dataset update

app.py

@@ -3,92 +3,6 @@ from datetime import datetime
 import gradio as gr
 from huggingface_hub import HfApi, hf_hub_download
 
-
-import hashlib, subprocess, tempfile, pathlib, urllib.parse
-
-# ▶ 동영상이 있는 허브 데이터셋(repo id)
-VIDEO_DATASET_REPO = os.getenv("VIDEO_DATASET_REPO", "SGTLIM/ucf101_eval_unified")
-
-# ▶ 캐시 폴더 (충분한 공간 경로 권장)
-LOCAL_CACHE_DIR = "/tmp/hf_video_cache"
-MUTED_CACHE_DIR = "/tmp/hf_video_cache_muted"
-
-pathlib.Path(LOCAL_CACHE_DIR).mkdir(parents=True, exist_ok=True)
-pathlib.Path(MUTED_CACHE_DIR).mkdir(parents=True, exist_ok=True)
-
-def _sha1_8(s: str) -> str:
-    return hashlib.sha1(s.encode("utf-8")).hexdigest()[:8]
-
-def _download_local(v: dict) -> str:
-    """
-    v: {"url": ".../resolve/main/<id>.mp4", "id": "<id>.mp4", ...}
-    - id가 있으면 VIDEO_DATASET_REPO에서 hf_hub_download 사용
-    - (비상) id가 없으면 url에서 repo/id를 파싱해서 사용 시도
-    """
-    vid = v.get("id")
-    if vid:
-        return hf_hub_download(
-            repo_id=VIDEO_DATASET_REPO,
-            filename=vid,
-            repo_type="dataset",
-            token=HF_TOKEN,
-            local_dir=LOCAL_CACHE_DIR,
-            local_dir_use_symlinks=False,
-        )
-    # fallback: url 파싱 (datasets/{owner/repo}/resolve/main/{path})
-    u = v.get("url", "")
-    try:
-        parts = urllib.parse.urlparse(u).path.strip("/").split("/")
-        i = parts.index("datasets")
-        repo_id = "/".join(parts[i+1:i+3])              # owner/repo
-        file_path = "/".join(parts[i+5:])               # after resolve/main/
-        return hf_hub_download(
-            repo_id=repo_id, filename=file_path, repo_type="dataset",
-            token=HF_TOKEN, local_dir=LOCAL_CACHE_DIR, local_dir_use_symlinks=False
-        )
-    except Exception:
-        # 마지막 폴백: 그냥 원격 URL 그대로(소리 날 수 있음)
-        return u
-
-def _ensure_muted(local_or_url: str, key_hint: str) -> str:
-    """
-    local_or_url이 로컬 파일이면 그걸 기반으로 무음 mp4 생성.
-    -c:v copy -an → 실패 시 재인코딩으로 폴백.
-    이미 만들어진 캐시가 있으면 재사용.
-    """
-    # 원격이면 그대로(가능하면 _download_local로 로컬을 넘기세요)
-    if not os.path.isfile(local_or_url):
-        return local_or_url
-
-    h8 = _sha1_8(key_hint + "|" + os.path.abspath(local_or_url))
-    out = os.path.join(MUTED_CACHE_DIR, f"muted_{h8}.mp4")
-    if os.path.exists(out):
-        return out
-
-    # 1차: 스트림 복사로 오디오만 제거(빠르고 손상 없음)
-    cmd = ["ffmpeg","-y","-i",local_or_url,"-c:v","copy","-an",out]
-    try:
-        subprocess.run(cmd, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
-        return out
-    except subprocess.CalledProcessError:
-        pass
-
-    # 2차: 재인코딩 폴백 (확실히 무음)
-    cmd = [
-        "ffmpeg","-y","-i",local_or_url,
-        "-vf","format=yuv420p","-movflags","+faststart",
-        "-c:v","libx264","-crf","18","-preset","veryfast",
-        "-an", out
-    ]
-    subprocess.run(cmd, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
-    return out
-
-def _video_src_muted(v: dict) -> str:
-    """평가 화면에 넣을 '확실한 무음' 소스 경로 반환"""
-    local = _download_local(v)
-    return _ensure_muted(local, v.get("id") or v.get("url",""))
-
-
 # -------------------- Config --------------------
 REPO_ID = os.getenv("RESULTS_REPO", "sgtlim/videoeval_results")  # 업로드한 리포와 일치
 HF_TOKEN = os.getenv("HF_TOKEN")
@@ -207,7 +121,7 @@ def push(participant_id, video_id, score, notes=""):
             token=HF_TOKEN,
             commit_message="append"
         )
-        return gr.update(visible=True, value=f"✅ Saved successfully.")
+        return gr.update(visible=True, value=f"✅ Saved for {video_id}.")
     except Exception as e:
         return gr.update(
             visible=True,
@@ -341,15 +255,7 @@ for cls, files in EXAMPLES.items():
     EX_CACHE[cls] = {"real": None, "bad": None}
     for kind, fname in files.items():
         try:
-            # EX_CACHE[cls][kind] = hf_hub_download(
-            #     repo_id=REPO_ID,
-            #     filename=fname,
-            #     repo_type="dataset",
-            #     token=HF_TOKEN,
-            #     local_dir="/tmp",
-            #     local_dir_use_symlinks=False,
-            # )
-            p = hf_hub_download(
+            EX_CACHE[cls][kind] = hf_hub_download(
                 repo_id=REPO_ID,
                 filename=fname,
                 repo_type="dataset",
@@ -357,101 +263,11 @@ for cls, files in EXAMPLES.items():
                 local_dir="/tmp",
                 local_dir_use_symlinks=False,
             )
-            # ⬇️ 받은 파일을 확실한 무음으로 변환/캐시
-            EX_CACHE[cls][kind] = _ensure_muted(p, f"example::{cls}::{kind}")
         except Exception as e:
             print(f"[WARN] example missing: {cls} {kind} -> {fname}: {e}")
 
