Add FastAPI transcriber with file upload, URL support, SSE progress

Dockerfile

@@ -0,0 +1,16 @@
+FROM python:3.11-slim
+
+RUN apt-get update && \
+    apt-get install -y ffmpeg && \
+    rm -rf /var/lib/apt/lists/*
+
+WORKDIR /app
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY app.py .
+
+EXPOSE 7860
+
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,12 +1,29 @@
 ---
-title: Vtranscript
-emoji: 👁
+title: VOD Transcriber
+emoji: 🎙
 colorFrom: gray
-colorTo: indigo
+colorTo: gray
 sdk: docker
+app_port: 7860
 pinned: false
-license: cc-by-nc-nd-4.0
-short_description: VT
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# VOD Transcriber
+This directory contains `transcribe_vod.sh` and the transcript outputs.
+
+## What was fixed
+- Added a dependency check for `ffmpeg` before extraction.
+  - If missing, the script now exits with:
+    - `Error: ffmpeg not found in PATH.`
+    - `Install ffmpeg with: brew install ffmpeg`
+- Removed suppression of `ffmpeg` extraction errors so failures are visible.
+- Updated Homebrew whisper executable path from `.../bin/whisper-cpp` to `.../bin/whisper-cli`.
+
+## Why extraction previously failed
+The script attempted to run `ffmpeg` without verifying it existed, and stderr was redirected away on extraction. When `ffmpeg` was missing, this produced exit code `127` with limited diagnostics.
+
+## Verification
+After installing `ffmpeg` and updating the script:
+- audio extraction from `vod.mp4` completed successfully
+- whisper transcription completed successfully
+- transcript output was written to `vod_transcript.txt`

app.py

@@ -0,0 +1,509 @@
+"""
+VOD Transcriber — Hugging Face Spaces
+FastAPI + faster-whisper + yt-dlp
+"""
+import asyncio
+import glob
+import json
+import os
+import secrets
+import shutil
+import threading
+import uuid
+from concurrent.futures import ThreadPoolExecutor
+from typing import Optional
+
+import yt_dlp
+from faster_whisper import WhisperModel
+from fastapi import Depends, FastAPI, File, Form, HTTPException, UploadFile
+from fastapi.responses import StreamingResponse
+from fastapi.security import HTTPBasic, HTTPBasicCredentials
+
+# ---------------------------------------------------------------------------
+# Config
+# ---------------------------------------------------------------------------
+PASSWORD = os.getenv("APP_PASSWORD", "changeme")
+MAX_UPLOAD_BYTES = 500 * 1024 * 1024  # 500 MB
+WHISPER_MODEL = "base.en"
+
+# ---------------------------------------------------------------------------
+# App + auth
+# ---------------------------------------------------------------------------
+app = FastAPI()
+security = HTTPBasic()
+executor = ThreadPoolExecutor(max_workers=2)
+
+_model: Optional[WhisperModel] = None
+_model_lock = threading.Lock()
+
+
+def get_model() -> WhisperModel:
+    global _model
+    if _model is None:
+        with _model_lock:
+            if _model is None:
+                _model = WhisperModel(WHISPER_MODEL, device="cpu", compute_type="int8")
+    return _model
+
+
+def check_auth(credentials: HTTPBasicCredentials = Depends(security)) -> str:
+    ok = secrets.compare_digest(credentials.password.encode(), PASSWORD.encode())
+    if not ok:
+        raise HTTPException(
+            status_code=401,
+            detail="Unauthorized",
+            headers={"WWW-Authenticate": "Basic"},
+        )
+    return credentials.username
+
+
+# ---------------------------------------------------------------------------
+# Job store
+# ---------------------------------------------------------------------------
+jobs: dict[str, dict] = {}
+
+
+def new_job() -> str:
+    jid = str(uuid.uuid4())
+    jobs[jid] = {"messages": [], "done": False, "transcript": None, "error": None}
+    return jid
+
+
+def push(jid: str, msg: str):
+    jobs[jid]["messages"].append(msg)
+
+
+# ---------------------------------------------------------------------------
+# Transcription worker (runs in thread)
+# ---------------------------------------------------------------------------
+def transcribe_file(jid: str, path: str, cleanup_paths: list[str]):
