Movie Company And Year Image Scanner Identifier

/**
 * Movie Company and Year Image Scanner Identifier
 * Scans an image for text using OCR to identify a movie company and a release year.
 * 
 * @param {HTMLImageElement} originalImg - The original image to be processed
 * @param {string} language - English by default ('eng')
 * @returns {Promise<HTMLCanvasElement>} - A promise that resolves to a Canvas containing the result
 */
async function processImage(originalImg, language = "eng") {
    // Canvas setup
    const canvas = document.createElement('canvas');
    const ctx = canvas.getContext('2d');

    const width = originalImg.width;
    const height = originalImg.height;
    
    // UI Panel config
    const textPanelHeight = 140;
    const minWidth = 600;
    const canvasWidth = Math.max(width, minWidth);
    const canvasHeight = height + textPanelHeight;

    canvas.width = canvasWidth;
    canvas.height = canvasHeight;

    // 1. Draw the background
    ctx.fillStyle = '#0a0a0a';
    ctx.fillRect(0, 0, canvasWidth, canvasHeight);

    // 2. Draw the original image centered horizontally
    const imgX = (canvasWidth - width) / 2;
    ctx.drawImage(originalImg, imgX, 0, width, height);

    // 3. Draw "Scanner" aesthetic overlays (cyan corners over the image)
    ctx.strokeStyle = "#00ffcc";
    ctx.lineWidth = 4;
    const padding = Math.min(20, width * 0.1, height * 0.1);
    const len = Math.min(40, width * 0.2, height * 0.2);

    // Top-left
    ctx.beginPath(); ctx.moveTo(imgX + padding, padding + len); ctx.lineTo(imgX + padding, padding); ctx.lineTo(imgX + padding + len, padding); ctx.stroke();
    // Top-right
    ctx.beginPath(); ctx.moveTo(imgX + width - padding - len, padding); ctx.lineTo(imgX + width - padding, padding); ctx.lineTo(imgX + width - padding, padding + len); ctx.stroke();
    // Bottom-left
    ctx.beginPath(); ctx.moveTo(imgX + padding, height - padding - len); ctx.lineTo(imgX + padding, height - padding); ctx.lineTo(imgX + padding + len, height - padding); ctx.stroke();
    // Bottom-right
    ctx.beginPath(); ctx.moveTo(imgX + width - padding - len, height - padding); ctx.lineTo(imgX + width - padding, height - padding); ctx.lineTo(imgX + width - padding, height - padding - len); ctx.stroke();

    // 4. Load Tesseract.js dynamically for OCR
    if (typeof window.Tesseract === 'undefined') {
        await new Promise((resolve, reject) => {
            const script = document.createElement('script');
            script.src = 'https://cdn.jsdelivr.net/npm/tesseract.js@4/dist/tesseract.min.js';
            script.crossOrigin = 'anonymous';
            script.onload = () => resolve();
            script.onerror = () => reject(new Error('Failed to load Tesseract.js'));
            document.head.appendChild(script);
        });
    }

    let foundYear = "None identified";
    let foundCompany = "None identified";
    let rawResultText = "";

    try {
        // Run OCR Engine on the original image
        const result = await window.Tesseract.recognize(originalImg, language);
        const text = result.data.text || "";
        rawResultText = text.replace(/\n\s*\n/g, ' ').replace(/\s+/g, ' ').trim();

        // Regex to find a standard movie release year (1880 - 2099)
        const yearRegex = /\b(18[8-9]\d|19\d{2}|20[0-2]\d)\b/g;
        const years = text.match(yearRegex);
        if (years && years.length > 0) {
            foundYear = years[0]; 
        }

        // Expanded list of popular production companies (ordered roughly by specificity)
        const companies = [
            "20th Century Fox", "Twentieth Century Fox", "Warner Bros", "Universal Pictures",
            "Paramount Pictures", "Walt Disney", "Sony Pictures", "Columbia Pictures",
            "Metro-Goldwyn-Mayer", "New Line Cinema", "DreamWorks", "Miramax",
            "Lionsgate", "Pixar", "Marvel Studios", "Lucasfilm", "A24", "Legendary",
            "TriStar", "Orion", "RKO", "United Artists", "Amblin", "Touchstone",
            "Universal", "Paramount", "Disney", "Columbia", "Sony", "Fox", "MGM"
        ];

        // Normalize OCR text (remove spaces/punctuation for a robust inner-match)
        const normalizedText = text.replace(/[^a-zA-Z0-9]/g, '').toUpperCase();

        for (let comp of companies) {
            const normalizedComp = comp.replace(/[^a-zA-Z0-9]/g, '').toUpperCase();
            if (normalizedText.includes(normalizedComp)) {
                foundCompany = comp;
                break;
            }
        }
    } catch (err) {
        console.error("Scanner Identifier Error:", err);
        foundCompany = "Error reading image";
        foundYear = "Error";
    }

    // 5. Draw Results Panel
    const panelY = height;

    // Panel Background
    ctx.fillStyle = "#1a1a1a";
    ctx.fillRect(0, panelY, canvasWidth, textPanelHeight);

    // Subtle neon border separator
    ctx.fillStyle = "#00ffcc";
    ctx.fillRect(0, panelY, canvasWidth, 2);

    // Panel Typography
    ctx.font = "bold 22px Arial, sans-serif";
    ctx.textAlign = "left";
    ctx.textBaseline = "top";
    
    // Panel Header
    ctx.fillStyle = "#ffffff";
    ctx.fillText("SCANNER IDENTIFIER RESULTS", 30, panelY + 15);

    // Data - Company
    ctx.font = "bold 18px Arial, sans-serif";
    ctx.fillStyle = "#00ffcc";
    ctx.fillText(`Production Company : ${foundCompany}`, 30, panelY + 55);

    // Data - Year
    ctx.fillStyle = "#ffcc00";
    ctx.fillText(`Release Year            : ${foundYear}`, 30, panelY + 85);

    // Show raw string preview if nothing gracefully identified
    if (foundCompany === "None identified" && rawResultText.length > 0) {
        ctx.fillStyle = "#888888";
        ctx.font = "italic 14px Arial, sans-serif";
        let displayRaw = rawResultText;
        if (displayRaw.length > 70) displayRaw = displayRaw.substring(0, 67) + '...';
        ctx.fillText(`Raw text segment: "${displayRaw}"`, 30, panelY + 115);
    }

    return canvas;
}