// server/utils/recognitionProcessor.js
import TextDirectionClassifier from './textDirectionClassifier.js';
import TextRecognizer from './textRecognizer.js';
import TextRegionCropper from './textRegionCropper.js';

class RecognitionProcessor {
    constructor() {
        this.recSession = null;
        this.clsSession = null;
        this.config = null;

        this.textDirectionClassifier = new TextDirectionClassifier();
        this.textRecognizer = new TextRecognizer();
        this.textRegionCropper = new TextRegionCropper();
    }

    initialize(recSession, clsSession, config) {
        this.recSession = recSession;
        this.clsSession = clsSession;
        this.config = config;

        this.textDirectionClassifier.initialize(clsSession, config);
        this.textRecognizer.initialize(recSession, config);
    }

    async loadCharacterSet(keysPath) {
        await this.textRecognizer.loadCharacterSet(keysPath);
    }

    getCharacterSetSize() {
        return this.textRecognizer.getCharacterSetSize();
    }

    async recognizeTextWithCls(processedImage, textBoxes) {
        const results = [];

        try {
            console.log(`🔄 开始处理 ${textBoxes.length} 个文本区域`);

            for (let i = 0; i < textBoxes.length; i++) {
                const box = textBoxes[i];

                try {
                    console.log(`\n📦 处理区域 ${i + 1}/${textBoxes.length}, 置信度: ${box.confidence.toFixed(4)}`);

                    const textRegion = await this.textRegionCropper.cropTextRegion(
                        processedImage.buffer, box, i + 1
                    );
                    if (!textRegion) {
                        console.log(`⏭️ 区域 ${i + 1}: 跳过无效区域`);
                        continue;
                    }

                    const { clsResult, clsConfidence } = await this.textDirectionClassifier.classifyTextDirection(
                        textRegion.buffer
                    );

                    let recognitionImage = textRegion.buffer;
                    if (clsResult === 180 && clsConfidence > this.config.clsThresh) {
                        console.log(`🔄 区域 ${i + 1}: 旋转 180°`);
                        recognitionImage = await this.textRegionCropper.rotateImage(textRegion.buffer, 180);
                    }

                    const textResult = await this.textRecognizer.recognizeText(recognitionImage, i + 1);

                    if (textResult.text && textResult.text.trim().length > 0 && textResult.confidence > 0.05) {
                        const originalBox = this.scaleBoxToOriginalImage(box, processedImage);

                        results.push({
                            text: textResult.text.trim(),
                            confidence: textResult.confidence * clsConfidence,
                            box: originalBox,
                            clsResult,
                            clsConfidence,
                            regionIndex: i + 1
                        });

                        console.log(`✅ 区域 ${i + 1}: 识别成功 "${textResult.text}"`);
                    } else {
                        console.log(`❌ 区域 ${i + 1}: 识别失败或置信度过低`);
                    }

                } catch (error) {
                    console.error(`💥 区域 ${i + 1}: 处理失败`, error.message);
                    continue;
                }
            }

            console.log(`\n🎯 识别完成: ${results.length}/${textBoxes.length} 个区域成功`);
            return results;

        } catch (error) {
            console.error('❌ 整体识别失败:', error);
            throw error;
        }
    }

    scaleBoxToOriginalImage(box, processedImage) {
        const {
            scaleX, scaleY,
            paddingX, paddingY,
            originalWidth, originalHeight
        } = processedImage;

        const paddedX1 = box.x1 * scaleX;
        const paddedY1 = box.y1 * scaleY;
        const paddedX3 = box.x3 * scaleX;
        const paddedY3 = box.y3 * scaleY;

        const originalX1 = paddedX1 - paddingX;
        const originalY1 = paddedY1 - paddingY;
        const originalX3 = paddedX3 - paddingX;
        const originalY3 = paddedY3 - paddingY;

        const clamp = (value, max) => Math.max(0, Math.min(max, value));

        return {
            x1: clamp(originalX1, originalWidth - 1),
            y1: clamp(originalY1, originalHeight - 1),
            x2: clamp(originalX3, originalWidth - 1),
            y2: clamp(originalY1, originalHeight - 1),
            x3: clamp(originalX3, originalWidth - 1),
            y3: clamp(originalY3, originalHeight - 1),
            x4: clamp(originalX1, originalWidth - 1),
            y4: clamp(originalY3, originalHeight - 1),
            confidence: box.confidence
        };
    }
}

export default RecognitionProcessor;