jecnarozvrh/scrape/parse.js

import XLSX from "xlsx";
import fs from "fs";
import parseAbsence from "./utils/parseAbsence.js";
import parseTeachers from "./utils/parseTeachers.js";

export default async function parseThisShit(downloadedFilePath) {
  const workbook = XLSX.readFile(downloadedFilePath);
  const sheetNames = workbook.SheetNames;

  const dateRegex = /^(pondělí|úterý|středa|čtvrtek|pátek)\s+(\d{1,2})\.\s*(\d{1,2})\.\s*(20\d{2})/i;

  // Get today's date for comparison
  function getCurrentDateObject() {
    const now = new Date();
    return new Date(now.getFullYear(), now.getMonth(), now.getDate());
  }

  const today = getCurrentDateObject();

  const upcomingSheets = sheetNames.filter(name => {
    const match = name.match(dateRegex);
    if (!match) return false;

    const day = parseInt(match[2], 10);
    const month = parseInt(match[3], 10) - 1; // JavaScript months are 0-indexed
    const year = parseInt(match[4], 10);

    const sheetDate = new Date(year, month, day);

    return sheetDate >= today;
  });

  const final = [];

  let finalIndex = 0;
  for (const key of upcomingSheets) {
    const currentSheet = workbook.Sheets[key];
    final.push({});

    const regex = /[AEC][0-4][a-c]?\s*\/.*/s;
    const prefixRegex = /[AEC][0-4][a-c]?/;
    const classes = [];
    const matchingKeys = Object.keys(currentSheet).filter(key => {
      const value = currentSheet[key];

      const testResult = regex.test(value.v);
      if (testResult) {
        const prefixMatch = value.v.match(prefixRegex);
        if (prefixMatch) {
          const prefix = prefixMatch[0];
          classes.push(prefix);
        }
      }

      return testResult;
    });

    function letterToNumber(letter) {
      return letter.toLowerCase().charCodeAt(0) - 'a'.charCodeAt(0);
    }

    // For each class
    let classI = 0;
    for (const matchingKey of matchingKeys) {
      const allKeys = Object.keys(currentSheet).filter(key => key !== matchingKey && key.replace(/[a-z]/gi, '') == matchingKey.replace(/[a-z]/gi, ''));
      let final2 = [];

      for (const key of allKeys) {
        const parsedKey = letterToNumber(key.replace(/[0-9]/gi, ''));
        let d = true;

        try {
          const regex = /^úklid\s+\d+\s+[A-Za-z]{2}$/;
          if (regex.test(currentSheet[key]["w"].trim()) || currentSheet[key]["w"].trim().length == 0) {
            d = false;
          }
        } catch {}

        if (d) {
          final2[parsedKey] = currentSheet[key]["w"];
        } else {
          final2[parsedKey] = null;
        }
      }

      final2 = Array.from(final2, item => item === undefined ? null : item);
      while (final2.length < 10) {
        final2.push(null);
      }

      final[finalIndex][classes[classI]] = final2.slice(1, 11);;

      classI++;
    }

    // ABSENCE
    final[finalIndex]["ABSENCE"] = [];
    const absenceKey = Object.keys(currentSheet).find(key => {
      const value = (typeof currentSheet[key].v == "string" ? currentSheet[key].v : "").trim().toLowerCase();
      return value == "absence";
    });
    const teacherMap = await parseTeachers();
    const allAbsenceKeys = Object.keys(currentSheet).filter(key => key !== absenceKey && key.replace(/[a-z]/gi, '') == absenceKey.replace(/[a-z]/gi, ''));
    let i = 0;
    for (const absenceKeyCur of allAbsenceKeys) {
      if (i >= 10) {
        break; // stop once 10 items are added
      }
      i++;

      const value = currentSheet[absenceKeyCur]["v"].trim();
      if (value.length === 0) {
        continue;
      }

      const data = parseAbsence(value, teacherMap);
      final[finalIndex]["ABSENCE"].push(data);
    }

    finalIndex++;
  }

  const data = {
    schedule: final,
    props: upcomingSheets.map(str => {
      const dateMatch = str.match(/(\d{1,2})\.\s*(\d{1,2})\.\s*(\d{4})/);

      let date = null;

      if (dateMatch) {
        const day = parseInt(dateMatch[1], 10);
        const month = parseInt(dateMatch[2], 10);
        const year = parseInt(dateMatch[3], 10);

        date = new Date(year, month - 1, day);
      }

      const isPriprava = str.toLowerCase().normalize("NFD").replace(/[\u0300-\u036f]/g, "").includes("priprava");

      return {
        date: date
          ? `${date.getFullYear()}-${String(date.getMonth() + 1).padStart(2, '0')}-${String(date.getDate()).padStart(2, '0')}`
          : null,
        priprava: isPriprava,
      };
    }),
  }

  fs.writeFileSync('db/current.json', JSON.stringify(data));
}

parseThisShit("downloads/table.xlsx");