jzitnik/jecnarozvrh
1
0
Fork 0
Code Actions Activity

fix: Parser once again
All checks were successful
Remote Deploy / deploy (push) Successful in 6s
Details

Browse Source 
Ti kokoti nyní místo "čtvrtek 23.10.2025 příprava" dali jen
"čt 23.10.2025 příprava", což samozřejmě zase rozbilo parser.

Nyní se v názvu stránky může objevit i zkrácená forma.
This commit is contained in:
Jakub Žitník
2025-10-21 19:29:05 +02:00
parent b600fe2669
commit 0a0b9ab436
1 changed files with 82 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files

164
scrape/parse.js
View File

@@ -4,96 +4,96 @@ import parseAbsence from "./utils/parseAbsence.js"
import parseTeachers from "./utils/parseTeachers.js" import parseTeachers from "./utils/parseTeachers.js"
export default async function parseThisShit(downloadedFilePath) { export default async function parseThisShit(downloadedFilePath) {
const workbook = new ExcelJS.Workbook() const workbook = new ExcelJS.Workbook();
await workbook.xlsx.readFile(downloadedFilePath) await workbook.xlsx.readFile(downloadedFilePath);
const sheetNames = workbook.worksheets.map((sheet) => sheet.name) 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 const dateRegex = /^(pondělí|úterý|středa|čtvrtek|pátek|po|út|ut|st|čt|ct|pa|pá)\s+(\d{1,2})\.\s*(\d{1,2})\.\s*(20\d{2})/i;
// Get today's date for comparison // Get today's date for comparison
function getCurrentDateObject() { function getCurrentDateObject() {
const now = new Date() const now = new Date();
return new Date(now.getFullYear(), now.getMonth(), now.getDate()) return new Date(now.getFullYear(), now.getMonth(), now.getDate());
} }
const today = getCurrentDateObject() const today = getCurrentDateObject();
const upcomingSheets = sheetNames.filter((name) => { const upcomingSheets = sheetNames.filter((name) => {
const match = name.match(dateRegex) const match = name.match(dateRegex);
if (!match) return false if (!match) return false;
const day = Number.parseInt(match[2], 10) const day = Number.parseInt(match[2], 10);
const month = Number.parseInt(match[3], 10) - 1 // JavaScript months are 0-indexed const month = Number.parseInt(match[3], 10) - 1; // JavaScript months are 0-indexed
const year = Number.parseInt(match[4], 10) const year = Number.parseInt(match[4], 10);
const sheetDate = new Date(year, month, day) const sheetDate = new Date(year, month, day);
return sheetDate >= today return sheetDate >= today;
}) })
const final = [] const final = [];
let finalIndex = 0 let finalIndex = 0
for (const key of upcomingSheets) { for (const key of upcomingSheets) {
const currentSheet = workbook.getWorksheet(key) const currentSheet = workbook.getWorksheet(key);
final.push({}) final.push({});
const regex = /[AEC][0-4][a-c]?\s*\/.*/s const regex = /[AEC][0-4][a-c]?\s*\/.*/s;
const prefixRegex = /[AEC][0-4][a-c]?/ const prefixRegex = /[AEC][0-4][a-c]?/;
const classes = [] const classes = [];
const matchingKeys = [] const matchingKeys = [];
currentSheet.eachRow((row, rowNumber) => { currentSheet.eachRow((row) => {
row.eachCell((cell, colNumber) => { row.eachCell((cell) => {
const cellAddress = cell.address const cellAddress = cell.address;
const value = cell.value const value = cell.value;
if (value && typeof value === "string") { if (value && typeof value === "string") {
const testResult = regex.test(value) const testResult = regex.test(value);
if (testResult) { if (testResult) {
const prefixMatch = value.match(prefixRegex) const prefixMatch = value.match(prefixRegex);
if (prefixMatch) { if (prefixMatch) {
const prefix = prefixMatch[0] const prefix = prefixMatch[0];
classes.push(prefix) classes.push(prefix);
} }
matchingKeys.push(cellAddress) matchingKeys.push(cellAddress);
} }
} }
}) })
}) })
function letterToNumber(letter) { function letterToNumber(letter) {
return letter.toLowerCase().charCodeAt(0) - "a".charCodeAt(0) return letter.toLowerCase().charCodeAt(0) - "a".charCodeAt(0);
} }
// For each class // For each class
let classI = 0 let classI = 0;
for (const matchingKey of matchingKeys) { for (const matchingKey of matchingKeys) {
const matchingCell = currentSheet.getCell(matchingKey) const matchingCell = currentSheet.getCell(matchingKey);
const rowNumber = matchingCell.row const rowNumber = matchingCell.row;
const allKeys = [] const allKeys = [];
// Get all cells in the same row // Get all cells in the same row
const row = currentSheet.getRow(rowNumber) const row = currentSheet.getRow(rowNumber);
row.eachCell((cell, colNumber) => { row.eachCell((cell) => {
if (cell.address !== matchingKey) { if (cell.address !== matchingKey) {
allKeys.push(cell.address) allKeys.push(cell.address);
} }
}) })
let final2 = [] let final2 = [];
for (const key of allKeys) { for (const key of allKeys) {
const cell = currentSheet.getCell(key) const cell = currentSheet.getCell(key);
