From b7a49efa88d5c711d1da32163e95734f2b02f758 Mon Sep 17 00:00:00 2001
From: Gani Georgiev <gani.georgiev@gmail.com>
Date: Fri, 27 Oct 2023 17:34:45 +0300
Subject: [PATCH] fixed excerpt modifier to properly add spaces after block
 tags

---
 CHANGELOG.md                        |  2 ++
 tools/rest/excerpt_modifier.go      | 43 +++++++++++++++++------------
 tools/rest/excerpt_modifier_test.go |  4 +--
 3 files changed, 30 insertions(+), 19 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 78474a7d..6a6262e1 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,8 @@
   An option to return also the empty found tokens was also added via `Tokenizer.KeepEmptyTokens(true)`.
   _This should fix the parsing of whitespace charactes around view query column names when no quotes are used ([#3616](https://github.com/pocketbase/pocketbase/discussions/3616#discussioncomment-7398564))._
 
+- Fixed the `:excerpt(max, withEllipsis?)` `field` query param modifier to properly add space to the generated text fragment after block tags.
+
 
 ## v0.19.0
 
diff --git a/tools/rest/excerpt_modifier.go b/tools/rest/excerpt_modifier.go
index 107a8fcf..07da1a06 100644
--- a/tools/rest/excerpt_modifier.go
+++ b/tools/rest/excerpt_modifier.go
@@ -78,8 +78,7 @@ func (m *excerptModifier) Modify(value any) (any, error) {
 		return "", err
 	}
 
-	var isNotEmpty bool
-	var needSpace bool
+	var hasPrevSpace bool
 
 	// for all node types and more details check
 	// https://pkg.go.dev/golang.org/x/net/html#Parse
@@ -87,37 +86,47 @@ func (m *excerptModifier) Modify(value any) (any, error) {
 	stripTags = func(n *html.Node) {
 		switch n.Type {
 		case html.TextNode:
-			if txt := strings.TrimSpace(whitespaceRegex.ReplaceAllString(n.Data, " ")); txt != "" {
-				if isNotEmpty && needSpace {
-					needSpace = false
-					builder.WriteString(" ")
-				}
+			// collapse multiple spaces into one
+			txt := whitespaceRegex.ReplaceAllString(n.Data, " ")
+
+			if hasPrevSpace {
+				txt = strings.TrimLeft(txt, " ")
+			}
+
+			if txt != "" {
+				hasPrevSpace = strings.HasSuffix(txt, " ")
 
 				builder.WriteString(txt)
-
-				if !isNotEmpty {
-					isNotEmpty = true
-				}
-			}
-		case html.ElementNode:
-			if !needSpace && !list.ExistInSlice(n.Data, inlineTags) {
-				needSpace = true
 			}
 		}
 
-		if builder.Len() > m.max {
+		// excerpt max has been reached => no need to further iterate
+		// (+2 for the extra whitespace suffix/prefix that will be trimmed later)
+		if builder.Len() > m.max+2 {
 			return
 		}
 
 		for c := n.FirstChild; c != nil; c = c.NextSibling {
 			if c.Type != html.ElementNode || !list.ExistInSlice(c.Data, excludeTags) {
+				isBlock := c.Type == html.ElementNode && !list.ExistInSlice(c.Data, inlineTags)
+
+				if isBlock && !hasPrevSpace {
+					builder.WriteString(" ")
+					hasPrevSpace = true
+				}
+
 				stripTags(c)
+
+				if isBlock && !hasPrevSpace {
+					builder.WriteString(" ")
+					hasPrevSpace = true
+				}
 			}
 		}
 	}
 	stripTags(doc)
 
-	result := builder.String()
+	result := strings.TrimSpace(builder.String())
 
 	if len(result) > m.max {
 		result = strings.TrimSpace(result[:m.max])
diff --git a/tools/rest/excerpt_modifier_test.go b/tools/rest/excerpt_modifier_test.go
index 67c69850..eedb25a8 100644
--- a/tools/rest/excerpt_modifier_test.go
+++ b/tools/rest/excerpt_modifier_test.go
@@ -86,9 +86,9 @@ func TestNewExcerptModifier(t *testing.T) {
 func TestExcerptModifierModify(t *testing.T) {
 	// plain text value: "Hello t est12 3 word"
 	html := ` <script>var a = 123;</script>   <p>Hello</p><div id="test_id">t   est<b>12
-	3</b></div> <h1>word  </h1> `
+	3</b><span>456</span></div><span>word <b>7</b> 89<span>!<b>?</b><b> a </b><b>b </b>c</span>#<h1>title</h1>`
 
-	plainText := "Hello t est12 3 word"
+	plainText := "Hello t est12 3456 word 7 89!? a b c# title"
 
 	scenarios := []struct {
 		name     string