initial commit

2025-08-07 00:17:31 +02:00
commit 47a7c73316
11 changed files with 2676 additions and 0 deletions
--- a/scrape/parse.js
+++ b/scrape/parse.js
@@ -0,0 +1,119 @@
+const XLSX = require('xlsx');
+const fs = require("fs");
+
+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;
+    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.endsWith(matchingKey.replace(/[a-z]/gi, '')));
+
+      const final2 = [];
+
+      for (const key of allKeys) {
+        const parsedKey = letterToNumber(key.replace(/[0-9]/gi, ''));
+        let d = true;
+
+        try {
+          if (currentSheet[key]["w"].startsWith("úklid")) {
+            d = false;
+          }
+        } catch {}
+
+        if (d) {
+          final2[parsedKey] = currentSheet[key]["w"];
+        } else {
+          final2[parsedKey] = null;
+        }
+      }
+
+      final[finalIndex][classes[classI]] = final2.slice(1, 11);;
+
+      classI++;
+    }
+
+    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 = /priprava/i.test(str);
+
+      return {
+        date: date ? date.toISOString().slice(0,10) : null, // ISO yyyy-mm-dd string for easy use
+        priprava: isPriprava,
+      };
+    }),
+  }
+
+  fs.writeFileSync('db/current.json', JSON.stringify(data));
+}
+
+module.exports = parseThisShit;