"use client";

import { useEffect, useRef, useState } from "react";
import Webcam from "react-webcam";
import * as faceapi from "face-api.js";
import { Button } from "@/components/ui/button";
import { Camera } from "lucide-react";
import { useToast } from "@/hooks/use-toast";

const MODEL_URL = "https://cdn.jsdelivr.net/npm/@vladmandic/face-api/model";
const PADDING = 50; // Padding around face in pixels

const RealtimeFaceDetection = () => {
  const webcamRef = useRef<Webcam>(null);
  const canvasRef = useRef<HTMLCanvasElement>(null);
  const [isModelLoaded, setIsModelLoaded] = useState(false);
  const [isDetecting, setIsDetecting] = useState(false);
  const { toast } = useToast();

  useEffect(() => {
    const loadModels = async () => {
      try {
        await faceapi.nets.tinyFaceDetector.loadFromUri(MODEL_URL);
        await faceapi.nets.faceLandmark68Net.loadFromUri(MODEL_URL);
        await faceapi.nets.faceRecognitionNet.loadFromUri(MODEL_URL);
        setIsModelLoaded(true);
      } catch (error) {
        console.error("Error loading models:", error);
        toast({
          title: "Error",
          description: "Failed to load face detection models.",
          variant: "destructive",
        });
      }
    };
    loadModels();
  }, [toast]);

  const extractFaceWithPadding = (
    video: HTMLVideoElement,
    box: faceapi.Box
  ): HTMLCanvasElement => {
    const canvas = document.createElement("canvas");
    const context = canvas.getContext("2d");

    // Calculate padded dimensions
    const x = Math.max(0, box.x - PADDING);
    const y = Math.max(0, box.y - PADDING);
    const width = Math.min(video.videoWidth - x, box.width + 2 * PADDING);
    const height = Math.min(video.videoHeight - y, box.height + 2 * PADDING);

    canvas.width = width;
    canvas.height = height;

    if (context) {
      // Extract face region with padding
      context.drawImage(
        video,
        x,
        y,
        width,
        height, // Source coordinates
        0,
        0,
        width,
        height // Destination coordinates
      );
    }

    return canvas;
  };

  const detectFace = async () => {
    if (!webcamRef.current?.video || !canvasRef.current) return;

    const video = webcamRef.current.video;
    const canvas = canvasRef.current;
    const context = canvas.getContext("2d");

    if (!context) return;

    canvas.width = video.videoWidth;
    canvas.height = video.videoHeight;
    context.clearRect(0, 0, canvas.width, canvas.height); // Clear previous drawings

    // Mirror the canvas context to match the mirrored video
    context.translate(canvas.width, 0); // Move the origin to the right side of the canvas
    context.scale(-1, 1); // Flip the context horizontally

    // Detect all faces
    const detections = await faceapi
      .detectAllFaces(video, new faceapi.TinyFaceDetectorOptions())
      .withFaceLandmarks()
      .withFaceDescriptors();

    for (const detection of detections) {
      // Draw box for visualization
      const { box } = detection.detection;
      context.strokeStyle = "#00FF00";
      context.lineWidth = 2;
      context.strokeRect(box.x, box.y, box.width, box.height);

      // Extract face with padding and send to API
      const faceCanvas = extractFaceWithPadding(video, box);
      faceCanvas.toBlob(
        (blob) => {
          if (blob) sendFaceDataToAPI(blob);
        },
        "image/jpeg",
        0.95
      );
    }
  };

  const sendFaceDataToAPI = async (imageBlob: Blob) => {
    try {
      const formData = new FormData();
      formData.append("image", imageBlob, "face.jpg");

      const response = await fetch(
        `${process.env.NEXT_PUBLIC_BASE_URL}/search`,
        {
          method: "POST",
          body: formData,
        }
      );

      const data = await response.json();
      toast({ title: data?.name, description: data.message });
    } catch (error) {
      console.error("Error sending face data:", error);
      toast({
        title: "Error",
        description: "Failed to send face data.",
        variant: "destructive",
      });
    }
  };

  const startDetection = () => {
    if (!isModelLoaded) return;
    setIsDetecting(true);
    setInterval(detectFace, 300);
  };
  return (
    <div className="max-w-3xl mx-auto">
      <div className="relative">
        <Webcam ref={webcamRef} mirrored className="w-full rounded-lg" />
        <canvas
          ref={canvasRef}
          className="absolute top-0 left-0 w-full h-full"
        />
      </div>
      <div className="mt-6 flex justify-center">
        <Button
          onClick={startDetection}
          disabled={!isModelLoaded || isDetecting}
        >
          <Camera className="mr-2 h-4 w-4" />
          {isDetecting ? "Detecting..." : "Start Realtime Detection"}
        </Button>
      </div>
    </div>
  );
};

export default RealtimeFaceDetection;