+    try:
+        push(jid, "Loading whisper model…")
+        model = get_model()
+
+        push(jid, f"Transcribing {os.path.basename(path)}…")
+        segments, info = model.transcribe(path, language="en", beam_size=5)
+
+        lines = []
+        for seg in segments:
+            ts = f"[{seg.start:.1f}s – {seg.end:.1f}s]"
+            lines.append(f"{ts} {seg.text.strip()}")
+            push(jid, f"{ts} {seg.text.strip()}")
+
+        jobs[jid]["transcript"] = "\n".join(lines)
+        push(jid, "✓ Done")
+    except Exception as e:
+        jobs[jid]["error"] = str(e)
+        push(jid, f"ERROR: {e}")
+    finally:
+        jobs[jid]["done"] = True
+        for p in cleanup_paths:
+            try:
+                os.remove(p)
+            except OSError:
+                pass
+
+
+# ---------------------------------------------------------------------------
+# Routes
+# ---------------------------------------------------------------------------
+@app.get("/", dependencies=[Depends(check_auth)])
+async def index():
+    from fastapi.responses import HTMLResponse
+    return HTMLResponse(HTML)
+
+
+@app.post("/transcribe/file")
+async def transcribe_upload(
+    file: UploadFile = File(...),
+    _: str = Depends(check_auth),
+):
+    # Size check via content-length header
+    size = int(file.headers.get("content-length", 0))
+    if size > MAX_UPLOAD_BYTES:
+        raise HTTPException(413, f"File too large. Max {MAX_UPLOAD_BYTES // 1024 // 1024} MB.")
+
+    jid = new_job()
+    ext = os.path.splitext(file.filename or "video.mp4")[1] or ".mp4"
+    tmp_path = f"/tmp/{jid}{ext}"
+
+    push(jid, f"Saving upload ({file.filename})…")
+    with open(tmp_path, "wb") as f:
+        shutil.copyfileobj(file.file, f)
+
+    # Double-check actual size
+    actual = os.path.getsize(tmp_path)
+    if actual > MAX_UPLOAD_BYTES:
+        os.remove(tmp_path)
+        raise HTTPException(413, f"File too large. Max {MAX_UPLOAD_BYTES // 1024 // 1024} MB.")
+
+    executor.submit(transcribe_file, jid, tmp_path, [tmp_path])
+    return {"job_id": jid}
+
+
+@app.post("/transcribe/url")
+async def transcribe_url(
+    url: str = Form(...),
+    _: str = Depends(check_auth),
+):
+    jid = new_job()
+    push(jid, f"Fetching URL: {url}")
+
+    def download_and_transcribe():
+        output_tpl = f"/tmp/{jid}.%(ext)s"
+        downloaded = []
+
+        class Hook:
+            def __call__(self, d):
+                if d["status"] == "downloading":
+                    pct = d.get("_percent_str", "").strip()
+                    spd = d.get("_speed_str", "").strip()
+                    if pct:
+                        push(jid, f"Downloading {pct} at {spd}")
+                elif d["status"] == "finished":
+                    downloaded.append(d["filename"])
+                    push(jid, "Download complete.")
+
+        opts = {
+            "format": "bestaudio/best",
+            "outtmpl": output_tpl,
+            "progress_hooks": [Hook()],
+            "quiet": True,
+            "no_warnings": True,
+        }
+
+        try:
+            with yt_dlp.YoutubeDL(opts) as ydl:
+                ydl.download([url])
+
+            # Find downloaded file (extension may differ from template)
+            files = glob.glob(f"/tmp/{jid}.*")
+            if not files:
+                raise RuntimeError("Download produced no output file.")
+            video_path = files[0]
+
+            # Check size
+            if os.path.getsize(video_path) > MAX_UPLOAD_BYTES:
+                os.remove(video_path)
+                raise RuntimeError(f"Downloaded file exceeds 500 MB limit.")
+
+            transcribe_file(jid, video_path, files)
+
+        except Exception as e:
+            jobs[jid]["error"] = str(e)
+            push(jid, f"ERROR: {e}")
+            jobs[jid]["done"] = True
+
+    executor.submit(download_and_transcribe)
+    return {"job_id": jid}
+
+
+@app.get("/progress/{jid}")
+async def progress(jid: str, _: str = Depends(check_auth)):
+    if jid not in jobs:
+        raise HTTPException(404, "Job not found.")
