Changeset 0.17.3 (#78)

server/notelinks/extract.go

@@ -0,0 +1,50 @@
+package notelinks
+
+import (
+	"regexp"
+	"strings"
+
+	"git.gobha.me/xcaliber/chat-switchboard/models"
+)
+
+// wikiRe matches [[Title]], [[Title|Display]], ![[Title]], and ![[Title|Display]].
+// Title first char rejects [ and ] to avoid matching inside [[[triple]]].
+var wikiRe = regexp.MustCompile(`(!?)\[\[([^\[\]|][^\]|]*?)(?:\|([^\]]+?))?\]\]`)
+
+// ExtractWikilinks parses [[Title]], [[Title|Display]], and ![[Title]] from markdown content.
+// Returns deduplicated links preserving first occurrence.
+// Ignores matches preceded by [ (e.g. [[[triple]]]) since Go regexp lacks lookbehind.
+func ExtractWikilinks(content string) []models.NoteLink {
+	indices := wikiRe.FindAllStringSubmatchIndex(content, -1)
+	seen := make(map[string]bool)
+	var links []models.NoteLink
+
+	for _, idx := range indices {
+		matchStart := idx[0] // start of full match (including optional !)
+
+		// Skip if preceded by [ — rejects [[[triple]]] patterns
+		if matchStart > 0 && content[matchStart-1] == '[' {
+			continue
+		}
+
+		// Extract submatches by index pairs: idx[2*n], idx[2*n+1]
+		title := strings.TrimSpace(content[idx[4]:idx[5]]) // group 2: title
+
+		key := strings.ToLower(title)
+		if key == "" || seen[key] {
+			continue
+		}
+		seen[key] = true
+
+		link := models.NoteLink{
+			TargetTitle:    title,
+			IsTransclusion: content[idx[2]:idx[3]] == "!", // group 1
+		}
+		// group 3: display text (optional — idx[6:7] may be -1)
+		if idx[6] >= 0 && idx[7] >= 0 {
+			link.DisplayText = strings.TrimSpace(content[idx[6]:idx[7]])
+		}
+		links = append(links, link)
+	}
+	return links
+}