Add OpenSearch and SearchPage functionality for DuckDuckGo

Introduced the `OpenSearch` method and `SearchPage` interface to streamline search operations and allow for loading more results dynamically. Updated dependencies and modified the DuckDuckGo CLI to utilize these enhancements.

sites/duckduckgo/page.go

@@ -0,0 +1,68 @@
+package duckduckgo
+
+import (
+	"fmt"
+	"gitea.stevedudenhoeffer.com/steve/go-extractor"
+	"io"
+	"log/slog"
+)
+
+type SearchPage interface {
+	io.Closer
+	GetResults() []Result
+	LoadMore() error
+}
+
+type searchPage struct {
+	doc extractor.Document
+}
+
+func (s searchPage) GetResults() []Result {
+	var res []Result
+	var err error
+
+	err = s.doc.ForEach(`article[id^="r1-"]`, func(n extractor.Node) error {
+		var r Result
+
+		links := n.Select(`a[href][target="_self"]`)
+
+		if len(links) == 0 {
+			return nil
+		}
+
+		r.URL, err = links[0].Attr(`href`)
+
+		if err != nil {
+			return fmt.Errorf("failed to get link: %w", err)
+		}
+
+		titles := n.Select("h2")
+
+		if len(titles) != 0 {
+			r.Title, _ = titles[0].Text()
+		}
+
+		descriptions := n.Select("span > span")
+
+		if len(descriptions) != 0 {
+			r.Description, _ = descriptions[0].Text()
+		}
+
+		res = append(res, r)
+
+		return nil
+	})
+
+	return res
+}
+
+func (s searchPage) LoadMore() error {
+	return s.doc.ForEach(`button#more-results`, func(n extractor.Node) error {
+		slog.Info("clicking load more", "node", n)
+		return n.Click()
+	})
+}
+
+func (s searchPage) Close() error {
+	return s.doc.Close()
+}