优化 Tesseract.js 处理多栏图像的文本识别

import React, { useState, useEffect } from "react";
import Tesseract from "tesseract.js";
import ClipboardJS from "clipboard";
import Select from "react-select";

// ... (languageOptions 和其他组件代码保持不变)

const ImagesToText = () => {
  const [isLoading, setIsLoading] = useState(false);
  const [images, setImages] = useState([]);
  const [texts, setTexts] = useState([]);
  const [progress, setProgress] = useState(0);
  const [currentImageIndex, setCurrentImageIndex] = useState(0);
  const [errorMessage, setErrorMessage] = useState("");
  const [errorLanguagesMessage, setErrorLanguagesMessage] = useState("");
  const [selectedLanguages, setSelectedLanguages] = useState([]);

  // ... (handleImageUpload, handleCopyText, handleDownloadText, useEffect, handleReset 等函数保持不变)

  const handleSubmit = async () => {
    if (images.length === 0) {
      setErrorMessage("Select an image to convert.");
      return;
    }

    if (selectedLanguages.length === 0) {
      setErrorLanguagesMessage("Select any language.");
      return;
    }

    setIsLoading(true);
    setProgress(0);
    setTexts([]);
    setCurrentImageIndex(0);
    setErrorMessage("");
    setErrorLanguagesMessage("");

    const totalImages = images.length;
    let processedImages = 0;

    if (Array.isArray(images)) {
      for (const [index, image] of images?.entries()) {
        setCurrentImageIndex(index + 1);

        try {
          const result = await Tesseract.recognize(
            image,
            selectedLanguages.map((lang) => lang.value).join("+"),
            {
              // 在这里添加 PSM 配置
              // 推荐使用 psm: 3 (PSM_AUTO_OSD) 来处理多栏布局
              // 你也可以尝试其他值，例如 psm: 1 (PSM_AUTO)
              logger: (m) => {
                // 可选：添加 logger 追踪 Tesseract 进程
                if (m.status === "recognizing") {
                  // console.log(`Progress: ${m.progress}`);
                }
              },
              // 关键配置：设置页面分段模式
              // PSM_AUTO_OSD 对应的值是 3
              // PSM_AUTO 对应的值是 1
              // 更多 PSM 值可以参考 Tesseract 官方文档
              psm: 3 // 尝试 PSM_AUTO_OSD
            }
          );
          // 原有的文本处理逻辑
          const paragraphs = result.data.text.split("\n\n");
          const formattedParagraphs = paragraphs.map((paragraph) => {
            const sentences = paragraph.split(/[.|?]\s/);
            return sentences.join(" ");
          });
          setTexts((prevTexts) => [...prevTexts, ...formattedParagraphs]);
        } catch (err) {
          console.error(err);
          // Clear texts and stop conversion process immediately on error
          setTexts([]);
          setProgress(0);
          setIsLoading(false);
          return;
        } finally {
          processedImages++;
          const currentProgress = (processedImages / totalImages) * 100;
          setProgress(currentProgress);
        }
      }
    } else {
      console.error("Images is not an array.");
    }

    setIsLoading(false);
  };

  // ... (return 语句保持不变)
};

export default ImagesToText;