-# GLOBAL_CSS = """
-# /* ===== 공통 변수 투명화 (v3/v4 둘다) ===== */
-# :root, .gradio-container {
-#   --body-background-fill: transparent !important;
-#   --background-fill-primary: transparent !important;
-#   --background-fill-secondary: transparent !important;
-#   --block-background-fill: transparent !important;
-#   --block-border-color: transparent !important;
-#   --panel-background-fill: transparent !important;
-#   --panel-border-color: transparent !important;
-#   --section-header-background-fill: transparent !important;
-#   --shadow-drop: 0 0 0 rgba(0,0,0,0) !important;
-#   --shadow-spread: 0 0 0 rgba(0,0,0,0) !important;
-# }
-
-# /* ===== v4(Tailwind 기반)에서 자주 쓰이는 배경/테두리/그림자 제거 ===== */
-# .gradio-container .bg-white,
-# .gradio-container .bg-gray-50,
-# .gradio-container .bg-gray-100,
-# .gradio-container .bg-slate-50,
-# .gradio-container .bg-neutral-50,
-# .gradio-container .bg-secondary,
-# .gradio-container .border,
-# .gradio-container .shadow,
-# .gradio-container .shadow-sm,
-# .gradio-container .shadow-md,
-# .gradio-container .ring-1,
-# .gradio-container .ring,
-# .gradio-container .gr-card,
-# .gradio-container .prose > *:where(hr) {
-#   background: transparent !important;
-#   box-shadow: none !important;
-#   border-color: transparent !important;
-# }
-
-# /* ===== v3 컴포넌트 계열 ===== */
-# .gradio-container .gr-panel,
-# .gradio-container .gr-group,
-# .gradio-container .gr-box,
-# .gradio-container .gr-row,
-# .gradio-container .gr-column,
-# .gradio-container .gr-accordion,
-# .gradio-container .gr-block,
-# .gradio-container .gr-form,
-# .gradio-container .gr-tabs,
-# .gradio-container .gr-tabitem,
-# .gradio-container .gr-section-header {
-#   background: transparent !important;
-#   box-shadow: none !important;
-#   border: none !important;
-# }
-
-# /* 구분선/헤더 바 제거 */
-# .gradio-container hr,
-# .gradio-container .gr-divider,
-# .gradio-container .gr-accordion .label {
-#   background: transparent !important;
-#   border: none !important;
-#   box-shadow: none !important;
-# }
-
-# /* 바깥쪽 페이지 배경도 강제로 투명/흰색으로 */
-# html, body, .gradio-container { background: transparent !important; }
-
-# /* 기존 CSS 아래에 추가 */
-# #eval [class*="bg-"],
-# #eval [class*="border"],
-# #eval [class*="shadow"],
-# #eval .gr-panel, #eval .gr-group, #eval .gr-box, #eval .gr-row, #eval .gr-column,
-# #eval .gr-block, #eval .gr-form, #eval .gr-section-header, #eval .gr-accordion {
-#   background: transparent !important;
-#   border-color: transparent !important;
-#   box-shadow: none !important;
-# }
-# #eval .gr-form, #eval .gr-panel { background: transparent !important; box-shadow:none !important; border:none !important; }
-
-# """
-
 GLOBAL_CSS = """
-/* ===== 스크롤 강제 허용 ===== */
-html, body, #root, .gradio-container {
-  height: auto !important;
-  min-height: 100% !important;
-  overflow: auto !important;
-}
-.gradio-container *[class*="overflow-hidden"] {
-  overflow: visible !important;
-}
-
-/* ===== 배경/테두리 제거(기존 내용) ===== */
+/* ===== 공통 변수 투명화 (v3/v4 둘다) ===== */
 :root, .gradio-container {
   --body-background-fill: transparent !important;
   --background-fill-primary: transparent !important;
@@ -465,6 +281,7 @@ html, body, #root, .gradio-container {
   --shadow-spread: 0 0 0 rgba(0,0,0,0) !important;
 }
 
