feat(parser): update group parsing to handle special cases and improve regex patterns

feat(parser): enhance details extraction for lessons, including handling of subgroups and room moves

feat(timetable): integrate room and teacher retrieval based on week context

test(parser): add tests for group parsing with special characters and room move scenarios

test(formats): implement tests for various lesson formats and their expected outputs

feat(rooms): introduce room management functions for handling moves and week-based plans

server/parser.js

@@ -7,8 +7,9 @@ import { tmpdir } from "node:os";
 import { join } from "node:path";
 import { createHash } from "node:crypto";
 import { parseWeekRanges, parseTextDates, DATE_LIST, weekday } from "../shared/weeks.js";
+import { roomMoves, onlyPlace, weekPlan, partsByWeek } from "./rooms.js";
 const run = promisify(execFile);
-export const PARSER_VERSION = 12;
+export const PARSER_VERSION = 13;
 const days = [
   "понедельник",
   "вторник",
@@ -24,19 +25,22 @@ const clean = (value) =>
     .replace(/[\t ]+/g, " ")
     .replace(/\s*\n\s*/g, "\n")
     .trim();
-// A group is a two-digit year and a capitalised abbreviation: "23-БА", "24-ДОи", "25-ИЯ(РА)",
-// "23-ИНС-1". Header cells also carry practice dates ("02.11-22.11-производственная практика",
-// "09.11 по 06.12"): a number right after a date's dot or another digit is never a group.
+// A group is a two-digit year and a capitalised abbreviation: "23-БА", "24-ДОи", "23-ИНС-1",
+// "24-ИЯ (РА)и-1" (a specialisation in brackets, then the group's letter and number). Header
+// cells also carry practice dates ("02.11-22.11-производственная практика", "09.11 по 06.12"):
+// a number right after a date's dot or another digit is never a group. Written as the university
+// writes it: "24-ИЯ (РА)и-1", with the space before the bracket.
 export const groupsIn = (value) =>
   [
     ...value.matchAll(
-      /(?<![\d.,:/])\b\d{2}\s*[-–]?\s*[А-ЯЁ][А-ЯЁа-яё]{1,7}(?:\s*\([А-ЯЁа-яё]+\))?(?:\s*[-–]\s*\d)?/gu,
+      /(?<![\d.,:/])\b\d{2}\s*[-–]?\s*[А-ЯЁ][А-ЯЁа-яё]{1,7}(?:\s*\([А-ЯЁа-яё]+\)[а-яё]{0,3})?(?:\s*[-–]\s*\d)?/gu,
     ),
   ].map((m) =>
     m[0]
       .replace(/\s+/g, "")
       .replace(/–/g, "-")
-      .replace(/^(\d{2})([А-ЯЁа-яё])/, "$1-$2"),
+      .replace(/^(\d{2})([А-ЯЁа-яё])/, "$1-$2")
+      .replace(/(\S)\(/, "$1 ("),
   );
 function timesIn(value) {
   const times = [...value.matchAll(/\b([0-2]?\d)[.:]([0-5]\d)\b/g)].map(
@@ -73,15 +77,20 @@ export function details(raw) {
   // Correspondence sessions: "Экзамен по учебной дисциплине «Базы данных» Скуковская А.А. ауд.219Д".
   const named = rest.match(/^по\s+(?:учебной\s+)?дисциплине\s*«([^»]+)»\s*/i);
   if (named) rest = named[1].trim() + " " + rest.slice(named[0].length);
-  const kind = rest.match(
-    /\((уст\.?\s*л\.?|уст\.?\s*п\.?\s*з\.?|л\.?\s*з\.?|п\.?\s*з\.?|л\.?|лаб\.?|пр\.?|лек\.?|сем\.?|экз\.?|зач\.?|конс\.?)\)/i,
-  );
+  // "Дисциплина по выбору: Социология": the elective's label is no part of its name.
+  rest = rest.replace(/^дисциплин[аы]\s+(?:по\s+выбору|с[пц]е[цш]иализаци[ийя])\s*:\s*/iu, "");
+  // "12н.2п/гр. Экономика…": the lesson of one subgroup, named in the cell.
+  const own = rest.match(/^(\d)\s*п\.?\s*\/?\s*гр?\.?\s*/i);
+  if (own) rest = rest.slice(own[0].length);
+  const kind = rest.match(KIND);
