Migrate search to Vertex AI and enhance preview modal

backend/handlers/api.go

@@ -67,6 +67,7 @@ func (h *Hub) Remove(conn *websocket.Conn) {
 type PreviewResponse struct {
 	Title           string            `json:"title"`
 	Thumbnail       string            `json:"thumbnail"`
+	PreviewStreamURL string           `json:"previewStreamUrl"`
 	Duration        string            `json:"duration"`
 	DurationSeconds int               `json:"durationSeconds"`
 	StartDefault    string            `json:"startDefault"`

backend/main.go

@@ -34,7 +34,13 @@ func main() {
 		DB:            db,
 		DownloadsDir:  downloadsDir,
 		WorkerScript:  workerScript,
-		SearchService: services.NewSearchService(os.Getenv("GOOGLE_CSE_API_KEY"), os.Getenv("GOOGLE_CSE_CX")),
+		SearchService: services.NewSearchService(
+			os.Getenv("VERTEX_AI_SEARCH_API_KEY"),
+			os.Getenv("VERTEX_AI_SEARCH_PROJECT_ID"),
+			os.Getenv("VERTEX_AI_SEARCH_LOCATION"),
+			os.Getenv("VERTEX_AI_SEARCH_DATA_STORE_ID"),
+			os.Getenv("VERTEX_AI_SEARCH_SERVING_CONFIG"),
+		),
 		GeminiService: services.NewGeminiService(os.Getenv("GEMINI_API_KEY")),
 		Hub:           handlers.NewHub(),
 	}

backend/services/cse.go

@@ -5,7 +5,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
-	"net/url"
+	neturl "net/url"
 	"strings"
 	"time"
 )
@@ -20,71 +20,82 @@ type SearchResult struct {
 }
 
 type SearchService struct {
-	APIKey string
+	APIKey        string
-	CX     string
+	ProjectID     string
-	Client *http.Client
+	Location      string
+	DataStoreID   string
+	ServingConfig string
+	Client        *http.Client
 }
 
-func NewSearchService(apiKey, cx string) *SearchService {
+func NewSearchService(apiKey, projectID, location, dataStoreID, servingConfig string) *SearchService {
+	if location == "" {
+		location = "global"
+	}
+	if servingConfig == "" {
+		servingConfig = "default_serving_config"
+	}
 	return &SearchService{
-		APIKey: apiKey,
+		APIKey:        apiKey,
-		CX:     cx,
+		ProjectID:     projectID,
-		Client: &http.Client{Timeout: 20 * time.Second},
+		Location:      location,
+		DataStoreID:   dataStoreID,
+		ServingConfig: servingConfig,
+		Client:        &http.Client{Timeout: 20 * time.Second},
 	}
 }
 
 func (s *SearchService) SearchMedia(query string) ([]SearchResult, error) {
-	if s.APIKey == "" || s.CX == "" {
+	if s.APIKey == "" || s.ProjectID == "" || s.DataStoreID == "" {
-		return nil, fmt.Errorf("google cse credentials are not configured")
+		return nil, fmt.Errorf("vertex ai search credentials are not configured")
 	}
-
+	results, err := s.searchLite(query, true)
-	domains := []string{"youtube.com", "tiktok.com", "envato.com", "artgrid.io"}
+	if err != nil {
-	siteQuery := strings.Join(domains, " OR site:")
+		results, err = s.searchLite(query, false)
-	fullQuery := fmt.Sprintf("%s (site:%s)", query, siteQuery)
-
-	values := url.Values{}
-	values.Set("key", s.APIKey)
-	values.Set("cx", s.CX)
-	values.Set("q", fullQuery)
-	values.Set("num", "10")
-	values.Set("safe", "off")
-
-	results := make([]SearchResult, 0, 30)
-	seen := map[string]bool{}
-	for _, start := range []string{"1", "11", "21"} {
-		pageResults, err := s.fetchPage(values, start, true)
 		if err != nil {
-			pageResults, err = s.fetchPage(values, start, false)
+			return nil, err
-			if err != nil {
-				return nil, err
-			}
-		}
-
-		for _, item := range pageResults {
-			if item.Link == "" || item.ThumbnailURL == "" || seen[item.Link] {
-				continue
-			}
-			seen[item.Link] = true
-			results = append(results, item)
 		}
 	}
 	return results, nil
 }
 
