안의 | 들을 추출
+ const trRegex = /]*>([\s\S]*?)<\/tr>/gi;
+ let trMatch: RegExpExecArray | null;
+
+ while ((trMatch = trRegex.exec(html)) !== null) {
+ const trContent = trMatch[1];
+ const tds: string[] = [];
+ const tdRegex = /]*>([\s\S]*?)<\/td>/gi;
+ let tdMatch: RegExpExecArray | null;
+ while ((tdMatch = tdRegex.exec(trContent)) !== null) {
+ tds.push(stripTags(decodeHtmlEntities(tdMatch[1])));
+ }
+
+ // 팀 순위 행은 12개 컬럼: 순위, 팀명, 경기, 승, 패, 무, 승률, 게임차, 최근10경기, 연속, 홈, 방문
+ if (tds.length >= 10) {
+ // 팀명이 한글/영문인지 간단 체크
+ const teamName = tds[1];
+ if (/^[A-Z가-힣]/.test(teamName)) {
+ teams.push({
+ 순위: tds[0],
+ 팀명: tds[1],
+ 경기: tds[2],
+ 승: tds[3],
+ 패: tds[4],
+ 무: tds[5],
+ 승률: tds[6],
+ 게임차: tds[7],
+ 최근10경기: tds[8],
+ 연속: tds[9],
+ 홈: tds[10] ?? "",
+ 방문: tds[11] ?? "",
+ });
+ }
+ }
+ }
+
+ return teams;
+}
+
+function parseWLD(s: string): WinLossDraw {
+ const parts = s.split("-").map(Number);
+ return [parts[0] ?? 0, parts[1] ?? 0, parts[2] ?? 0];
+}
+
+function parseVsTable(html: string): TeamVsRecord[] {
+ const tableMatch = html.match(
+ /]*summary="팀간승패표"[^>]*>([\s\S]*?)<\/table>/
+ );
+ if (!tableMatch) return [];
+
+ const tableHtml = tableMatch[1];
+
+ // 헤더에서 팀명 목록 추출
+ const theadMatch = tableHtml.match(/]*>([\s\S]*?)<\/thead>/);
+ const teamNames: string[] = [];
+ if (theadMatch) {
+ const thRegex = /]*>([\s\S]*?)<\/th>/gi;
+ let thMatch: RegExpExecArray | null;
+ while ((thMatch = thRegex.exec(theadMatch[1])) !== null) {
+ const text = stripTags(decodeHtmlEntities(thMatch[1])).replace(/\(승-패-무\)/g, "").trim();
+ if (text && text !== "팀명" && text !== "합계") {
+ teamNames.push(text);
+ }
+ }
+ }
+
+ // 바디에서 각 팀의 상대전적 추출
+ const records: TeamVsRecord[] = [];
+ const tbodyMatch = tableHtml.match(/]*>([\s\S]*?)<\/tbody>/);
+ if (!tbodyMatch) return [];
+
+ const trRegex = /]*>([\s\S]*?)<\/tr>/gi;
+ let trMatch: RegExpExecArray | null;
+
+ while ((trMatch = trRegex.exec(tbodyMatch[1])) !== null) {
+ const cells: string[] = [];
+ const tdRegex = /]*>([\s\S]*?)<\/td>/gi;
+ let tdMatch: RegExpExecArray | null;
+ while ((tdMatch = tdRegex.exec(trMatch[1])) !== null) {
+ cells.push(stripTags(decodeHtmlEntities(tdMatch[1])));
+ }
+
+ if (cells.length < teamNames.length + 2) continue;
+
+ const 팀명 = cells[0];
+ const 상대전적: Record = {};
+
+ for (let i = 0; i < teamNames.length; i++) {
+ const val = cells[i + 1];
+ if (val === "■" || val === "▲" || 팀명 === teamNames[i]) continue;
+ 상대전적[teamNames[i]] = parseWLD(val);
+ }
+
+ const 합계 = parseWLD(cells[cells.length - 1]);
+ records.push({ 팀명, 상대전적, 합계 });
+ }
+
+ return records;
+}
+
+// ── 네트워크 ──
+
+async function getInitialPage(): Promise<{
+ html: string;
+ viewState: string;
+ viewStateGenerator: string;
+ eventValidation: string;
+ currentYear: number;
+}> {
+ const res = await fetch(BASE_URL, {
+ headers: {
+ "User-Agent":
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/147.0.0.0 Safari/537.36",
+ },
+ });
+ const html = await res.text();
+
+ // 현재 선택된 연도 추출
+ const yearMatch = html.match(
+ /id="cphContents_cphContents_cphContents_lblSearchDateTitle">(\d{4})/
+ );
+ const currentYear = yearMatch ? parseInt(yearMatch[1], 10) : new Date().getFullYear();
+
+ return {
+ html,
+ viewState: extractHiddenField(html, "__VIEWSTATE"),