   // Whole words only: "консультирования" in a subject name is not a consultation.
   const sessionWord = (leading?.[1] ||
     rest.match(/(?<![а-яё])(экзамен|зач[её]т|консультация)(?![а-яё])/i)?.[1] ||
     "").toLowerCase();
   const person = rest.match(/[А-ЯЁ][а-яё-]+\s+[А-ЯЁ]\.\s*[А-ЯЁ]\./u);
   const room =
+    // "ауд. А1-06", "ауд. С0-03": a building letter and a number.
+    rest.match(/(?:ауд\.?|(?<![А-Яа-яЁё])а\.)\s*([А-ЯA-Z]\d-\d{2})/u) ||
     rest.match(/(?:ауд\.?\s*)([\d]+\s*[А-Яа-я]?(?:\s*\([^)]*\))?)/i) ||
     rest.match(/\b(\d{2,3}\s*[А-Яа-я]?)\s*$/u) ||
     // A room with a name instead of a number: "ауд.Медиа".
@@ -94,9 +103,13 @@ export function details(raw) {
   // A typo'd date before the subject ("19,11, 03.12 Физика") leaves a stray comma behind.
   subject = subject.replace(/^[\s,;.]+/, "");
   if (!subject) subject = rest;
+  // A long note under a capitalised heading ("ВВЕДЕНИЕ В ВЫСШЕЕ ОБРАЗОВАНИЕ / г. Полоцк, …"):
+  // the heading names it; the whole text stays in the details.
+  const heading = rest.split("\n")[0].trim();
+  if (subject === rest && rest.split("\n").length >= 3 && /[А-ЯЁ]{4}/u.test(heading) && !/[а-яё]/u.test(heading)) subject = heading;
   const defence = /^защита/i.test(leading?.[1] || "");
   const token =
-    kind?.[1].replace(/[.\s]/g, "").toLowerCase() ||
+    kind?.[1].replace(/[.\s/]/g, "").toLowerCase() ||
     (defence ? "защ" : undefined) ||
     (sessionWord.startsWith("экз")
       ? "экз"
@@ -110,7 +123,9 @@ export function details(raw) {
     teacher: person?.[0] || "",
     room: room?.[1].trim().replace(/\s+/g, "") || "",
     type:
-      token === "устл"
+      token === "лпз" || token === "липз"
+        ? "Лекция и практика"
+        : token === "устл"
         ? "Установочная лекция"
         : token === "устпз"
           ? "Установочное занятие"
@@ -132,11 +147,97 @@ export function details(raw) {
                     ? "Консультация"
                     : "Занятие",
     condition,
+    ...(own && { ownSubgroup: Number(own[1]) }),
   };
 }
-// Split only at an explicit new week condition. Keep unfamiliar wording intact.
-function alternatives(raw) {
-  return raw.split(new RegExp(String.raw`\n(?=${WEEK_MARK.source.slice(1)})`, "i")).filter(Boolean);
+// The kind of a lesson in brackets: "(л.)", "(п.з.)", "(уст.л.)", "(конс.)".
+const KIND =
+  /\((л\.?\s*(?:\/|и)\s*п\.?\s*з\.?|уст\.?\s*л\.?|уст\.?\s*п\.?\s*з\.?|л\.?\s*з\.?|п\.?\s*з\.?|л\.?|лаб\.?|пр\.?|лек\.?|сем\.?|экз\.?|зач\.?|конс\.?)\)/i;
+// Split only at an explicit new week condition. Keep unfamiliar wording intact. A line of an
+// elective's choice may put its weeks after the subject ("Социология 1-10 н. (л.) …"): it is
+// a lesson of its own, read with the weeks first.
+const SUBJECT_THEN_WEEKS = /^([А-ЯЁ][а-яё]+(?:[ -][а-яё]+){0,4})\s+((?:с\s*)?\d{1,2}(?:\s*[-–,]\s*\d{1,2})*\s*н(?:ед)?\.?)\s*(?=\()/u;
+// "Дисциплина специализации: 4-14 н. Преступления…": the heading hides the weeks.
+const HEADING = /^[^\n:]{3,40}:\s*(?=(?:с\s*)?\d{1,2}(?:\s*[-–,]\s*\d{1,2})*\s*н)/u;
+// "… (л.) Шидловский С.О. 11-16 н. (п.з.) Лапацуева А.С.": after a teacher, other weeks of
+// another kind begin on the same line.
+const NEXT_KIND = /(?<=[А-ЯЁ][а-яё-]+\s+[А-ЯЁ]\.\s*[А-ЯЁ]\.)\s+(?=(?:с\s*)?\d{1,2}(?:\s*[-–,]\s*\d{1,2})*\s*н(?:ед)?\.?\s*\()/u;
