diff options
Diffstat (limited to 'src/main/java')
3 files changed, 101 insertions, 12 deletions
diff --git a/src/main/java/org/zwobble/mammoth/Mammoth.java b/src/main/java/org/zwobble/mammoth/Mammoth.java index 93c639a..4963c3a 100644 --- a/src/main/java/org/zwobble/mammoth/Mammoth.java +++ b/src/main/java/org/zwobble/mammoth/Mammoth.java @@ -7,6 +7,7 @@ import org.zwobble.mammoth.documents.HasChildren; import org.zwobble.mammoth.documents.Paragraph; import org.zwobble.mammoth.documents.Text; import org.zwobble.mammoth.docx.DocxFile; +import org.zwobble.mammoth.docx.InMemoryDocxFile; import org.zwobble.mammoth.docx.ZippedDocxFile; import org.zwobble.mammoth.html.Html; import org.zwobble.mammoth.results.Result; @@ -18,6 +19,8 @@ import org.zwobble.mammoth.util.Casts; import java.io.File; import java.io.IOException; +import java.io.InputStream; +import java.nio.file.Path; import java.util.List; import java.util.Optional; import java.util.zip.ZipFile; @@ -61,30 +64,47 @@ public class Mammoth { .mapParagraph(ParagraphMatcher.styleName("Normal"), HtmlPath.element("p")) .build(); - public static Result<String> convertToHtml(File file) { + public static Result<String> convertToHtml(InputStream stream) throws IOException { + return convertToHtml(stream, Options.DEFAULT); + } + + public static Result<String> convertToHtml(InputStream stream, Options options) throws IOException { + return withDocxFile(stream, zipFile -> + convertToHtml(Optional.empty(), zipFile, options)); + } + + public static Result<String> convertToHtml(File file) throws IOException { return convertToHtml(file, Options.DEFAULT); } - public static Result<String> convertToHtml(File file, Options options) { + public static Result<String> convertToHtml(File file, Options options) throws IOException { return withDocxFile(file, zipFile -> - readDocument(Optional.of(file.toPath()), zipFile) - .flatMap(nodes -> DocumentConverter.convertToHtml(options.idPrefix, options.preserveEmptyParagraphs, DEFAULT_STYLE_MAP, nodes)) - .map(Html::stripEmpty) - .map(Html::collapse) - .map(Html::write)); + convertToHtml(Optional.of(file.toPath()), zipFile, options)); } - public static Result<String> extractRawText(File file) { + private static Result<String> convertToHtml(Optional<Path> path, DocxFile zipFile, Options options) { + return readDocument(path, zipFile) + .flatMap(nodes -> DocumentConverter.convertToHtml(options.idPrefix, options.preserveEmptyParagraphs, DEFAULT_STYLE_MAP, nodes)) + .map(Html::stripEmpty) + .map(Html::collapse) + .map(Html::write); + } + + public static Result<String> extractRawText(File file) throws IOException { return withDocxFile(file, zipFile -> readDocument(Optional.of(file.toPath()), zipFile) .map(Mammoth::extractRawTextOfChildren)); } - private static <T> T withDocxFile(File file, Function<DocxFile, T> function) { + private static <T> T withDocxFile(File file, Function<DocxFile, T> function) throws IOException { try (DocxFile zipFile = new ZippedDocxFile(new ZipFile(file))) { return function.apply(zipFile); - } catch (IOException e) { - throw new UnsupportedOperationException("Should return a result of failure"); + } + } + + private static <T> T withDocxFile(InputStream stream, Function<DocxFile, T> function) throws IOException { + try (DocxFile zipFile = InMemoryDocxFile.fromStream(stream)) { + return function.apply(zipFile); } } diff --git a/src/main/java/org/zwobble/mammoth/docx/DocumentReader.java b/src/main/java/org/zwobble/mammoth/docx/DocumentReader.java index 284dd89..1fc0fc8 100644 --- a/src/main/java/org/zwobble/mammoth/docx/DocumentReader.java +++ b/src/main/java/org/zwobble/mammoth/docx/DocumentReader.java @@ -7,6 +7,9 @@ import org.zwobble.mammoth.util.MammothLists; import org.zwobble.mammoth.xml.XmlElement; import java.io.IOException; +import java.io.InputStream; +import java.net.URI; +import java.net.URISyntaxException; import java.nio.file.Path; import java.util.Optional; @@ -24,7 +27,14 @@ public class DocumentReader { ContentTypes contentTypes = readContentTypes(zipFile); FileReader fileReader = uri -> { try { - return path.get().toUri().resolve(uri).toURL().openStream(); + Optional<URI> absoluteUri = asAbsoluteUri(uri); + if (absoluteUri.isPresent()) { + return open(absoluteUri.get()); + } else if (path.isPresent()) { + return open(path.get().toUri().resolve(uri)); + } else { + throw new IOException("path of document is unknown, but is required for relative URI"); + } } catch (IOException exception) { throw new IOException("could not open external image '" + uri + "': " + exception.getMessage()); } @@ -40,6 +50,19 @@ public class DocumentReader { }); } + private static InputStream open(URI uri) throws IOException { + return uri.toURL().openStream(); + } + + private static Optional<URI> asAbsoluteUri(String uriString) { + try { + URI uri = new URI(uriString); + return uri.isAbsolute() ? Optional.of(uri) : Optional.empty(); + } catch (URISyntaxException exception) { + return Optional.empty(); + } + } + private static Result<Notes> readNotes(DocxFile file, BodyReaders bodyReaders) { return Result.map( tryParseOfficeXml(file, "word/footnotes.xml") diff --git a/src/main/java/org/zwobble/mammoth/docx/InMemoryDocxFile.java b/src/main/java/org/zwobble/mammoth/docx/InMemoryDocxFile.java new file mode 100644 index 0000000..3c6d4b4 --- /dev/null +++ b/src/main/java/org/zwobble/mammoth/docx/InMemoryDocxFile.java @@ -0,0 +1,46 @@ +package org.zwobble.mammoth.docx; + +import com.google.common.collect.ImmutableMap; +import com.google.common.collect.Maps; +import com.google.common.io.ByteStreams; + +import java.io.ByteArrayInputStream; +import java.io.IOException; +import java.io.InputStream; +import java.nio.charset.StandardCharsets; +import java.util.Map; +import java.util.Optional; +import java.util.zip.ZipEntry; +import java.util.zip.ZipInputStream; + +public class InMemoryDocxFile implements DocxFile { + public static DocxFile fromStream(InputStream stream) throws IOException { + ZipInputStream zipStream = new ZipInputStream(stream); + ImmutableMap.Builder<String, byte[]> entries = ImmutableMap.builder(); + ZipEntry entry; + while ((entry = zipStream.getNextEntry()) != null) { + entries.put(entry.getName(), ByteStreams.toByteArray(zipStream)); + } + return new InMemoryDocxFile(entries.build()); + } + + public static DocxFile fromStrings(Map<String, String> entries) { + return new InMemoryDocxFile(Maps.transformValues(entries, value -> value.getBytes(StandardCharsets.UTF_8))); + } + + private final Map<String, byte[]> entries; + + public InMemoryDocxFile(Map<String, byte[]> entries) { + this.entries = entries; + } + + @Override + public Optional<InputStream> tryGetInputStream(String name) throws IOException { + return Optional.ofNullable(entries.get(name)) + .map(ByteArrayInputStream::new); + } + + @Override + public void close() throws IOException { + } +} |