectedContentTypesXml);
+            });
+        });
+    });
+    
+});
+
+var expectedRelationshipsXml = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>' +
+    '<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">' +
+    '<Relationship Id="rId3" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/settings" Target="settings.xml"/>' +
+    '<Relationship Id="rMammothStyleMap" Type="http://schemas.zwobble.org/mammoth/style-map" Target="/mammoth/style-map"/>' +
+    '</Relationships>';
+
+var expectedContentTypesXml = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>' +
+    '<Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">' +
+    '<Default Extension="png" ContentType="image/png"/>' +
+    '<Override PartName="/mammoth/style-map" ContentType="text/prs.mammoth.style-map"/>' +
+    '</Types>';
+
+function normalDocx() {
+    var zip = new JSZip();
+    var originalRelationshipsXml = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>' +
+        '<Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">' +
+        '<Relationship Id="rId3" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/settings" Target="settings.xml"/>' +
+        '</Relationships>';
+    var originalContentTypesXml = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?>' +
+        '<Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">' +
+        '<Default Extension="png" ContentType="image/png"/>' +
+        '</Types>';
+    zip.file("word/_rels/document.xml.rels", originalRelationshipsXml);
+    zip.file("[Content_Types].xml", originalContentTypesXml);
+    var buffer = zip.generate({type: "arraybuffer"});
+    return zipfile.openArrayBuffer(buffer);
+}
+
diff --git a/test/docx/styles-reader.tests.js b/test/docx/styles-reader.tests.js
new file mode 100644
index 0000000..4ff19c7
--- /dev/null
+++ b/test/docx/styles-reader.tests.js
@@ -0,0 +1,92 @@
+var assert = require("assert");
+
+var readStylesXml = require("../../lib/docx/styles-reader").readStylesXml;
+var XmlElement = require("../../lib/xml").Element;
+var test = require("../testing").test;
+
+
+describe('readStylesXml', function() {
+    test('paragraph style is null if no style with that ID exists', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [])
+        });
+        assert.equal(styles.findParagraphStyleById("Heading1"), null);
+    });
+    
+    test('paragraph style can be found by ID', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                paragraphStyleElement("Heading1", "Heading 1")
+            ])
+        });
+        assert.equal(styles.findParagraphStyleById("Heading1").styleId, "Heading1");
+    });
+    
+    test('character style can be found by ID', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                characterStyleElement("Heading1Char", "Heading 1 Char")
+            ])
+        });
+        assert.equal(styles.findCharacterStyleById("Heading1Char").styleId, "Heading1Char");
+    });
+    
+    test('paragraph and character styles are distinct', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                paragraphStyleElement("Heading1", "Heading 1"),
+                characterStyleElement("Heading1Char", "Heading 1 Char")
+            ])
+        });
+        assert.equal(styles.findCharacterStyleById("Heading1"), null);
+        assert.equal(styles.findParagraphStyleById("Heading1Char"), null);
+    });
+    
+    test('character and table styles are distinct', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                styleElement("table", "Heading1", "Heading 1")
+            ])
+        });
+        assert.equal(styles.findCharacterStyleById("Heading1"), null);
+    });
+    
+    test('styles include names', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                paragraphStyleElement("Heading1", "Heading 1")
+            ])
+        });
+        assert.equal(styles.findParagraphStyleById("Heading1").name, "Heading 1");
+    });
+
+    test('style name is null if w:name element does not exist', function() {
+        var styles = readStylesXml({
+            root: new XmlElement("w:styles", {}, [
+                styleWithoutWNameElement("paragraph", "Heading1"),
+                styleWithoutWNameElement("character", "Heading1Char")
+            ])
+        });
+        assert.equal(styles.findParagraphStyleById("Heading1").name, null);
+        assert.equal(styles.findCharacterStyleById("Heading1Char").name, null);
+    });
+
+});
+
+function paragraphStyleElement(id, name) {
+    return styleElement("paragraph", id, name);
+}
+
+function characterStyleElement(id, name) {
+    return styleElement("character", id, name);
+}
+
+function styleElement(type, id, name) {
+    return new XmlElement("w:style", {"w:type": type, "w:styleId": id}, [
+        new XmlElement("w:name", {"w:val": name}, [])
+    ]);
+}
+
+function styleWithoutWNameElement(type, id) {
+    return new XmlElement("w:style", {"w:type": type, "w:styleId": id}, []);
+}