-func (s *SearchService) fetchPage(values url.Values, start string, imageSearch bool) ([]SearchResult, error) {
+func (s *SearchService) searchLite(query string, imageSearch bool) ([]SearchResult, error) {
-	pageValues := url.Values{}
+	filteredQuery := strings.TrimSpace(query + " site:youtube.com OR site:tiktok.com OR site:envato.com OR site:artgrid.io")
-	for key, items := range values {
+	servingConfig := fmt.Sprintf(
-		for _, item := range items {
+		"projects/%s/locations/%s/dataStores/%s/servingConfigs/%s",
-			pageValues.Add(key, item)
+		s.ProjectID,
-		}
+		s.Location,
+		s.DataStoreID,
+		s.ServingConfig,
+	)
+
+	params := map[string]any{
+		"user_country_code": "us",
 	}
-	pageValues.Set("start", start)
 	if imageSearch {
-		pageValues.Set("searchType", "image")
+		params["searchType"] = 1
 	}
 
-	endpoint := "https://www.googleapis.com/customsearch/v1?" + pageValues.Encode()
+	requestBody := map[string]any{
-	resp, err := s.Client.Get(endpoint)
+		"query":    filteredQuery,
+		"pageSize": 25,
+		"safeSearch": false,
+		"languageCode": "ko-KR",
+		"params": params,
+		"contentSearchSpec": map[string]any{
+			"snippetSpec": map[string]any{
+				"returnSnippet": true,
+			},
+		},
+	}
+
+	body, _ := json.Marshal(requestBody)
+	endpoint := fmt.Sprintf(
+		"https://discoveryengine.googleapis.com/v1/%s:searchLite?key=%s",
+		servingConfig,
+		neturl.QueryEscape(s.APIKey),
+	)
+	resp, err := s.Client.Post(endpoint, "application/json", strings.NewReader(string(body)))
 	if err != nil {
 		return nil, err
 	}
@@ -92,69 +103,73 @@ func (s *SearchService) fetchPage(values url.Values, start string, imageSearch b
 
 	if resp.StatusCode >= 300 {
 		data, _ := io.ReadAll(io.LimitReader(resp.Body, 2048))
-		return nil, fmt.Errorf("google cse returned status %d: %s", resp.StatusCode, strings.TrimSpace(string(data)))
+		return nil, fmt.Errorf("vertex ai search returned status %d: %s", resp.StatusCode, strings.TrimSpace(string(data)))
 	}
 
 	var payload struct {
-		Items []struct {
+		Results []struct {
-			Title       string `json:"title"`
+			Document struct {
-			Link        string `json:"link"`
+				StructData        map[string]any `json:"structData"`
-			DisplayLink string `json:"displayLink"`
+				DerivedStructData map[string]any `json:"derivedStructData"`
-			Snippet     string `json:"snippet"`
+			} `json:"document"`
-			Image       struct {
+		} `json:"results"`
-				ThumbnailLink string `json:"thumbnailLink"`
-			} `json:"image"`
-			Pagemap struct {
-				CSEImage []struct {
-					Src string `json:"src"`
-				} `json:"cse_image"`
-				CSEThumbnail []struct {
-					Src string `json:"src"`
-				} `json:"cse_thumbnail"`
-				Metatags []map[string]string `json:"metatags"`
-			} `json:"pagemap"`
-		} `json:"items"`
 	}
 
 	if err := json.NewDecoder(resp.Body).Decode(&payload); err != nil {
 		return nil, err
 	}
 
