app.py

from fastapi import FastAPI
from pydantic import BaseModel
from typing import Optional
import subprocess, base64, os, uuid, shutil, whisper
   
app = FastAPI()
whisper_model = whisper.load_model("tiny")

# 🚨 DYNAMIC SCHEMA
class VideoJsonRequest(BaseModel):
    video_base64: str
    num_frames: Optional[int] = 15          # Default to 15
    get_transcript: Optional[bool] = True   # Default to True

def to_b64(path):
    with open(path, "rb") as f: return base64.b64encode(f.read()).decode('utf-8')

@app.get("/")
async def index():
    return {"success": True, "engine": "Dynamic Viral Cat Media Server"}
    
@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))
        
        # Get Duration
        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)
        
        # 🚨 DYNAMIC FRAME MATH: Spreads requested frames evenly across duration
        calc_fps = req.num_frames / max(dur, 1)
        
        print("-------------------")
        print(f"Requested Frames: {req.num_frames} | Duration: {dur:.2f}s | Calculated FPS: {calc_fps:.2f}")
        print(f"Transcript requested: {req.get_transcript}")
        print("-------------------")

        # Extract X frames
        subprocess.run(["ffmpeg", "-y", 
                        "-loglevel", "error", 
                        "-i", v_p, "-vf", f"fps={calc_fps}", 
                        "-vframes", str(req.num_frames), 
                        "-q:v", "5", f"{tmp}/f_%03d.jpg"])
        
        # 🚨 CONDITIONAL TRANSCRIPT
        txt = ""
        if req.get_transcript:
            subprocess.run(["ffmpeg", "-y",
                            "-loglevel", "error", 
                            "-i", v_p, "-vn", "-acodec", "pcm_s16le", "-ar", "16000", "-ac", "1", a_p])
            
            if os.path.exists(a_p):
                result = whisper_model.transcribe(a_p)
                lines = [f"[{s['start']:.2f}] {s['text'].strip()}" for s in result["segments"]]
                txt = "\n".join(lines)

        # Gather frames
        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]
        
        print("-------------------")
        print(f"Successfully extracted {len(imgs)} images.")
        print("-------------------")
        
        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)