usda-vision/media-api/main.py

import os
import pathlib
import subprocess
import urllib.parse
from typing import List, Optional

from fastapi import FastAPI, HTTPException, Response, Request
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse, StreamingResponse

MEDIA_DIR = pathlib.Path(os.getenv("MEDIA_VIDEOS_DIR", "/mnt/videos")).resolve()
THUMBS_DIR = pathlib.Path(os.getenv("MEDIA_THUMBS_DIR", MEDIA_DIR / ".thumbnails")).resolve()

app = FastAPI(title="Media API", version="0.1.0")

# CORS for dashboard at exp-dash:8080 (and localhost for convenience)
app.add_middleware(
    CORSMiddleware,
    allow_origins=["http://exp-dash:8080", "http://localhost:8080"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"]
)


def list_video_files() -> List[pathlib.Path]:
    exts = {".mp4", ".avi", ".mov", ".mkv"}
    files: List[pathlib.Path] = []
    for root, _, fnames in os.walk(MEDIA_DIR):
        # skip thumbnails dir
        if THUMBS_DIR.as_posix() in pathlib.Path(root).as_posix():
            continue
        for f in fnames:
            p = pathlib.Path(root) / f
            if p.suffix.lower() in exts:
                files.append(p)
    return sorted(files, key=lambda p: p.stat().st_mtime, reverse=True)


def file_id_from_path(p: pathlib.Path) -> str:
    return urllib.parse.quote_plus(p.relative_to(MEDIA_DIR).as_posix())


def path_from_file_id(fid: str) -> pathlib.Path:
    rel = urllib.parse.unquote_plus(fid)
    p = (MEDIA_DIR / rel).resolve()
    if not p.is_file() or MEDIA_DIR not in p.parents:
        raise HTTPException(status_code=404, detail="Video not found")
    return p


def ensure_thumbnail(p: pathlib.Path, width: int = 320, height: int = 180) -> pathlib.Path:
    THUMBS_DIR.mkdir(parents=True, exist_ok=True)
    thumb_name = file_id_from_path(p) + f"_{width}x{height}.jpg"
    out = THUMBS_DIR / thumb_name
    if out.exists():
        return out
    # generate from 1s offset or 0 if too short
    cmd = [
        "ffmpeg", "-y",
        "-ss", "1",
        "-i", str(p),
        "-vframes", "1",
        "-vf", f"scale='min({width},iw)':-2,scale={width}:{height}:force_original_aspect_ratio=decrease,pad={width}:{height}:(ow-iw)/2:(oh-ih)/2",
        str(out)
    ]
    try:
        subprocess.run(cmd, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
    except subprocess.CalledProcessError:
        raise HTTPException(status_code=500, detail="Failed to generate thumbnail")
    return out


@app.get("/health")
def health() -> dict:
    return {"status": "ok"}


@app.get("/videos/")
def list_videos(camera_name: Optional[str] = None, start_date: Optional[str] = None, end_date: Optional[str] = None,
                limit: int = 24, page: int = 1) -> dict:
    files = list_video_files()
    # basic filtering placeholder (camera_name can be inferred from path segments)
    if camera_name:
        files = [p for p in files if camera_name in p.as_posix()]
    total = len(files)
    start = max(0, (page - 1) * limit)
    slice_ = files[start:start + limit]
    items = []
    for p in slice_:
        stat = p.stat()
        items.append({
            "file_id": file_id_from_path(p),
            "camera_name": p.parent.name,
            "filename": p.name,
            "file_size_bytes": stat.st_size,
            "format": p.suffix.lstrip('.'),
            "status": "completed",
            "created_at": int(stat.st_mtime),
            "is_streamable": True,
            "needs_conversion": False,
        })
    total_pages = (total + limit - 1) // limit if limit else 1
    return {
        "videos": items,
        "total_count": total,
        "page": page,
        "total_pages": total_pages,
    }


