still moving codebase to new API (WIP)

core/api/request/page.go

@@ -16,11 +16,12 @@ import (
 	"strings"
 	"time"
 
+	"golang.org/x/net/html"
+
 	"github.com/documize/community/core/api/endpoint/models"
 	"github.com/documize/community/core/api/entity"
 	"github.com/documize/community/core/log"
 	"github.com/documize/community/core/streamutil"
-	"github.com/documize/community/domain/link"
 	"github.com/jmoiron/sqlx"
 )
 
@@ -249,7 +250,7 @@ func (p *Persister) UpdatePage(page entity.Page, refID, userID string, skipRevis
 	}
 
 	// find any content links in the HTML
-	links := link.GetContentLinks(page.Body)
+	links := GetContentLinks(page.Body)
 
 	// get a copy of previously saved links
 	previousLinks, _ := p.GetPageLinks(page.DocumentID, page.RefID)
@@ -497,3 +498,58 @@ func (p *Persister) GetNextPageSequence(documentID string) (maxSeq float64, err
 
 	return
 }
+
+// GetContentLinks returns Documize generated <a> links.
+// such links have an identifying attribute e.g. <a data-documize='true'...
+func GetContentLinks(body string) (links []entity.Link) {
+	z := html.NewTokenizer(strings.NewReader(body))
+
+	for {
+		tt := z.Next()
+
+		switch {
+		case tt == html.ErrorToken:
+			// End of the document, we're done
+			return
+		case tt == html.StartTagToken:
+			t := z.Token()
+
+			// Check if the token is an <a> tag
+			isAnchor := t.Data == "a"
+			if !isAnchor {
+				continue
+			}
+
+			// Extract the content link
+			ok, link := getLink(t)
+			if ok {
+				links = append(links, link)
+			}
+		}
+	}
+}
+
+// Helper function to pull the href attribute from a Token
+func getLink(t html.Token) (ok bool, link entity.Link) {
+	ok = false
+
+	// Iterate over all of the Token's attributes until we find an "href"
+	for _, a := range t.Attr {
+		switch a.Key {
+		case "data-documize":
+			ok = true
+		case "data-link-id":
+			link.RefID = strings.TrimSpace(a.Val)
+		case "data-link-space-id":
+			link.FolderID = strings.TrimSpace(a.Val)
+		case "data-link-target-document-id":
+			link.TargetDocumentID = strings.TrimSpace(a.Val)
+		case "data-link-target-id":
+			link.TargetID = strings.TrimSpace(a.Val)
+		case "data-link-type":
+			link.LinkType = strings.TrimSpace(a.Val)
+		}
+	}
+
+	return
+}