+export function alternatives(raw) {
+  return raw
+    // Lookaheads without capturing groups: split() would return what they capture.
+    .split(new RegExp(String.raw`\n(?=${WEEK_MARK.source.slice(1)}|${SUBJECT_THEN_WEEKS.source.slice(1).replace(/(?<!\\)\((?!\?)/g, "(?:")})`, "iu"))
+    .filter(Boolean)
+    .map((part) => part.replace(HEADING, "").replace(SUBJECT_THEN_WEEKS, "$2 $1 "))
+    .flatMap((part) => part.split(NEXT_KIND))
+    .reduce(
+      (state, part) => {
+        // "Дисциплины по выбору:" heads the lessons below; it is none itself. A heading with
+        // times ("18.35+20.10 Новополоцк:") gives those times to the lessons under it.
+        if (/^[^\n]{1,60}:\s*$/.test(part.trim())) state.times = part.match(HEADING_TIMES)?.[0] || state.times;
+        else state.parts.push(state.times ? `${state.times} ${part}` : part);
+        return state;
+      },
+      { parts: [], times: "" },
+    ).parts;
+}
+const HEADING_TIMES = /^[0-2]?\d[.:][0-5]\d(?:\s*(?:\+|и|,)\s*[0-2]?\d[.:][0-5]\d)*/u;
+// "1-11н. - л., 12-14 – п.з. Защита населения… Дмитриченко Н.С.": one kind for some weeks,
+// another for the others.
+const WEEKS_TEXT = String.raw`(?:с\s*)?\d{1,2}(?:\s*[-–,]\s*\d{1,2})*\s*(?:н(?:ед)?\.?)?`;
+const KIND_TEXT = String.raw`л|п\.?\s*з|л\.?\s*з|лаб|пр`;
+const BY_KIND = new RegExp(
+  String.raw`^(${WEEKS_TEXT})\s*[-–—]\s*(${KIND_TEXT})\.?\s*,\s*(${WEEKS_TEXT})\s*[-–—]\s*(${KIND_TEXT})\.?\s+([\s\S]+)$`,
+  "iu",
+);
+function byKind(text) {
+  const m = text.match(BY_KIND);
+  if (!m) return [text];
+  const weeks = (w) => (/н/i.test(w) ? w.trim() : `${w.trim()}н.`);
+  // The kind goes right after the subject, before the teacher and the room.
+  const rest = m[5];
+  const cut = [rest.search(/[А-ЯЁ][а-яё-]+\s+[А-ЯЁ]\.\s*[А-ЯЁ]\./u), rest.search(/(?:ауд\.?|(?<![А-Яа-яЁё])а\.)\s*\S/u)]
+    .filter((i) => i > 0)
+    .reduce((a, b) => Math.min(a, b), rest.length);
+  const at = (kind) => `${rest.slice(0, cut).trim()} (${kind}.) ${rest.slice(cut)}`.trim();
+  return [`${weeks(m[1])} ${at(m[2])}`, `${weeks(m[3])} ${at(m[4])}`];
+}
+
+/**
+ * A cell's entries, each `{text, main, plan}`: the whole text, its first line (subject, kind,
+ * teacher and room are read from it) and the plan. A line that names only rooms or teachers by week
+ * ("1-9н. 502н", "с 10н. – 359", "2-9н.-Клос О.Ю., 10-14н.-Конторов В.Э.") belongs to the lesson
+ * above: its plan of rooms and teachers by week. A line with weeks and a kind but no subject
+ * ("7,9,11н. – (п.з.) Савицкая К.Д. ауд. А1-06 8,12,14н. – (п.з.) …") is the subject above in
+ * other weeks. A stray "." is no lesson.
+ */
+export function entries(raw) {
+  const out = [];
+  for (const alternative of alternatives(raw).flatMap(byKind)) {
+    if (!/[\p{L}\d]/u.test(alternative)) continue;
+    const prev = out.at(-1);
+    if (prev && onlyPlace(alternative)) {
+      if (KIND.test(alternative)) {
+        const head = details(prev.main);
+        // "Дисциплина по выбору: Социология ауд. С0-03" over lines by kind: a heading, not a
+        // lesson; its room goes to the lessons under it.
+        if (!KIND.test(prev.main) && !head.teacher && !prev.plan.length) prev.heading = true;
+        for (const part of partsByWeek(alternative)) {
+          const mark = part.match(WEEK_MARK)?.[0] || "";
+          const rest = part.slice(mark.length);
