diff options
Diffstat (limited to 'src/main')
| -rw-r--r-- | src/main/java/org/zwobble/mammoth/Mammoth.java | 20 | ||||
| -rw-r--r-- | src/main/java/org/zwobble/mammoth/docx/OfficeXml.java | 26 |
2 files changed, 28 insertions, 18 deletions
diff --git a/src/main/java/org/zwobble/mammoth/Mammoth.java b/src/main/java/org/zwobble/mammoth/Mammoth.java index de0c45e..e2408e9 100644 --- a/src/main/java/org/zwobble/mammoth/Mammoth.java +++ b/src/main/java/org/zwobble/mammoth/Mammoth.java @@ -6,6 +6,7 @@ import com.google.common.collect.ImmutableMap; import org.zwobble.mammoth.documents.*; import org.zwobble.mammoth.docx.BodyXmlReader; import org.zwobble.mammoth.docx.DocumentXmlReader; +import org.zwobble.mammoth.docx.OfficeXml; import org.zwobble.mammoth.docx.Styles; import org.zwobble.mammoth.xml.XmlElement; import org.zwobble.mammoth.xml.parsing.XmlParser; @@ -19,23 +20,10 @@ import java.util.zip.ZipFile; import static com.google.common.collect.Iterables.transform; public class Mammoth { - private static ImmutableBiMap<String, String> XML_NAMESPACES = ImmutableBiMap.<String, String>builder() - .put("w", "http://schemas.openxmlformats.org/wordprocessingml/2006/main") - .put("wp", "http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing") - .put("a", "http://schemas.openxmlformats.org/drawingml/2006/main") - .put("pic", "http://schemas.openxmlformats.org/drawingml/2006/picture") - .put("content-types", "http://schemas.openxmlformats.org/package/2006/content-types") - .put("r", "http://schemas.openxmlformats.org/officeDocument/2006/relationships") - .put("relationships", "http://schemas.openxmlformats.org/package/2006/relationships") - .put("v", "urn:schemas-microsoft-com:vml") - .put("mc", "http://schemas.openxmlformats.org/markup-compatibility/2006") - .put("office-word", "urn:schemas-microsoft-com:office:word") - .build(); - public static String convertToHtml(File file) { try (ZipFile zipFile = new ZipFile(file)) { ZipEntry entry = zipFile.getEntry("word/document.xml"); - XmlElement documentXml = parseXml(zipFile.getInputStream(entry)); + XmlElement documentXml = OfficeXml.parseXml(zipFile.getInputStream(entry)); Styles styles = new Styles(ImmutableMap.of(), ImmutableMap.of()); Document document = new DocumentXmlReader(new BodyXmlReader(styles)).readElement(documentXml); @@ -78,8 +66,4 @@ public class Mammoth { } }); } - - private static XmlElement parseXml(InputStream inputStream) { - return new XmlParser(XML_NAMESPACES).parseStream(inputStream); - } } diff --git a/src/main/java/org/zwobble/mammoth/docx/OfficeXml.java b/src/main/java/org/zwobble/mammoth/docx/OfficeXml.java new file mode 100644 index 0000000..98ec1bf --- /dev/null +++ b/src/main/java/org/zwobble/mammoth/docx/OfficeXml.java @@ -0,0 +1,26 @@ +package org.zwobble.mammoth.docx; + +import com.google.common.collect.ImmutableBiMap; +import org.zwobble.mammoth.xml.XmlElement; +import org.zwobble.mammoth.xml.parsing.XmlParser; + +import java.io.InputStream; + +public class OfficeXml { + private static ImmutableBiMap<String, String> XML_NAMESPACES = ImmutableBiMap.<String, String>builder() + .put("w", "http://schemas.openxmlformats.org/wordprocessingml/2006/main") + .put("wp", "http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing") + .put("a", "http://schemas.openxmlformats.org/drawingml/2006/main") + .put("pic", "http://schemas.openxmlformats.org/drawingml/2006/picture") + .put("content-types", "http://schemas.openxmlformats.org/package/2006/content-types") + .put("r", "http://schemas.openxmlformats.org/officeDocument/2006/relationships") + .put("relationships", "http://schemas.openxmlformats.org/package/2006/relationships") + .put("v", "urn:schemas-microsoft-com:vml") + .put("mc", "http://schemas.openxmlformats.org/markup-compatibility/2006") + .put("office-word", "urn:schemas-microsoft-com:office:word") + .build(); + + public static XmlElement parseXml(InputStream inputStream) { + return new XmlParser(XML_NAMESPACES).parseStream(inputStream); + } +} |