-	results := make([]SearchResult, 0, len(payload.Items))
+	results := make([]SearchResult, 0, len(payload.Results))
-	for _, item := range payload.Items {
+	for _, item := range payload.Results {
-		thumb := item.Image.ThumbnailLink
+		link := firstString(item.Document.StructData, "link", "url", "uri")
+		title := firstString(item.Document.StructData, "title", "name")
+		displayLink := firstString(item.Document.StructData, "site_name", "displayLink")
+		snippet := firstString(item.Document.DerivedStructData, "snippets", "snippet")
+		thumb := firstString(item.Document.DerivedStructData, "link", "thumbnail", "image", "image_url")
 		if thumb == "" {
-			thumb = extractThumbnail(item.Pagemap)
+			thumb = firstString(item.Document.StructData, "thumbnail", "image", "image_url")
+		}
+		if thumb == "" || link == "" {
+			continue
 		}
 		results = append(results, SearchResult{
-			Title:        item.Title,
+			Title:        title,
-			Link:         item.Link,
+			Link:         link,
-			DisplayLink:  item.DisplayLink,
+			DisplayLink:  displayLink,
-			Snippet:      item.Snippet,
+			Snippet:      snippet,
 			ThumbnailURL: thumb,
-			Source:       inferSource(item.DisplayLink),
+			Source:       inferSource(displayLink + " " + link),
 		})
 	}
 	return results, nil
 }
 
-func extractThumbnail(pagemap struct {
+func firstString(values map[string]any, keys ...string) string {
-	CSEImage     []struct{ Src string "json:\"src\"" }            "json:\"cse_image\""
+	for _, key := range keys {
-	CSEThumbnail []struct{ Src string "json:\"src\"" }            "json:\"cse_thumbnail\""
+		value, ok := values[key]
-	Metatags     []map[string]string                              "json:\"metatags\""
+		if !ok {
-}) string {
+			continue
-	if len(pagemap.CSEThumbnail) > 0 && pagemap.CSEThumbnail[0].Src != "" {
-		return pagemap.CSEThumbnail[0].Src
-	}
-	if len(pagemap.CSEImage) > 0 && pagemap.CSEImage[0].Src != "" {
-		return pagemap.CSEImage[0].Src
-	}
-	for _, tag := range pagemap.Metatags {
-		if value := tag["og:image"]; value != "" {
-			return value
 		}
-		if value := tag["twitter:image"]; value != "" {
+		switch typed := value.(type) {
-			return value
+		case string:
+			if typed != "" {
+				return typed
+			}
+		case []any:
+			for _, item := range typed {
+				if text, ok := item.(string); ok && text != "" {
+					return text
+				}
+				if mapped, ok := item.(map[string]any); ok {
+					if text := firstString(mapped, "snippet", "htmlSnippet", "url"); text != "" {
+						return text
+					}
+				}
+			}
+		case map[string]any:
+			if text := firstString(typed, "snippet", "htmlSnippet", "url"); text != "" {
+				return text
+			}
 		}
 	}
 	return ""

frontend/app.js

@@ -15,6 +15,7 @@ const downloadResult = document.getElementById("downloadResult");
 const cardTemplate = document.getElementById("searchCardTemplate");
 const previewModal = document.getElementById("previewModal");
 const previewTitle = document.getElementById("previewTitle");
+const previewVideo = document.getElementById("previewVideo");
 const previewThumbnail = document.getElementById("previewThumbnail");
 const previewDuration = document.getElementById("previewDuration");
 const qualitySelect = document.getElementById("qualitySelect");
@@ -117,6 +118,17 @@ function openPreviewModal(preview) {
   previewTitle.textContent = preview.title;
   previewThumbnail.src = preview.thumbnail;
   previewThumbnail.alt = preview.title;
+  previewVideo.pause();
+  previewVideo.removeAttribute("src");
+  previewVideo.load();
+  if (preview.previewStreamUrl) {
+    previewVideo.src = preview.previewStreamUrl;
+    previewVideo.classList.remove("hidden");
+    previewThumbnail.classList.add("hidden");
+  } else {
+    previewVideo.classList.add("hidden");
+    previewThumbnail.classList.remove("hidden");
+  }
   previewDuration.textContent = preview.duration;
   qualitySelect.innerHTML = "";
   for (const item of preview.qualities || []) {
@@ -132,6 +144,9 @@ function openPreviewModal(preview) {
 }
 