+          const room = prev.heading && head.room && !/\d{2,}/.test(rest) ? ` ауд. ${head.room}` : "";
+          const text = `${mark} ${head.subject} ${rest}${room}`.replace(/[ \t]+/g, " ").trim();
+          out.push({ text, main: text, plan: [] });
+        }
+      } else {
+        prev.text += "\n" + alternative;
+        prev.plan.push(...weekPlan(alternative));
+      }
+      continue;
+    }
+    // "бел.н.504н/зел.н.524н": a room for white weeks and one for green ones.
+    out.push({ text: alternative, main: alternative, plan: weekPlan(alternative).filter((p) => p.color) });
+  }
+  return out.filter((e) => !e.heading);
 }
 export function readDocx(buffer) {
   let zip;
@@ -250,7 +351,7 @@ export function leadingDate(text, yearStart) {
  */
 export function leadingTime(text, loose = false) {
   const hm = (h, m) => `${h.padStart(2, "0")}:${m}`;
-  const range = text.match(/^([0-2]?\d)[.:]([0-5]\d)\s*[-–]\s*([0-2]?\d)[.:]([0-5]\d)(?=\s)/);
+  const range = text.match(/^([0-2]?\d)[.:]([0-5]\d)\s*[-–]\s*([0-2]?\d)[.:]([0-5]\d)(?=\s|[-–—])/);
   if (range && Number(range[1]) <= 23 && Number(range[3]) <= 23)
     return { starts: [hm(range[1], range[2])], end: hm(range[3], range[4]), length: range[0].length };
   const T = String.raw`[0-2]?\d[.:][0-5]\d`;
@@ -447,10 +548,11 @@ export function parseDocx(buffer, source = {}) {
         const spansBoth =
           cell.merge === "restart" &&
           rows[i + 1]?.find((c) => c.col === cell.col)?.merge === "continue";
-        for (const alternative of alternatives(raw)) {
-          // "Единый день информирования 10.09;15.10;19.11;17.12": the dates are not part of the subject.
+        for (const { text: whole, main: alternative, plan } of entries(raw)) {
+          // "Единый день информирования 10.09;15.10;19.11;17.12": the dates are not part of the
+          // subject, nor is a move to another room on a date ("412н 26.10.2026-333").
           const read = (text) => {
-            const data = details(text.replace(DATE_LIST, "").replace(/[ \t]{2,}/g, " ").trim());
+            const data = details(roomMoves(text).rest.replace(DATE_LIST, "").replace(/[ \t]{2,}/g, " ").trim());
             if (credits && data.type === "Занятие") data.type = "Зачёт";
             return data;
           };
@@ -483,6 +585,10 @@ export function parseDocx(buffer, source = {}) {
                   group: r.group,
                   label: `Подгруппа ${cell.col - r.col + 1}`,
                 }));
+          // A subgroup named in the cell ("2п/гр.") when the column does not tell one.
+          const own = data.ownSubgroup;
+          delete data.ownSubgroup;
+          if (own && !subgroups.length) subgroups.push(...targetGroups.map((group) => ({ group, label: `Подгруппа ${own}` })));
           // "17.00 и 18.35 Русский язык": one lesson at each time.
           for (const [n, start] of (lead?.starts || [null]).entries())
           lessons.push({
@@ -511,7 +617,9 @@ export function parseDocx(buffer, source = {}) {
               !["auto", "FFFFFF"].includes(cell.color.toUpperCase())
                 ? cell.color
                 : null,
-            raw: alternative,
+            raw: whole,
+            // Rooms and teachers by week, or by the week's colour.
+            ...(plan.length && { plan }),
           });
         }
       };