const parsedKey = letterToNumber(key.replace(/[0-9]/gi, "")) const parsedKey = letterToNumber(key.replace(/[0-9]/gi, ""));
let d = true let d = true;
try { try {
const regex = /^úklid\s+(?:\d+\s+)?[A-Za-z]{2}$/ const regex = /^úklid\s+(?:\d+\s+)?[A-Za-z]{2}$/;
const cellText = cell.text || "" const cellText = cell.text || "";
if (regex.test(cellText.trim()) || cellText.trim().length == 0 || cell.fill?.fgColor === undefined) { if (regex.test(cellText.trim()) || cellText.trim().length == 0 || cell.fill?.fgColor === undefined) {
d = false d = false;
} }
} catch {} } catch {}
@@ -102,101 +102,101 @@ export default async function parseThisShit(downloadedFilePath) {
if (cell.fill?.fgColor?.argb == "FFFFFF00") { if (cell.fill?.fgColor?.argb == "FFFFFF00") {
text += "\n(bude upřesněno)"; text += "\n(bude upřesněno)";
} }
final2[parsedKey] = text || "" final2[parsedKey] = text || "";
} else { } else {
final2[parsedKey] = null final2[parsedKey] = null;
} }
} }
final2 = Array.from(final2, (item) => (item === undefined ? null : item)) final2 = Array.from(final2, (item) => (item === undefined ? null : item));
while (final2.length < 10) { while (final2.length < 10) {
final2.push(null) final2.push(null);
} }
final[finalIndex][classes[classI]] = final2.slice(1, 11) final[finalIndex][classes[classI]] = final2.slice(1, 11);
classI++ classI++;
} }
// ABSENCE // ABSENCE
final[finalIndex]["ABSENCE"] = [] final[finalIndex]["ABSENCE"] = [];
let absenceKey = null let absenceKey = null;
currentSheet.eachRow((row, rowNumber) => { currentSheet.eachRow((row) => {
row.eachCell((cell, colNumber) => { row.eachCell((cell) => {
const value = (typeof cell.value === "string" ? cell.value : "").trim().toLowerCase() const value = (typeof cell.value === "string" ? cell.value : "").trim().toLowerCase();
if (value === "absence") { if (value === "absence") {
absenceKey = cell.address absenceKey = cell.address;
} }
}) })
}) });
if (absenceKey) { if (absenceKey) {
const teacherMap = await parseTeachers() const teacherMap = await parseTeachers();
const absenceCell = currentSheet.getCell(absenceKey) const absenceCell = currentSheet.getCell(absenceKey);
const rowNumber = absenceCell.row const rowNumber = absenceCell.row;
const allAbsenceKeys = [] const allAbsenceKeys = [];
// Get all cells in the same row as absence // Get all cells in the same row as absence
const row = currentSheet.getRow(rowNumber) const row = currentSheet.getRow(rowNumber);
row.eachCell((cell, colNumber) => { row.eachCell((cell) => {
if (cell.address !== absenceKey) { if (cell.address !== absenceKey) {
allAbsenceKeys.push(cell.address) allAbsenceKeys.push(cell.address);
} }
}) })
let i = 0 let i = 0;
for (const absenceKeyCur of allAbsenceKeys) { for (const absenceKeyCur of allAbsenceKeys) {
if (i >= 10) { if (i >= 10) {
break // stop once 10 items are added break; // stop once 10 items are added
} }
i++ i++;
const cell = currentSheet.getCell(absenceKeyCur) const cell = currentSheet.getCell(absenceKeyCur);
const value = (cell.value || "").toString().trim() const value = (cell.value || "").toString().trim();
if (value.length === 0) { if (value.length === 0) {
continue continue
} }
const data = parseAbsence(value, teacherMap) const data = parseAbsence(value, teacherMap);
final[finalIndex]["ABSENCE"].push(data) final[finalIndex]["ABSENCE"].push(data);
} }
} }
finalIndex++ finalIndex++;
} }
const data = { const data = {
schedule: final, schedule: final,
props: upcomingSheets.map((str) => { props: upcomingSheets.map((str) => {
const dateMatch = str.match(/(\d{1,2})\.\s*(\d{1,2})\.\s*(\d{4})/) const dateMatch = str.match(/(\d{1,2})\.\s*(\d{1,2})\.\s*(\d{4})/);
let date = null let date = null;
if (dateMatch) { if (dateMatch) {
const day = Number.parseInt(dateMatch[1], 10) const day = Number.parseInt(dateMatch[1], 10);
const month = Number.parseInt(dateMatch[2], 10) const month = Number.parseInt(dateMatch[2], 10);
const year = Number.parseInt(dateMatch[3], 10) const year = Number.parseInt(dateMatch[3], 10);
date = new Date(year, month - 1, day) date = new Date(year, month - 1, day);
} }
const isPriprava = str const isPriprava = str
.toLowerCase() .toLowerCase()
.normalize("NFD") .normalize("NFD")
.replace(/[\u0300-\u036f]/g, "") .replace(/[\u0300-\u036f]/g, "")
.includes("priprava") .includes("priprava");
return { return {
date: date date: date
? `${date.getFullYear()}-${String(date.getMonth() + 1).padStart(2, "0")}-${String(date.getDate()).padStart(2, "0")}` ? `${date.getFullYear()}-${String(date.getMonth() + 1).padStart(2, "0")}-${String(date.getDate()).padStart(2, "0")}`
: null, : null,
priprava: isPriprava, priprava: isPriprava,
} };
}), }),
} }
fs.writeFileSync("db/current.json", JSON.stringify(data)) fs.writeFileSync("db/current.json", JSON.stringify(data));
} }
// parseThisShit("downloads/table.xlsx") // parseThisShit("downloads/table.xlsx")