@app.get("/videos/{file_id:path}/thumbnail")
def get_thumbnail(file_id: str, width: int = 320, height: int = 180):
    p = path_from_file_id(file_id)
    thumb = ensure_thumbnail(p, width, height)
    return FileResponse(thumb, media_type="image/jpeg")

# Convenience endpoint: pass file_id via query instead of path (accepts raw or URL-encoded)
@app.get("/videos/thumbnail")
def get_thumbnail_q(file_id: str, width: int = 320, height: int = 180):
    p = path_from_file_id(file_id)
    thumb = ensure_thumbnail(p, width, height)
    return FileResponse(thumb, media_type="image/jpeg")


def get_video_mime_type(path: pathlib.Path) -> str:
    """Get MIME type based on file extension"""
    ext = path.suffix.lower()
    mime_types = {
        ".mp4": "video/mp4",
        ".avi": "video/x-msvideo",
        ".mov": "video/quicktime",
        ".mkv": "video/x-matroska",
        ".webm": "video/webm",
    }
    return mime_types.get(ext, "video/mp4")


def open_file_range(path: pathlib.Path, start: int, end: Optional[int]):
    file_size = path.stat().st_size
    if end is None or end >= file_size:
        end = file_size - 1
    length = end - start + 1
    with open(path, 'rb') as f:
        f.seek(start)
        chunk = f.read(length)
    return chunk, file_size, start, end


@app.head("/videos/{file_id:path}/stream")
@app.get("/videos/{file_id:path}/stream")
def stream_file(request: Request, file_id: str):
    p = path_from_file_id(file_id)
    file_size = p.stat().st_size
    content_type = get_video_mime_type(p)
    
    range_header = request.headers.get("range")
    
    # Base headers for all responses
    base_headers = {
        "Accept-Ranges": "bytes",
        "Content-Type": content_type,
        "Cache-Control": "public, max-age=3600",
        "Access-Control-Allow-Origin": "*",
        "Access-Control-Allow-Methods": "GET, HEAD, OPTIONS",
        "Access-Control-Allow-Headers": "Range, Content-Type",
        "Access-Control-Expose-Headers": "Content-Range, Accept-Ranges, Content-Length",
    }
    
    # For HEAD requests, just return headers without body
    if request.method == "HEAD":
        headers = {
            **base_headers,
            "Content-Length": str(file_size),
        }
        return Response(status_code=200, headers=headers)
    
    if not range_header:
        # No range request - return full file with proper headers
        headers = {
            **base_headers,
            "Content-Length": str(file_size),
        }
        return FileResponse(
            p,
            media_type=content_type,
            headers=headers,
            status_code=200
        )
    
    # Parse range request: bytes=START-END
    range_value = range_header.strip().lower().replace("bytes=", "")
    start_str, _, end_str = range_value.partition("-")
    try:
        start = int(start_str) if start_str else 0
        end = int(end_str) if end_str else None
    except ValueError:
        raise HTTPException(status_code=400, detail="Bad Range header")
    
    # Validate range
    if start < 0:
        start = 0
    if end is None or end >= file_size:
        end = file_size - 1
    if start > end:
        raise HTTPException(status_code=416, detail="Range Not Satisfiable")
    
    chunk, size, actual_start, actual_end = open_file_range(p, start, end)
    
    headers = {
        **base_headers,
        "Content-Range": f"bytes {actual_start}-{actual_end}/{size}",
        "Content-Length": str(len(chunk)),
    }
    return Response(content=chunk, status_code=206, headers=headers)

# Convenience endpoint: pass file_id via query instead of path (accepts raw or URL-encoded)
@app.head("/videos/stream")
@app.get("/videos/stream")
def stream_file_q(request: Request, file_id: str):
    return stream_file(request, file_id)


