From 3287e71ce896193eaf2bba8e6da9bf1827587ab3 Mon Sep 17 00:00:00 2001 From: Michael Williamson Date: Sun, 1 Aug 2021 11:17:17 +0100 Subject: Extract RawText class --- .../mammoth/internal/conversion/RawText.java | 35 ++++++++++++++++++++++ 1 file changed, 35 insertions(+) create mode 100644 src/main/java/org/zwobble/mammoth/internal/conversion/RawText.java (limited to 'src/main/java/org/zwobble/mammoth/internal/conversion/RawText.java') diff --git a/src/main/java/org/zwobble/mammoth/internal/conversion/RawText.java b/src/main/java/org/zwobble/mammoth/internal/conversion/RawText.java new file mode 100644 index 0000000..0834874 --- /dev/null +++ b/src/main/java/org/zwobble/mammoth/internal/conversion/RawText.java @@ -0,0 +1,35 @@ +package org.zwobble.mammoth.internal.conversion; + +import org.zwobble.mammoth.internal.documents.*; + +import java.util.List; + +import static org.zwobble.mammoth.internal.util.Casts.tryCast; +import static org.zwobble.mammoth.internal.util.Iterables.lazyMap; +import static org.zwobble.mammoth.internal.util.Lists.list; + +public class RawText { + public static String extractRawText(Document document) { + return extractRawTextOfChildren(document); + } + + private static String extractRawTextOfChildren(HasChildren parent) { + return extractRawText(parent.getChildren()); + } + + private static String extractRawText(List nodes) { + return String.join("", lazyMap(nodes, node -> extractRawText(node))); + } + + public static String extractRawText(DocumentElement node) { + return tryCast(Text.class, node) + .map(Text::getValue) + .orElseGet(() -> { + List children = tryCast(HasChildren.class, node) + .map(HasChildren::getChildren) + .orElse(list()); + String suffix = tryCast(Paragraph.class, node).map(paragraph -> "\n\n").orElse(""); + return extractRawText(children) + suffix; + }); + } +} -- cgit v1.2.3