+
+    async def stream():
+        sent = 0
+        while True:
+            job = jobs[jid]
+            msgs = job["messages"]
+            while sent < len(msgs):
+                yield f"data: {json.dumps({'msg': msgs[sent]})}\n\n"
+                sent += 1
+            if job["done"]:
+                yield f"data: {json.dumps({'done': True, 'transcript': job['transcript'], 'error': job['error']})}\n\n"
+                del jobs[jid]
+                break
+            await asyncio.sleep(0.4)
+
+    return StreamingResponse(stream(), media_type="text/event-stream")
+
+
+# ---------------------------------------------------------------------------
+# HTML
+# ---------------------------------------------------------------------------
+HTML = r"""<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1">
+<title>VOD Transcriber</title>
+<style>
+  *, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }
+  body {
+    font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;
+    background: #0f0f11;
+    color: #e2e2e5;
+    min-height: 100vh;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    padding: 2rem;
+  }
+  .card {
+    background: #1a1a1f;
+    border: 1px solid #2a2a32;
+    border-radius: 16px;
+    padding: 2.5rem;
+    width: 100%;
+    max-width: 660px;
+    box-shadow: 0 8px 40px rgba(0,0,0,0.4);
+  }
+  .card-header { margin-bottom: 2rem; }
+  h1 { font-size: 1.4rem; font-weight: 600; color: #fff; margin-bottom: 0.3rem; letter-spacing: -0.02em; }
+  .subtitle { font-size: 0.85rem; color: #555; }
+
+  /* Tabs */
+  .tabs { display: flex; gap: 0.25rem; background: #111115; border-radius: 10px; padding: 0.25rem; margin-bottom: 1.25rem; }
+  .tab {
+    flex: 1; text-align: center; padding: 0.55rem; border-radius: 8px;
+    font-size: 0.85rem; font-weight: 500; cursor: pointer; color: #666;
+    transition: background 0.15s, color 0.15s; user-select: none;
+  }
+  .tab.active { background: #2a2a32; color: #e2e2e5; }
+
+  /* Input areas */
+  .input-panel { display: none; }
+  .input-panel.active { display: block; }
+
+  label { display: block; font-size: 0.8rem; font-weight: 500; color: #888; margin-bottom: 0.5rem; text-transform: uppercase; letter-spacing: 0.05em; }
+
+  .upload-area {
+    background: #111115;
+    border: 1.5px dashed #2a2a32;
+    border-radius: 10px;
+    padding: 2rem;
+    text-align: center;
+    cursor: pointer;
+    transition: border-color 0.15s;
+    margin-bottom: 1rem;
+  }
+  .upload-area:hover, .upload-area.drag { border-color: #5b5bf6; }
+  .upload-area input { display: none; }
+  .upload-icon { font-size: 2rem; margin-bottom: 0.5rem; }
+  .upload-hint { font-size: 0.82rem; color: #555; }
+  .upload-hint span { color: #5b5bf6; }
+  .file-selected { font-size: 0.82rem; color: #e2e2e5; margin-top: 0.5rem; }
+
+  .url-row { display: flex; gap: 0.75rem; margin-bottom: 1rem; }
+  input[type="text"] {
+    flex: 1; background: #111115; border: 1px solid #2a2a32; border-radius: 10px;
+    padding: 0.7rem 1rem; font-size: 0.9rem; color: #e2e2e5; outline: none;
+    transition: border-color 0.15s;
+  }
+  input[type="text"]:focus { border-color: #5b5bf6; }
+  input[type="text"]::placeholder { color: #444; }
+
+  .url-hint { font-size: 0.78rem; color: #444; margin-bottom: 1rem; }
+
+  button.primary {
+    width: 100%; background: #5b5bf6; color: #fff; border: none; border-radius: 10px;
+    padding: 0.75rem; font-size: 0.95rem; font-weight: 600; cursor: pointer;
+    transition: background 0.15s, opacity 0.15s;
+  }
+  button.primary:hover:not(:disabled) { background: #4a4ae0; }
+  button.primary:disabled { opacity: 0.4; cursor: not-allowed; }
+
+  /* Status */
+  .status { display: none; align-items: center; gap: 0.6rem; font-size: 0.85rem; color: #888; margin-top: 1.25rem; }
+  .status.visible { display: flex; }
