Read defaults from content types XML

7 files changed, 143 insertions, 2 deletions

diff --git a/src/main/java/org/zwobble/mammoth/docx/ContentTypes.java b/src/main/java/org/zwobble/mammoth/docx/ContentTypes.java
new file mode 100644
index 0000000..18aa8e4
--- /dev/null
+++ b/src/main/java/org/zwobble/mammoth/docx/ContentTypes.java
@@ -0,0 +1,19 @@
+package org.zwobble.mammoth.docx;
+
+import com.google.common.io.Files;
+
+import java.util.Map;
+import java.util.Optional;
+
+public class ContentTypes {
+    private final Map<String, String> extensionDefaults;
+
+    public ContentTypes(Map<String, String> extensionDefaults) {
+        this.extensionDefaults = extensionDefaults;
+    }
+
+    public Optional<String> findContentType(String path) {
+        String extension = Files.getFileExtension(path);
+        return Optional.ofNullable(extensionDefaults.get(extension));
+    }
+}
diff --git a/src/main/java/org/zwobble/mammoth/docx/ContentTypesXml.java b/src/main/java/org/zwobble/mammoth/docx/ContentTypesXml.java
new file mode 100644
index 0000000..6ecf8e2
--- /dev/null
+++ b/src/main/java/org/zwobble/mammoth/docx/ContentTypesXml.java
@@ -0,0 +1,25 @@
+package org.zwobble.mammoth.docx;
+
+import org.zwobble.mammoth.xml.XmlElement;
+import org.zwobble.mammoth.xml.XmlElementList;
+
+import java.util.AbstractMap;
+import java.util.Map;
+
+import static org.zwobble.mammoth.util.MammothMaps.toMap;
+
+public class ContentTypesXml {
+    public static ContentTypes readContentTypesXmlElement(XmlElement element) {
+        return new ContentTypes(readDefaults(element.findChildren("content-types:Default")));
+    }
+
+    private static Map<String, String> readDefaults(XmlElementList children) {
+        return toMap(children, ContentTypesXml::readDefault);
+    }
+
+    private static Map.Entry<String, String> readDefault(XmlElement element) {
+        return new AbstractMap.SimpleImmutableEntry<>(
+            element.getAttribute("Extension"),
+            element.getAttribute("ContentType"));
+    }
+}
diff --git a/src/main/java/org/zwobble/mammoth/util/MammothMaps.java b/src/main/java/org/zwobble/mammoth/util/MammothMaps.java
new file mode 100644
index 0000000..6e988dd
--- /dev/null
+++ b/src/main/java/org/zwobble/mammoth/util/MammothMaps.java
@@ -0,0 +1,16 @@
+package org.zwobble.mammoth.util;
+
+import com.google.common.collect.ImmutableMap;
+
+import java.util.Map;
+import java.util.function.Function;
+
+public class MammothMaps {
+    public static <T, K, V> Map<K, V> toMap(Iterable<T> iterable, Function<T, Map.Entry<K, V>> function) {
+        ImmutableMap.Builder<K, V> builder = ImmutableMap.builder();
+        for (T element : iterable) {
+            builder.put(function.apply(element));
+        }
+        return builder.build();
+    }
+}
diff --git a/src/main/java/org/zwobble/mammoth/xml/XmlElement.java b/src/main/java/org/zwobble/mammoth/xml/XmlElement.java
index f138917..a42a1b7 100644
--- a/src/main/java/org/zwobble/mammoth/xml/XmlElement.java
+++ b/src/main/java/org/zwobble/mammoth/xml/XmlElement.java
@@ -1,10 +1,14 @@
 package org.zwobble.mammoth.xml;
 
+import com.google.common.collect.ImmutableList;
+import com.google.common.collect.ImmutableMap;
+import com.google.common.collect.Iterables;
+
 import java.util.List;
 import java.util.Map;
+import java.util.Optional;
 
-import com.google.common.collect.ImmutableList;
-import com.google.common.collect.ImmutableMap;
+import static com.google.common.collect.Iterables.filter;
 
 public class XmlElement implements XmlNode {
     public XmlElement(String name) {
@@ -33,6 +37,11 @@ public class XmlElement implements XmlNode {
         return attributes;
     }
 
