From 6005c9dc75a8d17761d775b6a86d79892aef1a56 Mon Sep 17 00:00:00 2001
From: Michael Williamson <mike@zwobble.org>
Date: Sat, 26 Sep 2015 18:11:53 +0100
Subject: Add conversion of single paragraph documents

---
 src/main/java/org/zwobble/mammoth/Mammoth.java | 50 ++++++++++++++++++++++++--
 1 file changed, 48 insertions(+), 2 deletions(-)

(limited to 'src/main/java/org/zwobble/mammoth/Mammoth.java')

diff --git a/src/main/java/org/zwobble/mammoth/Mammoth.java b/src/main/java/org/zwobble/mammoth/Mammoth.java
index 5a2ed17..aa3514f 100644
--- a/src/main/java/org/zwobble/mammoth/Mammoth.java
+++ b/src/main/java/org/zwobble/mammoth/Mammoth.java
@@ -1,11 +1,57 @@
 package org.zwobble.mammoth;
 
+import java.io.File;
+import java.io.IOException;
 import java.io.InputStream;
+import java.util.zip.ZipFile;
+
+import org.zwobble.mammoth.xml.XmlElement;
+import org.zwobble.mammoth.xml.XmlNodeVisitor;
+import org.zwobble.mammoth.xml.XmlParser;
+import org.zwobble.mammoth.xml.XmlTextNode;
+
+import com.google.common.base.Joiner;
+import com.google.common.base.Strings;
+import com.google.common.collect.ImmutableBiMap;
+
+import static com.google.common.collect.Iterables.transform;
+
+import lombok.val;
 
 public class Mammoth {
 
-    public static String convertToHtml(InputStream stream) {
-        return "";
+    public static String convertToHtml(File file) {
+        try (val zipFile = new ZipFile(file)) {
+            val entry = zipFile.getEntry("word/document.xml");
+            val documentXml = parseXml(zipFile.getInputStream(entry));
+            val text = extractText(documentXml);
+            if (Strings.isNullOrEmpty(text)) {
+                return "";
+            } else {
+                return "<p>" + text + "</p>";                
+            }
+        } catch (IOException e) {
+            throw new UnsupportedOperationException("Should return a result of failure");   
+        }
+    }
+
+    private static XmlElement parseXml(InputStream inputStream) {
+        return new XmlParser(ImmutableBiMap.of()).parseStream(inputStream);
+    }
+
+    private static String extractText(XmlElement documentXml) {
+        return documentXml.accept(new XmlNodeVisitor<String>() {
+            @Override
+            public String visit(XmlElement element) {
+                return Joiner.on("").join(
+                    transform(element.children(), child -> child.accept(this)));
+            }
+
+            @Override
+            public String Visit(XmlTextNode textNode) {
+                return textNode.getValue();
+            }
+        });
     }
 
 }
-- 
cgit v1.2.3