 function closeModal() {
+  previewVideo.pause();
+  previewVideo.removeAttribute("src");
+  previewVideo.load();
   previewModal.classList.add("hidden");
   previewModal.classList.remove("flex");
   pendingDownload = null;

frontend/index.html

@@ -60,10 +60,6 @@
             <h2 class="text-xl font-semibold text-white">Direct Downloader & Crop</h2>
             <form id="downloadForm" class="mt-4 space-y-3">
               <input id="downloadUrl" type="url" placeholder="https://..." class="w-full rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white placeholder:text-zinc-500" />
-              <div class="grid grid-cols-2 gap-3">
-                <input id="startTime" type="text" value="00:00:00" class="rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white" />
-                <input id="endTime" type="text" value="00:00:00" class="rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white" />
-              </div>
               <button class="w-full rounded-2xl border border-white px-5 py-3 text-sm font-medium text-white transition hover:bg-white hover:text-black">Preview & Queue</button>
             </form>
             <p id="downloadResult" class="mt-3 text-sm text-zinc-400"></p>
@@ -83,6 +79,7 @@
         </div>
         <div class="mt-5 grid gap-5 md:grid-cols-[1.1fr_0.9fr]">
           <div class="overflow-hidden rounded-3xl border border-white/10 bg-black/30">
+            <video id="previewVideo" class="hidden aspect-video h-full w-full bg-black object-cover" controls playsinline></video>
             <img id="previewThumbnail" class="aspect-video h-full w-full object-cover" alt="" />
           </div>
           <div class="space-y-4">
@@ -92,6 +89,16 @@
                 <span id="previewDuration"></span>
               </div>
             </div>
+            <div class="grid grid-cols-2 gap-3">
+              <label class="block space-y-2">
+                <span class="text-sm text-zinc-400">Start</span>
+                <input id="startTime" type="text" value="00:00:00" class="w-full rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white" />
+              </label>
+              <label class="block space-y-2">
+                <span class="text-sm text-zinc-400">End</span>
+                <input id="endTime" type="text" value="00:00:00" class="w-full rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white" />
+              </label>
+            </div>
             <label class="block space-y-2">
               <span class="text-sm text-zinc-400">Quality</span>
               <select id="qualitySelect" class="w-full rounded-2xl border border-white/10 bg-black/40 px-4 py-3 text-sm text-white"></select>