+/* ===== v4(Tailwind 기반)에서 자주 쓰이는 배경/테두리/그림자 제거 ===== */
 .gradio-container .bg-white,
 .gradio-container .bg-gray-50,
 .gradio-container .bg-gray-100,
@@ -484,6 +301,7 @@ html, body, #root, .gradio-container {
   border-color: transparent !important;
 }
 
+/* ===== v3 컴포넌트 계열 ===== */
 .gradio-container .gr-panel,
 .gradio-container .gr-group,
 .gradio-container .gr-box,
@@ -500,6 +318,7 @@ html, body, #root, .gradio-container {
   border: none !important;
 }
 
+/* 구분선/헤더 바 제거 */
 .gradio-container hr,
 .gradio-container .gr-divider,
 .gradio-container .gr-accordion .label {
@@ -508,6 +327,10 @@ html, body, #root, .gradio-container {
   box-shadow: none !important;
 }
 
+/* 바깥쪽 페이지 배경도 강제로 투명/흰색으로 */
+html, body, .gradio-container { background: transparent !important; }
+
+/* 기존 CSS 아래에 추가 */
 #eval [class*="bg-"],
 #eval [class*="border"],
 #eval [class*="shadow"],
@@ -518,68 +341,70 @@ html, body, #root, .gradio-container {
   box-shadow: none !important;
 }
 #eval .gr-form, #eval .gr-panel { background: transparent !important; box-shadow:none !important; border:none !important; }
-.gradio-container .h-screen,
-.gradio-container .min-h-screen { height: auto !important; min-height: auto !important; }
-.gradio-container .overflow-hidden { overflow: visible !important; }
-html, body { position: static !important; }
-"""
 
+"""
 
