refactor: Rewrite to typescript

scrape/utils/parseAbsence.ts

@@ -0,0 +1,247 @@
+/*
+ * Copyright (C) 2025 Jakub Žitník
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+const LAST_HOUR = 10;
+
+// -------------------------------
+// Helpers
+// -------------------------------
+export const cleanInput = (input: string | null | undefined): string => (input ?? "").trim().replace(/\s+/g, " ");
+export const isTeacherToken = (t: string): boolean => /^[A-Za-z]+$/.test(t);
+
+interface Spec {
+  kind: "range" | "single";
+  value: { from: number; to: number } | number;
+}
+
+interface TeacherMap {
+  [key: string]: string;
+}
+
+export interface AbsenceResult {
+  teacher: string | null;
+  teacherCode: string | null;
+  type: string;
+  hours: { from: number; to: number } | number | null;
+  zastupuje?: { teacher: string | null; teacherCode: string | null };
+  original?: string;
+}
+
+export const parseSpec = (spec: string | null): Spec | null => {
+  if (!spec) return null;
+  let m;
+
+  // Handle "6,7"
+  if ((m = spec.match(/^(\d+),(\d+)$/))) {
+    const from = Number(m[1]);
+    const to = Number(m[2]);
+    return from >= 1 && to >= from && to <= LAST_HOUR
+      ? { kind: "range", value: { from, to } }
+      : null;
+  }
+
+  if ((m = spec.match(/^(\d+)\+$/))) {
+    const from = Number(m[1]);
+    return from >= 1 && from <= LAST_HOUR
+      ? { kind: "range", value: { from, to: LAST_HOUR } }
+      : null;
+  }
+  if ((m = spec.match(/^(\d+)-(\d+)$/))) {
+    const from = Number(m[1]);
+    const to = Number(m[2]);
+    return from >= 1 && to >= from && to <= LAST_HOUR
+      ? { kind: "range", value: { from, to } }
+      : null;
+  }
+  if ((m = spec.match(/^(\d+)$/))) {
+    const hour = Number(m[1]);
+    return hour >= 1 && hour <= LAST_HOUR
+      ? { kind: "single", value: hour }
+      : null;
+  }
+  return null;
+};
+
+export const resolveTeacher = (teacherCode: string, teacherMap: TeacherMap = {}): { code: string; name: string | null } => ({
+  code: teacherCode,
+  name: teacherMap?.[teacherCode.toLowerCase()] ?? null,
+});
+
+const makeResult = (teacherCode: string, spec: Spec | null, teacherMap: TeacherMap): AbsenceResult => {
+  const { name } = resolveTeacher(teacherCode, teacherMap);
+  const type = spec ? (spec.kind === "range" ? "range" : "single") : "wholeDay";
+  const hours = spec ? spec.value : null;
+  return { teacher: name, teacherCode: teacherCode.toLowerCase(), type, hours };
+};
+
+// -------------------------------
+// Teacher list processing (modular)
+// -------------------------------
+const processTeacherList = (teacherListStr: string, spec: Spec | null, teacherMap: TeacherMap): AbsenceResult[] => {
+  let results: AbsenceResult[] = [];
+  const teachers = teacherListStr.split(/[,;]\s*/).filter(Boolean);
+
+  if (teacherListStr.includes(";")) {
+    teachers.forEach((t, i) => {
+      const resSpec = i === teachers.length - 1 ? spec : null;
+      results.push(makeResult(t, resSpec, teacherMap));
+    });
+  } else {
+    results = teachers.map((t) => makeResult(t, spec, teacherMap));
+  }
+
+  return results;
+};
+
+// -------------------------------
+// Main parser
+// -------------------------------
+export default function parseAbsence(input: string, teacherMap: TeacherMap = {}): AbsenceResult[] {
+  const s = cleanInput(input);
+  if (!s) return [];
+
+  const results: AbsenceResult[] = [];
+  const consumed: [number, number][] = [];
+  const markConsumed = (start: number, end: number) => consumed.push([start, end]);
+  const isConsumed = (i: number) => consumed.some(([a, b]) => i >= a && i < b);
+
+  // 1. Teachers with specific hours (e.g. "Ab 1-4")
+  const teacherListThenSpecRe =
+    /([A-Za-z]+(?:[,;]\s?[A-Za-z]+)*)(?:\s*)(\d+(?:\+|-\d+|,\d+)?)(?:\.\s*h)?(?![A-Za-z])/g;
+
+  let m;
