Spaces:

ulduldp
/

test-ffmpeg

Running

App Files Files Community

ulduldp commited on 13 days ago

Commit

4cd5d22

verified ·

1 Parent(s): 3172482

Create app.py

Browse files

Files changed (1) hide show

app.py +370 -0

app.py ADDED Viewed

	@@ -0,0 +1,370 @@

+from flask import Flask, render_template_string, request, jsonify
+import os
+import uuid
+import subprocess
+import tempfile
+import shutil
+import textwrap
+from werkzeug.utils import secure_filename
+from faster_whisper import WhisperModel
+app = Flask(__name__)
+UPLOAD_FOLDER = "uploads"
+OUTPUT_FOLDER = "static/videos"
+SUBTITLE_FOLDER = "subtitles"
+os.makedirs(UPLOAD_FOLDER, exist_ok=True)
+os.makedirs(OUTPUT_FOLDER, exist_ok=True)
+os.makedirs(SUBTITLE_FOLDER, exist_ok=True)
+# Load model once
+model = WhisperModel(
+    "base",
+    device="cpu",
+    compute_type="int8"
+)
+HTML = """
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Photo + Audio To Video</title>
+<style>
+*{
+    margin:0;
+    padding:0;
+    box-sizing:border-box;
+    font-family:Arial;
+}
+body{
+    background:#0f0f0f;
+    color:white;
+    min-height:100vh;
+    display:flex;
+    justify-content:center;
+    align-items:center;
+    padding:20px;
+}
+.container{
+    width:100%;
+    max-width:500px;
+    background:#1b1b1b;
+    border-radius:20px;
+    padding:25px;
+    box-shadow:0 0 20px rgba(0,0,0,0.4);
+}
+h1{
+    text-align:center;
+    margin-bottom:25px;
+    font-size:28px;
+}
+.upload-box{
+    border:2px dashed #444;
+    padding:20px;
+    border-radius:15px;
+    margin-bottom:20px;
+}
+label{
+    display:block;
+    margin-bottom:8px;
+    color:#ccc;
+}
+input{
+    width:100%;
+    padding:12px;
+    background:#2a2a2a;
+    border:none;
+    border-radius:10px;
+    color:white;
+    margin-bottom:15px;
+}
+button{
+    width:100%;
+    padding:15px;
+    border:none;
+    border-radius:12px;
+    background:#00aaff;
+    color:white;
+    font-size:18px;
+    cursor:pointer;
+    transition:0.3s;
+}
+button:hover{
+    opacity:0.9;
+}
+#loading{
+    display:none;
+    text-align:center;
+    margin-top:20px;
+}
+video{
+    width:100%;
+    margin-top:20px;
+    border-radius:15px;
+    display:none;
+}
+.download-btn{
+    display:none;
+    margin-top:15px;
+    text-align:center;
+}
+.download-btn a{
+    display:inline-block;
+    background:#22c55e;
+    color:white;
+    text-decoration:none;
+    padding:12px 20px;
+    border-radius:10px;
+}
+.preview{
+    margin-top:15px;
+    width:100%;
+    border-radius:15px;
+    display:none;
+}
+</style>
+</head>
+<body>
+<div class="container">
+    <h1>Photo + Audio → Video</h1>
+    <form id="form">
+        <div class="upload-box">
+            <label>Select Photo</label>
+            <input type="file" id="image" name="image" accept="image/*" required>
+            <img id="preview" class="preview">
+            <label>Select Audio (mp3/wav)</label>
+            <input type="file" name="audio" accept="audio/*" required>
+        </div>
+        <button type="submit">Generate Video</button>
+    </form>
+    <div id="loading">Generating Video...</div>
+    <video id="video" controls></video>
+    <div class="download-btn" id="downloadDiv">
+        <a id="downloadBtn" download>Download Video</a>
+    </div>
+</div>
+<script>
+const form = document.getElementById("form");
+const loading = document.getElementById("loading");
+const video = document.getElementById("video");
+const downloadBtn = document.getElementById("downloadBtn");
+const downloadDiv = document.getElementById("downloadDiv");
+const preview = document.getElementById("preview");
+document.getElementById("image").addEventListener("change", function(e){
+    const file = e.target.files[0];
+    if(file){
+        preview.src = URL.createObjectURL(file);
+        preview.style.display = "block";
+    }
+});
+form.addEventListener("submit", async (e)=>{
+    e.preventDefault();
+    loading.style.display = "block";
+    video.style.display = "none";
+    downloadDiv.style.display = "none";
+    const formData = new FormData(form);
+    try{
+        const response = await fetch("/generate",{
+            method:"POST",
+            body:formData
+        });
+        const data = await response.json();
+        loading.style.display = "none";
+        if(data.video_url){
+            video.src = data.video_url + "?t=" + new Date().getTime();
+            video.style.display = "block";
+            downloadBtn.href = data.video_url;
+            downloadDiv.style.display = "block";
+        }else{
+            alert(data.error || "Failed");
+        }
+    }catch(err){
+        loading.style.display = "none";
