pdf-reader-web/helpers/pdf.ts

import pdfjs from 'pdfjs-dist';
import { delay } from './time';

pdfjs.GlobalWorkerOptions.workerSrc = '/static/build/pdf.worker.min.js';
pdfjs.cMapUrl = '/static/cmaps/';
pdfjs.cMapPacked = true;

let normalizationRegex: any = null;
const CHARACTERS_TO_NORMALIZE: { [index: string]: any } = {
  '\u2018': "'", // Left single quotation mark
  '\u2019': "'", // Right single quotation mark
  '\u201A': "'", // Single low-9 quotation mark
  '\u201B': "'", // Single high-reversed-9 quotation mark
  '\u201C': '"', // Left double quotation mark
  '\u201D': '"', // Right double quotation mark
  '\u201E': '"', // Double low-9 quotation mark
  '\u201F': '"', // Double high-reversed-9 quotation mark
  '\u00BC': '1/4', // Vulgar fraction one quarter
  '\u00BD': '1/2', // Vulgar fraction one half
  '\u00BE': '3/4', // Vulgar fraction three quarters
};

export const fetchPdf = async (
  src: string,
  cb?: (progress: ProgressType) => void
): Promise<any> => {
  try {
    const loadingTask = pdfjs.getDocument({
      url: src,
      cMapUrl: '/static/cmaps/',
      cMapPacked: true,
    });

    if (cb) {
      loadingTask.onProgress = (progress: ProgressType): void => {
        cb(progress);
      };
    }

    const pdf = await loadingTask.promise;
    return pdf;
  } catch (e) {
    console.log(e);
  }

  return {};
};

export const renderTextLayer = async ({
  pdfPage,
  textLayer,
  viewport,
  setTextDivs,
}: {
  pdfPage: any;
  textLayer: HTMLElement;
  viewport: ViewportType;
  setTextDivs?: (elements: HTMLElement[]) => void;
}): Promise<any> => {
  const textContent = await pdfPage.getTextContent();
  const textDivs: any[] = [];
  await pdfjs.renderTextLayer({
    textContent,
    container: textLayer,
    viewport,
    textDivs,
  });
  if (setTextDivs) {
    setTextDivs(textDivs);
  }
};

export const renderPdfPage = async ({
  rootEle,
  pdfPage,
  viewport,
  setRenderTask,
  setTextDivs,
}: {
  rootEle: HTMLElement;
  pdfPage: any;
  viewport: ViewportType;
  setRenderTask: any;
  setTextDivs: (elements: HTMLElement[]) => void;
}): Promise<any> => {
  if (rootEle) {
    const canvas: HTMLCanvasElement = rootEle.querySelectorAll(
      'canvas'
    )[0] as HTMLCanvasElement;
    const textLayer: HTMLDivElement = rootEle.querySelector(
      '[data-id="text-layer"]'
    ) as HTMLDivElement;

    if (canvas) {
      const context: CanvasRenderingContext2D = canvas.getContext(
        '2d'
      ) as CanvasRenderingContext2D;
      canvas.height = viewport.height;
      canvas.width = viewport.width;

      const renderContext = {
        canvasContext: context,
        viewport,
      };

      if (pdfPage) {
        const renderTask = pdfPage.render(renderContext);

        setRenderTask(renderTask);
        await renderTask.promise.catch((reason: string) => {
          console.log(`stopped ${reason}`);
        });
      }

      textLayer.innerHTML = '';
      await delay(200);

      await renderTextLayer({
        pdfPage,
        textLayer,
        viewport,
        setTextDivs,
      });
    }
  }
};

export const normalize = (text: string): string => {
  if (!normalizationRegex) {
    // Compile the regular expression for text normalization once.
    const replace = Object.keys(CHARACTERS_TO_NORMALIZE).join('');
    normalizationRegex = new RegExp(`[${replace}]`, 'g');
  }
  return text.replace(normalizationRegex, ch => CHARACTERS_TO_NORMALIZE[ch]);
};

export const calculatePhraseMatch = (
  pageContent: string,
  query: string
): number[] => {
  const matches = [];
  const queryLen = query.length;
  let matchIdx = -queryLen;

  if (pageContent) {
    while (query) {
      matchIdx = pageContent.indexOf(query, matchIdx + queryLen);
      if (matchIdx === -1) break;
      matches.push(matchIdx);
    }
  }

  return matches;
};

export const convertMatches = (
  queryString: string,
  matchIndex: number,
  textContentItem: any[]
): Record<string, any> => {
  let i = 0;
  let iIndex = 0;
  const end = textContentItem.length - 1;
  const queryLen = queryString.length;

  // Loop over the divIdxs.
  while (i !== end && matchIndex >= iIndex + textContentItem[i].length) {
    iIndex += textContentItem[i].length;
    i += 1;
  }

  if (i === textContentItem.length) {
    console.error('Could not find a matching mapping');
  }

  const match: Record<string, any> = {
    begin: {
      divIdx: i,
      offset: matchIndex - iIndex,
    },
  };

  // Calculate the end position.
  // eslint-disable-next-line no-param-reassign
  matchIndex += queryLen;

  // Somewhat the same array as above, but use > instead of >= to get
  // the end position right.
  while (i !== end && matchIndex > iIndex + textContentItem[i].length) {
    iIndex += textContentItem[i].length;
    i += 1;
  }

  match.end = {
    divIdx: i,
    offset: matchIndex - iIndex,
  };

  return match;
};

export const getPdfPage = async (pdf: any, pageNum: number): Promise<any> => {
  const page = await pdf.getPage(pageNum);
  return page;
};