unraid-template.xml

@@ -16,7 +16,10 @@
   <Config Name="WebUI Port" Target="8080" Default="8080" Mode="tcp" Description="Dashboard port" Type="Port" Display="always" Required="true" Mask="false">8080</Config>
   <Config Name="Downloads" Target="/app/downloads" Default="/mnt/user/appdata/ai-media-hub/downloads" Mode="rw" Description="Media output directory" Type="Path" Display="always" Required="true" Mask="false">/mnt/user/appdata/ai-media-hub/downloads</Config>
   <Config Name="Database" Target="/app/db" Default="/mnt/user/appdata/ai-media-hub/db" Mode="rw" Description="SQLite database directory" Type="Path" Display="always" Required="true" Mask="false">/mnt/user/appdata/ai-media-hub/db</Config>
-  <Config Name="Google CSE API Key" Target="GOOGLE_CSE_API_KEY" Default="" Mode="" Description="Google Custom Search API key" Type="Variable" Display="always" Required="true" Mask="true"/>
+  <Config Name="Vertex Search API Key" Target="VERTEX_AI_SEARCH_API_KEY" Default="" Mode="" Description="Vertex AI Search API key for searchLite" Type="Variable" Display="always" Required="true" Mask="true"/>
-  <Config Name="Google CSE CX" Target="GOOGLE_CSE_CX" Default="" Mode="" Description="Google Custom Search Engine ID" Type="Variable" Display="always" Required="true" Mask="false"/>
+  <Config Name="Vertex Project ID" Target="VERTEX_AI_SEARCH_PROJECT_ID" Default="" Mode="" Description="Google Cloud project ID hosting Vertex AI Search" Type="Variable" Display="always" Required="true" Mask="false"/>
+  <Config Name="Vertex Location" Target="VERTEX_AI_SEARCH_LOCATION" Default="global" Mode="" Description="Vertex AI Search location" Type="Variable" Display="always" Required="true" Mask="false">global</Config>
+  <Config Name="Vertex Data Store ID" Target="VERTEX_AI_SEARCH_DATA_STORE_ID" Default="" Mode="" Description="Public website data store ID" Type="Variable" Display="always" Required="true" Mask="false"/>
+  <Config Name="Vertex Serving Config" Target="VERTEX_AI_SEARCH_SERVING_CONFIG" Default="default_serving_config" Mode="" Description="Serving config name for website searchLite" Type="Variable" Display="always" Required="true" Mask="false">default_serving_config</Config>
   <Config Name="Gemini API Key" Target="GEMINI_API_KEY" Default="" Mode="" Description="Gemini API key" Type="Variable" Display="always" Required="true" Mask="true"/>
 </Container>

worker/downloader.py

@@ -28,7 +28,7 @@ def run(cmd):
 
 def parse_duration(value):
     if value is None:
-        return "00:00:10"
+        return "00:00:00"
     total = int(float(value))
     hours = total // 3600
     minutes = (total % 3600) // 60
@@ -61,6 +61,24 @@ def build_quality_options(formats: List[dict]):
     return options
 
 
+def preview_stream_url(url):
+    candidates = [
+        "best[ext=mp4]/best",
+        "best",
+    ]
+    for selector in candidates:
+        proc = subprocess.run(
+            ["yt-dlp", "-g", "--no-playlist", "-f", selector, url],
+            capture_output=True,
+            text=True,
+        )
+        if proc.returncode == 0:
+            lines = [line.strip() for line in proc.stdout.splitlines() if line.strip()]
+            if lines:
+                return lines[0]
+    return ""
+
+
 def probe(url):
     cmd = ["yt-dlp", "--dump-single-json", "--no-playlist", url]
     proc = run(cmd)
@@ -71,6 +89,7 @@ def probe(url):
     preview = {
         "title": payload.get("title") or "Untitled",
         "thumbnail": thumbnail,
+        "previewStreamUrl": preview_stream_url(url),
         "durationSeconds": duration or 0,
         "duration": parse_duration(duration),
         "startDefault": "00:00:00",
@@ -85,8 +104,8 @@ def main():
     parser.add_argument("--mode", choices=["probe", "download"], default="download")
     parser.add_argument("--url", required=True)
     parser.add_argument("--start", default="00:00:00")
-    parser.add_argument("--end", default="00:00:10")
+    parser.add_argument("--end", default="00:00:00")
-    parser.add_argument("--output", required=True)
+    parser.add_argument("--output", default="")
     parser.add_argument("--quality", default="best")
     args = parser.parse_args()
 
@@ -94,6 +113,8 @@ def main():
         probe(args.url)
         return
 
+    if not args.output:
+        raise RuntimeError("output path is required for download mode")
     os.makedirs(os.path.dirname(args.output), exist_ok=True)
     emit("starting", 5, "Resolving media stream")