+  .spinner { width: 16px; height: 16px; border: 2px solid #2a2a32; border-top-color: #5b5bf6; border-radius: 50%; animation: spin 0.7s linear infinite; flex-shrink: 0; }
+  @keyframes spin { to { transform: rotate(360deg); } }
+
+  /* Log */
+  .log-box {
+    display: none; background: #111115; border: 1px solid #2a2a32; border-radius: 10px;
+    padding: 1rem; font-family: "SF Mono", "Fira Code", monospace; font-size: 0.75rem;
+    line-height: 1.6; color: #666; max-height: 220px; overflow-y: auto;
+    margin-top: 1.25rem; white-space: pre-wrap; word-break: break-all;
+  }
+  .log-box.visible { display: block; }
+
+  /* Result */
+  .result { display: none; background: #0d1f17; border: 1px solid #1a3a28; border-radius: 10px; padding: 1rem 1.25rem; margin-top: 1.25rem; }
+  .result.visible { display: block; }
+  .result-label { font-size: 0.75rem; font-weight: 600; color: #3d9e6a; text-transform: uppercase; letter-spacing: 0.05em; margin-bottom: 0.75rem; }
+  .result-actions { display: flex; gap: 0.75rem; }
+  .btn-dl {
+    background: #1a3a28; color: #5bf6a0; border: 1px solid #1a3a28; border-radius: 8px;
+    padding: 0.5rem 1rem; font-size: 0.82rem; font-weight: 600; cursor: pointer;
+    text-decoration: none; display: inline-block; transition: background 0.15s;
+  }
+  .btn-dl:hover { background: #224d36; }
+
+  /* Error */
+  .error-box { display: none; background: #1f0d0d; border: 1px solid #3a1a1a; border-radius: 10px; padding: 1rem 1.25rem; font-size: 0.85rem; color: #f66; margin-top: 1.25rem; }
+  .error-box.visible { display: block; }
+</style>
+</head>
+<body>
+<div class="card">
+  <div class="card-header">
+    <h1>VOD Transcriber</h1>
+    <p class="subtitle">whisper · local processing · no data retained</p>
+  </div>
+
+  <div class="tabs">
+    <div class="tab active" onclick="switchTab('file')">File Upload</div>
+    <div class="tab" onclick="switchTab('url')">Video URL</div>
+  </div>
+
+  <!-- File panel -->
+  <div class="input-panel active" id="panel-file">
+    <div class="upload-area" id="drop-zone" onclick="document.getElementById('file-input').click()"
+         ondragover="event.preventDefault(); this.classList.add('drag')"
+         ondragleave="this.classList.remove('drag')"
+         ondrop="onDrop(event)">
+      <input type="file" id="file-input" accept="video/*,audio/*" onchange="onFileSelect(this)">
+      <div class="upload-icon">🎬</div>
+      <div class="upload-hint">Drop video or audio file here, or <span>browse</span></div>
+      <div class="upload-hint">MP4, MOV, MKV, WebM, MP3, WAV · max 500 MB</div>
+      <div class="file-selected" id="file-name"></div>
+    </div>
+  </div>
+
+  <!-- URL panel -->
+  <div class="input-panel" id="panel-url">
+    <div class="url-row">
+      <input type="text" id="url-input" placeholder="https://youtube.com/watch?v=..." autocomplete="off" spellcheck="false">
+    </div>
+    <p class="url-hint">YouTube, Twitter/X, Vimeo, Twitch, and 1000+ sites via yt-dlp</p>
+  </div>
+
+  <button class="primary" id="btn" onclick="run()">Transcribe</button>
+
+  <div class="status" id="status"><div class="spinner"></div><span id="status-text">Processing…</span></div>
+  <pre class="log-box" id="log"></pre>
+
+  <div class="result" id="result">
+    <div class="result-label">Transcript ready</div>
+    <div class="result-actions">
+      <a class="btn-dl" id="dl-link" download="transcript.txt">Download .txt</a>
+    </div>
+  </div>
+
+  <div class="error-box" id="error"></div>
+</div>
+
+<script>
+let activeTab = 'file';
+let selectedFile = null;
+let transcriptText = null;
+
+function switchTab(tab) {
+  activeTab = tab;
+  document.querySelectorAll('.tab').forEach((t, i) => t.classList.toggle('active', (i === 0) === (tab === 'file')));
+  document.getElementById('panel-file').classList.toggle('active', tab === 'file');
+  document.getElementById('panel-url').classList.toggle('active', tab === 'url');