# Handle OPTIONS requests for CORS preflight
@app.options("/videos/{file_id:path}/stream")
@app.options("/videos/stream")
@app.options("/videos/{file_id:path}/stream-transcoded")
@app.options("/videos/stream-transcoded")
def stream_options():
    return Response(
        status_code=200,
        headers={
            "Access-Control-Allow-Origin": "*",
            "Access-Control-Allow-Methods": "GET, HEAD, OPTIONS",
            "Access-Control-Allow-Headers": "Range, Content-Type",
            "Access-Control-Max-Age": "86400",
        }
    )


def generate_transcoded_stream(file_path: pathlib.Path, start_time: float = 0.0):
    """
    Transcode video to H.264 on-the-fly using FFmpeg.
    Streams H.264/MP4 that browsers can actually play.
    """
    if not file_path.exists():
        raise HTTPException(status_code=404, detail="Video file not found")
    
    if file_path.stat().st_size == 0:
        raise HTTPException(status_code=500, detail="Video file is empty (0 bytes)")
    
    # FFmpeg command to transcode to H.264 with web-optimized settings
    cmd = [
        "ffmpeg",
        "-i", str(file_path),
        "-c:v", "libx264",  # H.264 codec
        "-preset", "ultrafast",  # Fast encoding for real-time
        "-tune", "zerolatency",  # Low latency
        "-crf", "23",  # Quality (18-28, lower = better)
        "-c:a", "aac",  # AAC audio if present
        "-movflags", "+faststart",  # Web-optimized (moov atom at beginning)
        "-f", "mp4",  # MP4 container
        "-"  # Output to stdout
    ]
    
    # If seeking to specific time
    if start_time > 0:
        cmd.insert(-2, "-ss")
        cmd.insert(-2, str(start_time))
    
    try:
        process = subprocess.Popen(
            cmd,
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
            bufsize=8192
        )
        
        # Stream chunks
        chunk_size = 8192
        bytes_yielded = 0
        while True:
            chunk = process.stdout.read(chunk_size)
            if not chunk:
                break
            bytes_yielded += len(chunk)
            yield chunk
        
        # Check for errors
        process.wait()
        if process.returncode != 0:
            stderr = process.stderr.read().decode('utf-8', errors='ignore')
            print(f"FFmpeg error (code {process.returncode}): {stderr}")
            if bytes_yielded == 0:
                raise HTTPException(status_code=500, detail=f"FFmpeg transcoding failed: {stderr[:200]}")
        
    except HTTPException:
        raise
    except Exception as e:
        print(f"FFmpeg transcoding error: {e}")
        raise HTTPException(status_code=500, detail=f"Transcoding error: {str(e)}")


@app.head("/videos/{file_id:path}/stream-transcoded")
@app.get("/videos/{file_id:path}/stream-transcoded")
@app.head("/videos/stream-transcoded")
@app.get("/videos/stream-transcoded")
def stream_transcoded(request: Request, file_id: str, start_time: float = 0.0):
    """
    Stream video transcoded to H.264 on-the-fly.
    This endpoint converts MPEG-4 Part 2 videos to H.264 for browser compatibility.
    """
    p = path_from_file_id(file_id)
    content_type = "video/mp4"
    
    # Base headers
    headers = {
        "Content-Type": content_type,
        "Cache-Control": "no-cache, no-store, must-revalidate",
        "Access-Control-Allow-Origin": "*",
        "Access-Control-Allow-Methods": "GET, HEAD, OPTIONS",
        "Access-Control-Allow-Headers": "Range, Content-Type",
        "Access-Control-Expose-Headers": "Content-Range, Accept-Ranges, Content-Length",
    }
    
    # For HEAD requests, just return headers (we can't know size without transcoding)
    if request.method == "HEAD":
        return Response(status_code=200, headers=headers)
    
    # Note: Range requests are complex with transcoding, so we'll transcode from start
    # For better performance with range requests, we'd need to cache transcoded segments
    return StreamingResponse(
        generate_transcoded_stream(p, start_time),
        media_type=content_type,
        headers=headers
    )