 # -------------------- UI --------------------
-with gr.Blocks(css=GLOBAL_CSS) as demo:
+with gr.Blocks(fill_height=True, css=GLOBAL_CSS) as demo:
+    # Blocks 안, 어디서든 한 번만 추가(권장 위치: Blocks 시작 직후)
     gr.HTML("""
     <script>
     (function(){
-      function allowScroll(){
-        // 최상위들 강제 오버라이드
-        [document.documentElement, document.body].forEach(el=>{
-          if (!el) return;
-          el.style.overflow = 'auto';
-          el.style.height = 'auto';
-          el.style.minHeight = '100%';
-          el.style.position = 'static';
-        });
-        const root = document.getElementById('root') || document.querySelector('#root,.gradio-container');
-        if (root){
-          root.style.overflow = 'auto';
-          root.style.height = 'auto';
-          root.style.minHeight = '100%';
+    // 모든 일반 DOM + Shadow DOM 안의 <video>까지 찾아 음소거
+    function eachNodeWithShadow(root, fn){
+        const walker = document.createTreeWalker(root, NodeFilter.SHOW_ELEMENT);
+        fn(root);
+        while (walker.nextNode()){
+        const el = walker.currentNode;
+        fn(el);
+        if (el.shadowRoot){
+            eachNodeWithShadow(el.shadowRoot, fn);
+        }
+        }
+    }
+
+    function muteVideoEl(v){
+        try{
+        // 재생 전에 반드시 muted 속성이 있어야 브라우저가 소리 차단
+        v.muted = true;
+        v.volume = 0.0;
+        v.setAttribute('muted','');
+        v.setAttribute('playsinline','');
+        v.setAttribute('preload','metadata');
+
+        // 사용자가 볼륨/음소거를 바꿔도 다시 0으로
+        if (!v._muteHooked){
+            v.addEventListener('volumechange', () => {
+            if (!v.muted || v.volume > 0) { v.muted = true; v.volume = 0.0; }
+            });
+            v.addEventListener('play', () => {
+            v.muted = true; v.volume = 0.0;
+            });
+            v._muteHooked = true;
         }
-        // 고정 레이아웃/오버플로우 차단 클래스 무력화
-        document.querySelectorAll('.overflow-hidden, .h-screen, .min-h-screen')
-          .forEach(el=>{
-            el.classList.remove('overflow-hidden','h-screen','min-h-screen');
-            el.style.overflow = 'auto';
-            if (getComputedStyle(el).height === '100vh') el.style.height = 'auto';
-            if (getComputedStyle(el).minHeight === '100vh') el.style.minHeight = 'auto';
-          });
-      }
-
-      const apply = () => requestAnimationFrame(allowScroll);
-
-      // 초기 1회
-      apply();
-
-      // 레이아웃/클래스가 바뀔 때마다 재적용
-      const mo = new MutationObserver(apply);
-      mo.observe(document.documentElement, {
-        subtree: true,
-        childList: true,
-        attributes: true,
-        attributeFilter: ['class','style']
-      });
-
-      // 가끔 놓치는 경우를 위해 주기적 백업 적용
-      const interval = setInterval(apply, 1500);
-
-      // 페이지 이벤트에도 훅
-      window.addEventListener('load', apply, { once:false });
-      document.addEventListener('visibilitychange', apply);
-      window.addEventListener('resize', apply);
+        }catch(e){}
+    }
+
+    function muteAll(root){
+        eachNodeWithShadow(root || document, (el)=>{
+        if (el.tagName && el.tagName.toLowerCase() === 'video'){
+            muteVideoEl(el);
+        }else if (el.querySelectorAll){
+            el.querySelectorAll('video').forEach(muteVideoEl);
+        }
+        });
+    }
+
+    // 초기 렌더에서
+    muteAll(document);
+
+    // 이후 DOM 변화 감시(Shadow DOM 내부 변화도 포착)
+    const obs = new MutationObserver(() => muteAll(document));
+    obs.observe(document, {subtree:true, childList:true, attributes:false});
     })();
     </script>
     """)
-    # Blocks 안, 어디서든 한 번만 추가(권장 위치: Blocks 시작 직후)
+
 
     order_state = gr.State(value=[])   # v4에서는 value= 권장
     ptr_state   = gr.State(value=0)
@@ -796,7 +621,7 @@ with gr.Blocks(css=GLOBAL_CSS) as demo:
                 _progress_html(TOTAL_PER_PARTICIPANT, TOTAL_PER_PARTICIPANT),
                 5.0,                                         # score 리셋
                 ptr,
-                cur_video_id
+                video_id
             )
 
         # 다음 영상 로드
@@ -804,8 +629,7 @@ with gr.Blocks(css=GLOBAL_CSS) as demo:
         v = V[next_idx]
         return (
             status_msg,
-            # v["url"],
-            _video_src_muted(v),
+            v["url"],
             _extract_action(v),
             new_done,
             _progress_html(new_done, TOTAL_PER_PARTICIPANT),