+  while ((m = teacherListThenSpecRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    const matchEnd = teacherListThenSpecRe.lastIndex;
+    if (isConsumed(matchStart)) continue;
+
+    const [_, teacherListStr, specStr] = m;
+    const spec = parseSpec(specStr);
+    if (!spec) continue;
+
+    results.push(...processTeacherList(teacherListStr, spec, teacherMap));
+    markConsumed(matchStart, matchEnd);
+  }
+
+  // 1a. Teachers with "-startHour-exk" suffix (e.g. "Sv-5-exk")
+  const teacherStartExkRe = /([A-Za-z]+)-(\d+)-exk/gi;
+  while ((m = teacherStartExkRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    const matchEnd = teacherStartExkRe.lastIndex;
+    if (isConsumed(matchStart)) continue;
+
+    const teacherCode = m[1];
+    const from = Number(m[2]);
+    const { name } = resolveTeacher(teacherCode, teacherMap);
+
+    if (from >= 1 && from <= LAST_HOUR) {
+      results.push({
+        teacher: name,
+        teacherCode: teacherCode.toLowerCase(),
+        type: "exkurze",
+        hours: { from, to: LAST_HOUR },
+      });
+      markConsumed(matchStart, matchEnd);
+    }
+  }
+
+  // 2. Teachers with "-exk" suffix
+  const teacherExkRe = /([A-Za-z]+)-exk/gi;
+  while ((m = teacherExkRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    const matchEnd = teacherExkRe.lastIndex;
+    if (isConsumed(matchStart)) continue;
+
+    const teacherCode = m[1];
+    const { name } = resolveTeacher(teacherCode, teacherMap);
+    results.push({
+      teacher: name,
+      teacherCode: teacherCode.toLowerCase(),
+      type: "exkurze",
+      hours: null,
+    });
+    markConsumed(matchStart, matchEnd);
+  }
+
+  // ---------------------------------------------------------
+  // 3. Substitution Pattern: "za Vn zastupuje Jk"
+  // ---------------------------------------------------------
+  const substitutionRe = /za\s+([A-Za-z]+)\s+zastupuje\s+([A-Za-z]+)/gi;
+  while ((m = substitutionRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    const matchEnd = substitutionRe.lastIndex;
+    if (isConsumed(matchStart)) continue;
+
+    const missingCode = m[1];
+    const substituteCode = m[2];
+    
+    const missingResolved = resolveTeacher(missingCode, teacherMap);
+    const subResolved = resolveTeacher(substituteCode, teacherMap);
+
+    results.push({
+      teacher: missingResolved.name,
+      teacherCode: missingResolved.code.toLowerCase(),
+      type: "zastoupen",
+      hours: null,
+      zastupuje: {
+        teacher: subResolved.name,
+        teacherCode: subResolved.code.toLowerCase()
+      }
+    });
+
+    markConsumed(matchStart, matchEnd);
+  }
+
+  const teacherOnlyRe = /([A-Za-z]+(?:[,;]\s?[A-Za-z]+)*)/g;
+  while ((m = teacherOnlyRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    const matchEnd = teacherOnlyRe.lastIndex;
+    if (isConsumed(matchStart)) continue;
+
+    const tList = m[1].split(/[,;]\s*/).filter(Boolean);
+    tList.forEach((t) => {
+      const lowerT = t.toLowerCase();
+      if (lowerT === 'za' || lowerT === 'zastupuje') return;
+
+      if (isTeacherToken(t)) results.push(makeResult(t, null, teacherMap));
+      else
+        results.push({
+          type: "invalid",
+          teacher: null,
+          teacherCode: null,
+          hours: null,
+          original: t,
+        });
+    });
+    markConsumed(matchStart, matchEnd);
+  }
+
+  // 5. Bare specs without teacher → invalid
+  const specOnlyRe = /\b(\d+(?:\+|-\d+|,\d+)?)\b/g;
+  while ((m = specOnlyRe.exec(s)) !== null) {
+    const matchStart = m.index;
+    if (isConsumed(matchStart)) continue;
+    if (m[1].trim() == "exk") continue; 
+
+    results.push({
+      type: "invalid",
+      teacher: null,
+      teacherCode: null,
+      hours: null,
+      original: m[1],
+    });
+  }
+
+  return results;
+}