+    public String getAttribute(String name) {
+        // TODO: throw more informative error message
+        return Optional.ofNullable(attributes.get(name)).get();
+    }
+
     public List<XmlNode> getChildren() {
         return children;
     }
@@ -51,4 +60,11 @@ public class XmlElement implements XmlNode {
         return "XmlElement(name=" + name + ", attributes=" + attributes
                + ", children=" + children + ")";
     }
+
+    public XmlElementList findChildren(String name) {
+        Iterable<XmlElement> filtered = filter(
+            filter(children, XmlElement.class),
+            child -> child.getName().equals(name));
+        return new XmlElementList(ImmutableList.copyOf(filtered));
+    }
 }
diff --git a/src/main/java/org/zwobble/mammoth/xml/XmlElementList.java b/src/main/java/org/zwobble/mammoth/xml/XmlElementList.java
new file mode 100644
index 0000000..9e46314
--- /dev/null
+++ b/src/main/java/org/zwobble/mammoth/xml/XmlElementList.java
@@ -0,0 +1,17 @@
+package org.zwobble.mammoth.xml;
+
+import java.util.Iterator;
+import java.util.List;
+
+public class XmlElementList implements Iterable<XmlElement> {
+    private final List<XmlElement> elements;
+
+    public XmlElementList(List<XmlElement> elements) {
+        this.elements = elements;
+    }
+
+    @Override
+    public Iterator<XmlElement> iterator() {
+        return elements.iterator();
+    }
+}
diff --git a/src/main/java/org/zwobble/mammoth/xml/XmlNodes.java b/src/main/java/org/zwobble/mammoth/xml/XmlNodes.java
new file mode 100644
index 0000000..272a710
--- /dev/null
+++ b/src/main/java/org/zwobble/mammoth/xml/XmlNodes.java
@@ -0,0 +1,21 @@
+package org.zwobble.mammoth.xml;
+
+import com.google.common.collect.ImmutableList;
+import com.google.common.collect.ImmutableMap;
+
+import java.util.List;
+import java.util.Map;
+
+public class XmlNodes {
+    public static XmlElement element(String name, Map<String, String> attributes) {
+        return element(name, attributes, ImmutableList.of());
+    }
+
+    public static XmlElement element(String name, List<XmlNode> children) {
+        return element(name, ImmutableMap.of(), children);
+    }
+
+    public static XmlElement element(String name, Map<String, String> attributes, List<XmlNode> children) {
+        return new XmlElement(name, attributes, children);
+    }
+}
diff --git a/src/test/java/org/zwobble/mammoth/tests/docx/ContentTypesXmlTests.java b/src/test/java/org/zwobble/mammoth/tests/docx/ContentTypesXmlTests.java
new file mode 100644
index 0000000..6b46456
--- /dev/null
+++ b/src/test/java/org/zwobble/mammoth/tests/docx/ContentTypesXmlTests.java
@@ -0,0 +1,27 @@
+package org.zwobble.mammoth.tests.docx;
+
+import com.google.common.collect.ImmutableList;
+import com.google.common.collect.ImmutableMap;
+import org.junit.Test;
+import org.zwobble.mammoth.docx.ContentTypes;
+import org.zwobble.mammoth.xml.XmlElement;
+
+import java.util.Optional;
+
+import static org.junit.Assert.assertEquals;
+import static org.zwobble.mammoth.docx.ContentTypesXml.readContentTypesXmlElement;
+import static org.zwobble.mammoth.xml.XmlNodes.element;
+
+public class ContentTypesXmlTests {
+    @Test
+    public void contentTypesBasedOnDefaultForExtensionIfThereIsNoOverride() {
+        XmlElement element = element("content-types:Types", ImmutableList.of(
+            element("content-types:Default", ImmutableMap.of(
+                "Extension", "png",
+                "ContentType", "image/png"))));
+        ContentTypes contentTypes = readContentTypesXmlElement(element);
+        assertEquals(
+            Optional.of("image/png"),
+            contentTypes.findContentType("word/media/hat.png"));
+    }
+}