+        alert("Server Error");
+    }
+});
+</script>
+</body>
+</html>
+"""
+def ass_time(seconds: float) -> str:
+    if seconds < 0:
+        seconds = 0
+    h = int(seconds // 3600)
+    m = int((seconds % 3600) // 60)
+    s = seconds % 60
+    return f"{h}:{m:02d}:{s:05.2f}"
+def ass_escape(text: str) -> str:
+    # Escape ASS special chars
+    text = text.replace("\\", "\\\\")
+    text = text.replace("{", "\\{").replace("}", "\\}")
+    text = text.replace("\n", " ")
+    return text
+def make_ass_subtitles(segments, ass_path):
+    header = """[Script Info]
+ScriptType: v4.00+
+PlayResX: 1280
+PlayResY: 720
+ScaledBorderAndShadow: yes
+[V4+ Styles]
+Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
+Style: Default,Arial,26,&H00FFFFFF,&H000000FF,&H00000000,&HEE000000,0,0,0,0,100,100,0,0,3,0,0,2,40,40,35,1
+[Events]
+Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
+"""
+    lines = [header]
+    for seg in segments:
+        start = ass_time(seg["start"])
+        end = ass_time(seg["end"])
+        text = textwrap.fill(seg["text"].strip(), width=36)
+        text = ass_escape(text).replace("\n", r"\N")
+        lines.append(
+            f"Dialogue: 0,{start},{end},Default,,0,0,0,,{text}\n"
+        )
+    with open(ass_path, "w", encoding="utf-8") as f:
+        f.writelines(lines)
+@app.route("/")
+def home():
+    return render_template_string(HTML)
+@app.route("/generate", methods=["POST"])
+def generate():
+    if "image" not in request.files or "audio" not in request.files:
+        return jsonify({"error": "Missing files"})
+    image = request.files["image"]
+    audio = request.files["audio"]
+    if not image.filename or not audio.filename:
+        return jsonify({"error": "Please upload both image and audio"})
+    uid = str(uuid.uuid4())
+    image_name = secure_filename(image.filename)
+    audio_name = secure_filename(audio.filename)
+    image_path = os.path.join(UPLOAD_FOLDER, uid + "_" + image_name)
+    audio_path = os.path.join(UPLOAD_FOLDER, uid + "_" + audio_name)
+    output_filename = uid + ".mp4"
+    output_path = os.path.join(OUTPUT_FOLDER, output_filename)
+    ass_path = os.path.join(SUBTITLE_FOLDER, uid + ".ass")
+    image.save(image_path)
+    audio.save(audio_path)
+    try:
+        # 1) Transcribe audio
+        segments_iter, info = model.transcribe(
+            audio_path,
+            beam_size=5,
+            vad_filter=True
+        )
+        transcript = []
+        full_text_parts = []
+        for segment in segments_iter:
+            text = segment.text.strip()
+            if not text:
+                continue
+            transcript.append({
+                "start": round(segment.start, 2),
+                "end": round(segment.end, 2),
+                "text": text
+            })
+            full_text_parts.append(text)
+        full_text = " ".join(full_text_parts).strip()
+        # 2) Create ASS subtitles
+        make_ass_subtitles(transcript, ass_path)
+        # 3) FFmpeg command with smooth animation + subtitles
+        # Subtle zoom/pan effect + subtitles at bottom
+        filter_complex = (
+            f"[0:v]scale=1280:720:force_original_aspect_ratio=increase,"
+            f"crop=1280:720,"
+            f"zoompan=z='min(zoom+0.0008,1.08)':"
+            f"x='iw/2-(iw/zoom/2)':y='ih/2-(ih/zoom/2)':"
+            f"d=1:s=1280x720:fps=30,"
+            f"subtitles='{ass_path.replace(\"'\", \"\\\\'\")}'[v]"
+        )
+        cmd = [
+            "ffmpeg",
+            "-y",
+            "-loop", "1",
+            "-i", image_path,
+            "-i", audio_path,
+            "-filter_complex", filter_complex,
+            "-map", "[v]",
+            "-map", "1:a:0",
+            "-c:v", "libx264",
+            "-pix_fmt", "yuv420p",
+            "-c:a", "aac",
+            "-b:a", "192k",
+            "-shortest",
+            output_path
+        ]
+        subprocess.run(
+            cmd,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            check=True
+        )
+        return jsonify({
+            "video_url": f"/static/videos/{output_filename}",
+            "transcript": transcript,
+            "full_text": full_text,
+            "language": getattr(info, "language", None)
+        })
+    except subprocess.CalledProcessError as e:
+        return jsonify({
+            "error": "FFmpeg failed",
+            "details": e.stderr.decode("utf-8", errors="ignore")
+        })
+    except Exception as e:
+        return jsonify({
+            "error": "Processing failed",
+            "details": str(e)
+        })
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=7860)