Merge pull request #45 from kc3hack/feat/#17-3

SplitBraceの修正、whiteSpace文の挿入阻止
kc3hack · Feb 24, 2024 · 6396cb3 · 6396cb3
2 parents 20eb8cf + 59325b4
commit 6396cb3
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 7 deletions.
diff --git a/Epub/KoeBook.Epub/ScrapingAozora.cs b/Epub/KoeBook.Epub/ScrapingAozora.cs
@@ -447,7 +447,7 @@ public async Task<EpubDocument> ScrapingAsync(string url, string coverFilePath,
                 {
                     if (nextNode.NodeType == NodeType.Text)
                     {
-                        if (nextNode.Text() != "\n")
+                        if (!string.IsNullOrWhiteSpace(nextNode.Text()))
                         {
                             previous = true;
 
@@ -522,7 +522,7 @@ private static string TextProcess(IElement element)
                     {
                         if (node.NodeType == NodeType.Text)
                         {
-                            if (node.Text() != "\n")
+                            if (!string.IsNullOrWhiteSpace(node.Text()))
                             {
                                 text += TextReplace(node.Text());
                             }
@@ -546,14 +546,14 @@ private static string TextProcess(IElement element)
                         }
                         else
                         {
-                            if ((item.TextContent != "\n") && (!string.IsNullOrEmpty(item.TextContent)))
+                            if (!string.IsNullOrWhiteSpace(item.TextContent) && (!string.IsNullOrEmpty(item.TextContent)))
                             {
                                 text += TextReplace(item.TextContent);
                             }
                         }
                         if (item.NextSibling != null)
                         {
-                            if ((item.NextSibling.TextContent != "\n") && (!string.IsNullOrEmpty(item.NextSibling.TextContent)))
+                            if (!string.IsNullOrWhiteSpace(item.NextSibling.TextContent) && (!string.IsNullOrEmpty(item.NextSibling.TextContent)))
                             {
                                 text += TextReplace(item.NextSibling.Text());
                             }

diff --git a/Epub/KoeBook.Epub/ScrapingHelper.cs b/Epub/KoeBook.Epub/ScrapingHelper.cs
@@ -3,7 +3,7 @@
 
 namespace KoeBook.Epub;
 
-internal static class ScrapingHelper
+public static class ScrapingHelper
 {
     internal static void checkChapter(EpubDocument document)
     {
@@ -46,6 +46,10 @@ internal static void checkParagraph(EpubDocument document, int chapterNum, int s
 
     public static List<string> SplitBrace(string text)
     {
+        if (text.Length == 1 && text != "「" && text != "」")
+        {
+            return new List<string>() { text };
+        }
         var result = new List<string>();
         int bracket = 0;
         var brackets = new List<int>();
@@ -62,19 +66,23 @@ public static List<string> SplitBrace(string text)
             brackets[i] -= mn;
             if (text[i] == '「' && brackets[i] == 1 && i != 0)
             {
-                result.Add(text[startIdx..(i - startIdx)]);
+                result.Add(text[startIdx..i]);
                 startIdx = i;
             }
             if (text[i] == '」' && brackets[i] == 0 && i != 0)
             {
-                result.Add(text[startIdx..(i - startIdx + 1)]);
+                result.Add(text[startIdx..(i + 1)]);
                 startIdx = i + 1;
             }
         }
         if (startIdx != text.Length - 1)
         {
             result.Add(text[startIdx..]);
         }
+        if (result[^1] == "")
+        {
+            result.RemoveAt(result.Count - 1);
+        }
 
         return result;
     }