tweak outline template rendering

proxychain/responsemodifers/generate_readable_outline.go

@@ -1,14 +1,17 @@
 package responsemodifers
 
 import (
+	"bytes"
 	"embed"
 	"fmt"
+	"golang.org/x/net/html"
+	"golang.org/x/net/html/atom"
 	"html/template"
 	"io"
 	"ladder/proxychain"
 	"log"
 
-	"github.com/go-shiori/dom"
+	//"github.com/go-shiori/dom"
 	"github.com/markusmobius/go-trafilatura"
 )
 
@@ -32,32 +35,42 @@ func GenerateReadableOutline() proxychain.ResponseModification {
 		// 1. extract dom contents using reading mode algo
 		// ===========================================================
 		opts := trafilatura.Options{
-			IncludeImages: true,
+			IncludeImages:      true,
-			IncludeLinks:  true,
+			IncludeLinks:       true,
-			//FavorPrecision:     true,
+			FavorRecall:        true,
+			Deduplicate:        true,
 			FallbackCandidates: nil, // TODO: https://github.com/markusmobius/go-trafilatura/blob/main/examples/chained/main.go
 			// implement fallbacks from	"github.com/markusmobius/go-domdistiller" and 	"github.com/go-shiori/go-readability"
 			OriginalURL: chain.Request.URL,
 		}
 
-		result, err := trafilatura.Extract(chain.Response.Body, opts)
+		extract, err := trafilatura.Extract(chain.Response.Body, opts)
 		if err != nil {
 			return err
 		}
 
-		doc := trafilatura.CreateReadableDocument(result)
-		distilledHTML := dom.OuterHTML(doc)
-
 		// ============================================================================
 		// 2. render generate_readable_outline.html template using metadata from step 1
 		// ============================================================================
+
+		// render DOM to string without H1 title
+		removeFirstH1(extract.ContentNode)
+		var b bytes.Buffer
+		html.Render(&b, extract.ContentNode)
+		distilledHTML := b.String()
+
+		// populate template parameters
 		data := map[string]interface{}{
-			"Success": true,
+			"Success":     true,
-			"Params":  chain.Request.URL,
+			"Footer":      extract.Metadata.License,
-			//"Title":   result.Metadata.Title, // todo: modify CreateReadableDocument so we don't have <h1> titles duplicated?
+			"Image":       extract.Metadata.Image,
-			"Date":   result.Metadata.Date.String(),
+			"Description": extract.Metadata.Description,
-			"Author": result.Metadata.Author,
+			"Hostname":    extract.Metadata.Hostname,
-			"Body":   distilledHTML,
+			"Url":         chain.Request.URL,
+			"Title":       extract.Metadata.Title, // todo: modify CreateReadableDocument so we don't have <h1> titles duplicated?
+			"Date":        extract.Metadata.Date.String(),
+			"Author":      extract.Metadata.Author,
+			"Body":        distilledHTML,
 		}
 
 		// ============================================================================
@@ -82,3 +95,20 @@ func GenerateReadableOutline() proxychain.ResponseModification {
 		return nil
 	}
 }
+
+func removeFirstH1(n *html.Node) {
+	var recurse func(*html.Node) bool
+	recurse = func(n *html.Node) bool {
+		if n.Type == html.ElementNode && n.DataAtom == atom.H1 {
+			return true // Found the first H1, return true to stop
+		}
+		for c := n.FirstChild; c != nil; c = c.NextSibling {
+			if recurse(c) {
+				n.RemoveChild(c)
+				return false // Removed first H1, no need to continue
+			}
+		}
+		return false
+	}
+	recurse(n)
+}

proxychain/responsemodifers/generate_readable_outline.html

@@ -4,6 +4,7 @@
     <meta charset="UTF-8" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <link rel="stylesheet" href="/styles.css" />
+    <script src="/script.js" defer></script>
     <script>
       const handleThemeChange = () => {
         const theme = localStorage.getItem("theme");
@@ -50,7 +51,7 @@
               />
             </svg>
             <span class="text-3xl mr-1 text-[#7AA7D1] leading-8 align-middle"
-              >ladder</span
+                  >ladder | {{.Hostname}}</span
             >
           </a>
         </div>
@@ -327,19 +328,10 @@
         </code>
         {{else}}
         <div class="flex flex-col gap-1 mt-3">
-          <div>
-            <small class="text-sm font-medium leading-none">
-              <a
-                href="{{.Url}}"
-                class="text-slate-600 dark:text-slate-400 hover:text-blue-500 hover:underline underline-offset-2 transition-colors duration-300"
-                >{{.Url}}</a
-              >
-            </small>
-          </div>
           <h1
             class="text-3xl sm:text-4xl font-extrabold text-slate-900 tracking-tight dark:text-slate-200"
           >
-            {{.Title}}
+            <a href="{{.Url}}"> {{.Title}} </a>
           </h1>
           {{ if ne .Date "" }}
           <small
@@ -355,7 +347,13 @@
         </div>
 
         <div class="flex flex-col space-y-3">
-        <div>{{ .Params }}</div>
+          <div>
+            <div class="grid grid-cols-1 justify-items-center">
+            <div><img src="{{.Image}}"/></div>
+            <div class="text-xs text-gray-800">{{.Description}}</div>
+          </div>
+        </div>
+
         <div>{{ .Body }}</div>
         {{ end }}
         <!-- Trick Tailwind into compiling these styles into styles.css -->
@@ -382,6 +380,5 @@
       <div class="my-2"></div>
     </div>
 
-    <script src="/script.js"></script>
   </body>
 </html>