jecnarozvrh/scrape/parse.js

import ExcelJS from "exceljs"
import fs from "fs"
import parseAbsence from "./utils/parseAbsence.js"
import parseTeachers from "./utils/parseTeachers.js"

export default async function parseThisShit(downloadedFilePath) {
  const workbook = new ExcelJS.Workbook()
  await workbook.xlsx.readFile(downloadedFilePath)

  const sheetNames = workbook.worksheets.map((sheet) => sheet.name)

  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 = Number.parseInt(match[2], 10)
    const month = Number.parseInt(match[3], 10) - 1 // JavaScript months are 0-indexed
    const year = Number.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.getWorksheet(key)
    final.push({})

    const regex = /[AEC][0-4][a-c]?\s*\/.*/s
    const prefixRegex = /[AEC][0-4][a-c]?/
    const classes = []
    const matchingKeys = []

    currentSheet.eachRow((row, rowNumber) => {
      row.eachCell((cell, colNumber) => {
        const cellAddress = cell.address
        const value = cell.value

        if (value && typeof value === "string") {
          const testResult = regex.test(value)
          if (testResult) {
            const prefixMatch = value.match(prefixRegex)
            if (prefixMatch) {
              const prefix = prefixMatch[0]
              classes.push(prefix)
            }
            matchingKeys.push(cellAddress)
          }
        }
      })
    })

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

    // For each class
    let classI = 0
    for (const matchingKey of matchingKeys) {
      const matchingCell = currentSheet.getCell(matchingKey)
      const rowNumber = matchingCell.row
      const allKeys = []

      // Get all cells in the same row
      const row = currentSheet.getRow(rowNumber)
      row.eachCell((cell, colNumber) => {
        if (cell.address !== matchingKey) {
          allKeys.push(cell.address)
        }
      })

      let final2 = []

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

        try {
          const regex = /^úklid\s+(?:\d+\s+)?[A-Za-z]{2}$/
          const cellText = cell.text || ""
          if (regex.test(cellText.trim()) || cellText.trim().length == 0 || cell.fill?.fgColor === undefined) {
            d = false
          }
        } catch {}

        if (d) {
          let text = cell.text;
          if (cell.fill?.fgColor?.argb == "FFFFFF00") {
            text += "\n(bude upřesněno)";
          }
          final2[parsedKey] = text || ""
        } 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"] = []
    let absenceKey = null

    currentSheet.eachRow((row, rowNumber) => {
      row.eachCell((cell, colNumber) => {
        const value = (typeof cell.value === "string" ? cell.value : "").trim().toLowerCase()
        if (value === "absence") {
          absenceKey = cell.address
        }
      })
    })

    if (absenceKey) {
      const teacherMap = await parseTeachers()
      const absenceCell = currentSheet.getCell(absenceKey)
      const rowNumber = absenceCell.row
      const allAbsenceKeys = []

      // Get all cells in the same row as absence
      const row = currentSheet.getRow(rowNumber)
      row.eachCell((cell, colNumber) => {
        if (cell.address !== absenceKey) {
          allAbsenceKeys.push(cell.address)
        }
      })

      let i = 0
      for (const absenceKeyCur of allAbsenceKeys) {
        if (i >= 10) {
          break // stop once 10 items are added
        }
        i++

        const cell = currentSheet.getCell(absenceKeyCur)
        const value = (cell.value || "").toString().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 = Number.parseInt(dateMatch[1], 10)
        const month = Number.parseInt(dateMatch[2], 10)
        const year = Number.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")