+ viewStateGenerator: extractHiddenField(html, "__VIEWSTATEGENERATOR"),
+ eventValidation: extractHiddenField(html, "__EVENTVALIDATION"),
+ currentYear,
+ };
+}
+
+interface PageState {
+ viewState: string;
+ viewStateGenerator: string;
+ eventValidation: string;
+ currentYear: number;
+}
+
+async function fetchYear(
+ year: number,
+ state: PageState
+): Promise<{ teams: TeamRank[]; vsRecords: TeamVsRecord[]; newState: PageState }> {
+ const formData = new URLSearchParams();
+ formData.set(`${PREFIX}ScriptManager`, `${PREFIX}udpRecord|${PREFIX}ddlYear`);
+ formData.set(`${PREFIX}ddlYear`, String(year));
+ formData.set(`${PREFIX}ddlSeries`, "0");
+ // hfSearchYear/Date는 "현재 페이지가 보여주던 연도" (변경 전)
+ formData.set(`${PREFIX}hfSearchYear`, String(state.currentYear));
+ formData.set(`${PREFIX}hfSearchDate`, `${state.currentYear}1231`);
+ formData.set(`${PREFIX}hfSearchSeries`, "0");
+ formData.set("__EVENTTARGET", `${PREFIX}ddlYear`);
+ formData.set("__EVENTARGUMENT", "");
+ formData.set("__LASTFOCUS", "");
+ formData.set("__VIEWSTATE", state.viewState);
+ formData.set("__VIEWSTATEGENERATOR", state.viewStateGenerator);
+ formData.set("__EVENTVALIDATION", state.eventValidation);
+ formData.set("__ASYNCPOST", "true");
+
+ const res = await fetch(BASE_URL, {
+ method: "POST",
+ headers: {
+ "Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
+ "X-MicrosoftAjax": "Delta=true",
+ "X-Requested-With": "XMLHttpRequest",
+ "User-Agent":
+ "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/147.0.0.0 Safari/537.36",
+ Referer: BASE_URL,
+ Origin: "https://www.koreabaseball.com",
+ },
+ body: formData.toString(),
+ });
+
+ const text = await res.text();
+
+ // ASP.NET AJAX 델타 응답에서 HTML 부분 추출 (첫 번째 updatePanel 섹션)
+ const panelMatch = text.match(
+ /updatePanel\|[^|]*\|([\s\S]*?)\|[0-9]+\|hiddenField\|/
+ );
+ const htmlContent = panelMatch?.[1] ?? text;
+
+ // 업데이트된 hidden field 추출
+ const vsMatch = text.match(/__VIEWSTATE\|([^|]*)\|/);
+ const vsgMatch = text.match(/__VIEWSTATEGENERATOR\|([^|]*)\|/);
+ const evMatch = text.match(/__EVENTVALIDATION\|([^|]*)\|/);
+
+ const newState: PageState = {
+ viewState: vsMatch?.[1] ?? state.viewState,
+ viewStateGenerator: vsgMatch?.[1] ?? state.viewStateGenerator,
+ eventValidation: evMatch?.[1] ?? state.eventValidation,
+ currentYear: year,
+ };
+
+ const teams = parseRankTable(htmlContent);
+ const vsRecords = parseVsTable(htmlContent);
+
+ return { teams, vsRecords, newState };
+}
+
+// ── 출력 ──
+
+function printTable(year: number, teams: TeamRank[]) {
+ console.log(`\n${"═".repeat(90)}`);
+ console.log(` KBO ${year} 정규시즌 팀 순위`);
+ console.log(`${"═".repeat(90)}`);
+
+ if (teams.length === 0) {
+ console.log(" 데이터를 찾을 수 없습니다.");
+ return;
+ }
+
+ const header = [
+ "순위",
+ "팀명",
+ "경기",
+ "승",
+ "패",
+ "무",
+ "승률",
+ "게임차",
+ "최근10경기",
+ "연속",
+ ];
+ const widths = [4, 6, 4, 3, 3, 3, 6, 6, 12, 8];
+
+ const padCell = (val: string, w: number) => {
+ // 한글 글자는 폭 2로 계산
+ const displayWidth = [...val].reduce(
+ (sum, ch) => sum + (ch.charCodeAt(0) > 0x7f ? 2 : 1),
+ 0
+ );
+ return val + " ".repeat(Math.max(0, w - displayWidth));
+ };
+
+ console.log(
+ " " + header.map((h, i) => padCell(h, widths[i])).join(" │ ")
+ );
+ console.log(" " + widths.map((w) => "─".repeat(w)).join("─┼─"));
+
+ for (const t of teams) {
+ const row = [
