app.py

import gradio as gr
import cv2
import easyocr
import numpy as np
import os
from PIL import Image
from ultralytics import YOLO
from datetime import datetime

# Load YOLO model
model = YOLO("/home/user/app/best.pt")

# Label map
label_map = {0: "Analog", 1: "Digital", 2: "Non-LP"}

# EasyOCR Bengali
reader = easyocr.Reader(['bn'])

def annotate_frame(frame):
    input_img = cv2.resize(frame, (640, 640))
    results = model(input_img)[0]
    detections = results.boxes.data.cpu().numpy()

    for det in detections:
        if len(det) < 6:
            continue

        x1, y1, x2, y2, conf, cls = det
        x1, y1, x2, y2 = map(int, [x1, y1, x2, y2])
        label = label_map.get(int(cls), "Unknown")
        percent = f"{conf * 100:.2f}%"

        # Draw box and label
        cv2.rectangle(input_img, (x1, y1), (x2, y2), (255, 0, 0), 2)
        cv2.putText(input_img, f"{label}: {percent}", (x1, y1 - 10),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)

        # OCR
        cropped = frame[y1:y2, x1:x2]
        if cropped.size > 0:
            ocr_result = reader.readtext(cropped)
            for i, item in enumerate(ocr_result):
                text = item[1].strip()
                conf = item[2]
                cv2.putText(input_img, text, (x1, y2 + 20 + i*25),
                            cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 255), 2)

    return cv2.cvtColor(input_img, cv2.COLOR_BGR2RGB)

def process_input(input_file):
    file_path = input_file.name
    ext = os.path.splitext(file_path)[-1].lower()

    if ext in ['.mp4', '.avi', '.mov']:
        cap = cv2.VideoCapture(file_path)
        if not cap.isOpened():
            return None, None, "Could not open video file", ""

        fps = cap.get(cv2.CAP_PROP_FPS)
        frame_skip = 5
        frame_id = 0
        output_path = "annotated_output.mp4"
        out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (640, 640))

        while True:
            ret, frame = cap.read()
            if not ret:
                break
            if frame_id % frame_skip != 0:
                frame_id += 1
                continue

            frame_id += 1
            frame = cv2.resize(frame, (640, 640))
            annotated = annotate_frame(frame)
            out.write(cv2.cvtColor(annotated, cv2.COLOR_RGB2BGR))

        cap.release()
        out.release()

        return output_path, None, "Bangla text in video (see frames)", "OCR confidence displayed"

    else:
        frame = cv2.imread(file_path)
        if frame is None:
            return None, None, "Invalid image", ""

        frame = cv2.resize(frame, (640, 640))
        annotated = annotate_frame(frame)
        pil_img = Image.fromarray(annotated)
        return None, pil_img, "Bangla text in image", "OCR confidence in image"



interface = gr.Interface(
    fn=process_input,
    inputs=gr.File(type="filepath", label="Upload Image or Video"),
    outputs=[
        gr.Video(label="Output Video"),
        gr.Image(type="pil", label="Output Image"),
        gr.Textbox(label="Detected Text (Bangla)"),
        gr.Textbox(label="Confidence (%)")
    ],
    title="YOLOv5 License Plate Detector (Bangla OCR)",
    description="Upload an image or video. Detects license plates and extracts Bangla text using EasyOCR."
)

interface.launch()