+}
+
+function onFileSelect(input) {
+  selectedFile = input.files[0] || null;
+  document.getElementById('file-name').textContent = selectedFile ? selectedFile.name : '';
+}
+
+function onDrop(e) {
+  e.preventDefault();
+  document.getElementById('drop-zone').classList.remove('drag');
+  const f = e.dataTransfer.files[0];
+  if (f) {
+    selectedFile = f;
+    document.getElementById('file-name').textContent = f.name;
+  }
+}
+
+function reset() {
+  ['log','result','error'].forEach(id => document.getElementById(id).classList.remove('visible'));
+  document.getElementById('log').textContent = '';
+  document.getElementById('status').classList.remove('visible');
+  transcriptText = null;
+}
+
+function appendLog(msg) {
+  const el = document.getElementById('log');
+  el.classList.add('visible');
+  el.textContent += msg + '\n';
+  el.scrollTop = el.scrollHeight;
+}
+
+async function run() {
+  reset();
+  const btn = document.getElementById('btn');
+  btn.disabled = true;
+  document.getElementById('status').classList.add('visible');
+
+  let jobId;
+  try {
+    if (activeTab === 'file') {
+      if (!selectedFile) { showError('Select a file first.'); btn.disabled = false; return; }
+      if (selectedFile.size > 500 * 1024 * 1024) { showError('File exceeds 500 MB limit.'); btn.disabled = false; return; }
+      const form = new FormData();
+      form.append('file', selectedFile);
+      document.getElementById('status-text').textContent = 'Uploading…';
+      const res = await fetch('/transcribe/file', { method: 'POST', body: form });
+      if (!res.ok) { showError(await res.text()); btn.disabled = false; return; }
+      jobId = (await res.json()).job_id;
+    } else {
+      const url = document.getElementById('url-input').value.trim();
+      if (!url) { showError('Enter a URL first.'); btn.disabled = false; return; }
+      document.getElementById('status-text').textContent = 'Submitting…';
+      const form = new FormData();
+      form.append('url', url);
+      const res = await fetch('/transcribe/url', { method: 'POST', body: form });
+      if (!res.ok) { showError(await res.text()); btn.disabled = false; return; }
+      jobId = (await res.json()).job_id;
+    }
+  } catch (e) {
+    showError('Request failed: ' + e.message);
+    btn.disabled = false;
+    return;
+  }
+
+  document.getElementById('status-text').textContent = 'Processing…';
+
+  // Stream progress via SSE
+  const es = new EventSource(`/progress/${jobId}`);
+  es.onmessage = (e) => {
+    const data = JSON.parse(e.data);
+    if (data.msg) appendLog(data.msg);
+    if (data.done) {
+      es.close();
+      document.getElementById('status').classList.remove('visible');
+      btn.disabled = false;
+      if (data.error) {
+        showError(data.error);
+      } else {
+        transcriptText = data.transcript;
+        showResult(data.transcript);
+      }
+    }
+  };
+  es.onerror = () => {
+    es.close();
+    document.getElementById('status').classList.remove('visible');
+    btn.disabled = false;
+    showError('Connection lost during processing.');
+  };
+}
+
+function showResult(text) {
+  const blob = new Blob([text], { type: 'text/plain' });
+  const url = URL.createObjectURL(blob);
+  const link = document.getElementById('dl-link');
+  link.href = url;
+  link.download = 'transcript.txt';
+  document.getElementById('result').classList.add('visible');
+}
+
+function showError(msg) {
+  const el = document.getElementById('error');
+  el.textContent = msg;
+  el.classList.add('visible');
+  document.getElementById('status').classList.remove('visible');
+}
+
+document.getElementById('url-input').addEventListener('keydown', e => { if (e.key === 'Enter') run(); });
+</script>
+</body>
+</html>"""

requirements.txt

@@ -0,0 +1,5 @@
+fastapi
+uvicorn[standard]
+faster-whisper
+yt-dlp
+python-multipart