+ t.순위,
+ t.팀명,
+ t.경기,
+ t.승,
+ t.패,
+ t.무,
+ t.승률,
+ t.게임차,
+ t.최근10경기,
+ t.연속,
+ ];
+ console.log(
+ " " + row.map((val, i) => padCell(val, widths[i])).join(" │ ")
+ );
+ }
+}
+
+function printVsTable(year: number, vsRecords: TeamVsRecord[]) {
+ if (vsRecords.length === 0) return;
+
+ console.log(`\n${"─".repeat(90)}`);
+ console.log(` KBO ${year} 팀간 승패표 (승-패-무)`);
+ console.log(`${"─".repeat(90)}`);
+
+ const teamNames = vsRecords.map((r) => r.팀명);
+ const colW = 7;
+
+ const padCell = (val: string, w: number) => {
+ const displayWidth = [...val].reduce(
+ (sum, ch) => sum + (ch.charCodeAt(0) > 0x7f ? 2 : 1),
+ 0
+ );
+ return val + " ".repeat(Math.max(0, w - displayWidth));
+ };
+
+ const fmtWLD = (wld: WinLossDraw) => `${wld[0]}-${wld[1]}-${wld[2]}`;
+
+ // 헤더
+ console.log(
+ " " +
+ padCell("", 6) +
+ " │ " +
+ teamNames.map((n) => padCell(n, colW)).join("│ ") +
+ "│ " +
+ padCell("합계", colW)
+ );
+ console.log(
+ " " +
+ "─".repeat(6) +
+ "─┼─" +
+ teamNames.map(() => "─".repeat(colW)).join("┼─") +
+ "┼─" +
+ "─".repeat(colW)
+ );
+
+ for (const rec of vsRecords) {
+ const cells = teamNames.map((name) => {
+ if (name === rec.팀명) return padCell(" ■", colW);
+ const wld = rec.상대전적[name];
+ return padCell(wld ? fmtWLD(wld) : "-", colW);
+ });
+ console.log(
+ " " +
+ padCell(rec.팀명, 6) +
+ " │ " +
+ cells.join("│ ") +
+ "│ " +
+ padCell(fmtWLD(rec.합계), colW)
+ );
+ }
+}
+
+// ── JSON 출력 ──
+
+interface YearResult {
+ year: number;
+ teams: TeamRank[];
+ vsRecords: TeamVsRecord[];
+}
+
+function printJson(results: YearResult[]) {
+ console.log(JSON.stringify(results, null, 2));
+}
+
+// ── Main ──
+
+async function main() {
+ const args = process.argv.slice(2);
+
+ let jsonMode = false;
+ const years: number[] = [];
+
+ for (const arg of args) {
+ if (arg === "--json") {
+ jsonMode = true;
+ } else {
+ const y = parseInt(arg, 10);
+ if (y >= 1982 && y <= 2026) {
+ years.push(y);
+ } else {
+ console.error(`유효하지 않은 연도: ${arg} (1982~2026)`);
+ process.exit(1);
+ }
+ }
+ }
+
+ if (years.length === 0) {
+ years.push(new Date().getFullYear());
+ }
+
+ console.log("KBO 팀 순위 데이터를 가져오는 중...\n");
+
+ // 1단계: 초기 페이지 가져오기
+ const initial = await getInitialPage();
+ let state: PageState = {
+ viewState: initial.viewState,
+ viewStateGenerator: initial.viewStateGenerator,
+ eventValidation: initial.eventValidation,
+ currentYear: initial.currentYear,
+ };
+
+ const results: YearResult[] = [];
+
+ // 2단계: 연도별 데이터 가져오기 (순차 — 상태 의존)
+ for (const year of years) {
+ process.stdout.write(` ${year}년 데이터 요청 중...`);
+
+ if (year === initial.currentYear && results.length === 0) {
+ const teams = parseRankTable(initial.html);
+ const vsRecords = parseVsTable(initial.html);
+ results.push({ year, teams, vsRecords });
+ console.log(` ${teams.length}개 팀 로드 완료 (초기 페이지)`);
+ } else {
+ const { teams, vsRecords, newState } = await fetchYear(year, state);
+ state = newState;
+ results.push({ year, teams, vsRecords });
+ console.log(` ${teams.length}개 팀 로드 완료`);
+ }
+ }
+
+ // 3단계: 출력
+ if (jsonMode) {
+ printJson(results);
+ } else {
+ for (const { year, teams, vsRecords } of results) {
+ printTable(year, teams);
+ printVsTable(year, vsRecords);
+ }
+ }
+}
+
+main().catch((err) => {
+ console.error("오류 발생:", err);
+ process.exit(1);
+});
|