Create app.py

app.py

@@ -0,0 +1,34 @@
+from fastapi import FastAPI
+from pydantic import BaseModel
+import subprocess, base64, os, uuid, shutil, whisper
+
+app = FastAPI()
+whisper_model = whisper.load_model("tiny")
+
+class VideoJsonRequest(BaseModel):
+    video_base64: str
+
+def to_b64(path):
+    with open(path, "rb") as f: return base64.b64encode(f.read()).decode('utf-8')
+
+@app.post("/process-video")
+async def process(req: VideoJsonRequest):
+    uid = str(uuid.uuid4())
+    tmp = f"/tmp/{uid}"
+    os.makedirs(tmp)
+    v_p = f"{tmp}/v.mp4"
+    a_p = f"{tmp}/a.wav"
+    try:
+        with open(v_p, "wb") as f: f.write(base64.b64decode(req.video_base64))
+        probe = subprocess.run(["ffprobe", "-v", "error", "-show_entries", "format=duration", "-of", "default=noprint_wrappers=1:nokey=1", v_p], capture_output=True, text=True).stdout
+        dur = float(probe.strip() or 0)
+        # Extract 15 frames
+        subprocess.run(["ffmpeg", "-y", "-i", v_p, "-vf", f"fps={15/max(dur,1)}", "-vframes", "15", "-q:v", "5", f"{tmp}/f_%03d.jpg"])
+        # Transcribe
+        subprocess.run(["ffmpeg", "-y", "-i", v_p, "-vn", "-acodec", "pcm_s16le", "-ar", "16000", "-ac", "1", a_p])
+        txt = whisper_model.transcribe(a_p)["text"].strip()
+        f_names = sorted([f"{tmp}/{f}" for f in os.listdir(tmp) if f.startswith("f_")])
+        imgs = [to_b64(f) for f in f_names]
+        return {"success": True, "transcript": txt, "frames": imgs, "thumbnail": imgs[0] if imgs else None}
+    except Exception as e: return {"success": False, "error": str(e)}
+    finally: shutil.rmtree(tmp, ignore_errors=True)