init

bizmax-download-excel.ts

@@ -0,0 +1,451 @@
+import axios, {
+  type AxiosRequestConfig,
+  type AxiosResponse,
+  type RawAxiosResponseHeaders,
+} from "axios";
+import * as cheerio from "cheerio";
+import { createHash } from "node:crypto";
+import { mkdir, writeFile } from "node:fs/promises";
+import path from "node:path";
+
+const DEFAULT_BASE_URLS = ["https://hyun.bizmax.net", "https://hyun2.bizmax.net"];
+const BASE_URLS = (
+  process.env.BIZMAX_BASE_URLS ??
+  process.env.BIZMAX_BASE_URL ??
+  DEFAULT_BASE_URLS.join(",")
+)
+  .split(",")
+  .map((value) => value.trim())
+  .filter(Boolean);
+const LOGIN_ID = process.env.BIZMAX_ID;
+const LOGIN_PASSWORD = process.env.BIZMAX_PASSWORD;
+const OUTPUT_DIR = process.env.BIZMAX_OUTPUT_DIR ?? path.resolve(process.cwd(), "output");
+const CUSTOMER_LIST_PATH = "/customer/tpl/customer_list.php";
+const RANGE_START = process.env.BIZMAX_RANGE_START ?? "2016-01-01";
+const BATCH_MONTHS = Number(process.env.BIZMAX_BATCH_MONTHS ?? "3");
+
+type LoginResponse = {
+  ret_code?: number;
+  ret_msg?: string;
+  login_succ?: string | number;
+  u_sms_a?: string;
+  f_url?: string;
+  lt?: string;
+};
+
+class SimpleCookieJar {
+  private readonly cookies = new Map<string, string>();
+
+  addFromHeaders(setCookie: string[] = []): void {
+    for (const cookieLine of setCookie) {
+      const [pair] = cookieLine.split(";", 1);
+      if (!pair) {
+        continue;
+      }
+
+      const eqIndex = pair.indexOf("=");
+      if (eqIndex <= 0) {
+        continue;
+      }
+
+      const name = pair.slice(0, eqIndex).trim();
+      const value = pair.slice(eqIndex + 1).trim();
+      this.cookies.set(name, value);
+    }
+  }
+
+  toHeader(): string {
+    return Array.from(this.cookies.entries())
+      .map(([name, value]) => `${name}=${value}`)
+      .join("; ");
+  }
+}
+
+function md5(value: string): string {
+  return createHash("md5").update(value).digest("hex");
+}
+
+function getRequiredEnv(name: string, value: string | undefined): string {
+  if (!value) {
+    throw new Error(`Missing required environment variable: ${name}`);
+  }
+
+  return value;
+}
+
+function getSetCookieHeaders(headers: RawAxiosResponseHeaders | Record<string, unknown>): string[] {
+  const headerValue = headers["set-cookie"];
+  if (Array.isArray(headerValue)) {
+    return headerValue.filter((item): item is string => typeof item === "string");
+  }
+
+  return [];
+}
+
+async function request<T = unknown>(
+  jar: SimpleCookieJar,
+  config: AxiosRequestConfig,
+): Promise<AxiosResponse<T>> {
+  const headers: Record<string, string> = {
+    "User-Agent":
+      "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36",
+    ...(config.headers as Record<string, string> | undefined),
+  };
+
+  const cookieHeader = jar.toHeader();
+  if (cookieHeader) {
+    headers.Cookie = cookieHeader;
+  }
+
+  const response = await axios<T>({
+    ...config,
+    headers,
+    validateStatus: () => true,
+    maxRedirects: 0,
+  });
+
+  jar.addFromHeaders(getSetCookieHeaders(response.headers));
+
+  if (response.status >= 400) {
+    throw new Error(`Request failed: ${config.method ?? "GET"} ${config.url} -> ${response.status}`);
+  }
+
+  return response;
+}
+
+function absoluteUrl(baseUrl: string, urlOrPath: string): string {
+  return new URL(urlOrPath, baseUrl).toString();
+}
+
+function parseLoginPage(html: string): { appName: string; authStep: string } {
+  const $ = cheerio.load(html);
+  const appName = $('input[name="app_name"]').val()?.toString() ?? "Chrome";
+  const authStep = $('input[name="auth_step"]').val()?.toString() ?? "1";
+  return { appName, authStep };
+}
+
+function parseExcelForm(html: string): { action: string; fields: Record<string, string> } {
+  const $ = cheerio.load(html);
+  const form = $('form[name="excel_form"]');
+  if (!form.length) {
+    throw new Error("excel_form was not found in the customer list page");
+  }
+
+  const action =
+    form.attr("action") ||
+    (() => {
+      const tpl = form.attr("ctm_tpl_excel") || "customer_excel.php";
+      return `/customer/tpl/${tpl}`;
+    })();
+  if (!action) {
+    throw new Error("excel_form action was not found");
+  }
+
+  const fields: Record<string, string> = {};
+  form.find("input[name]").each((_, element) => {
+    const input = $(element);
+    const name = input.attr("name");
+    if (!name) {
+      return;
+    }
+
+    fields[name] = input.val()?.toString() ?? "";
+  });
+
+  return { action, fields };
+}
+
+function parseTotalCount(html: string): number {
+  const $ = cheerio.load(html);
+  const explicitValue =
+    $('form[name="excel_form"] input[name="ctm_list_total_cnt"]').val()?.toString() ??
+    $('form[name="search_form"] input[name="ctm_list_total_cnt"]').val()?.toString();
+
+  if (explicitValue) {
+    const parsed = Number(explicitValue.replace(/,/g, ""));
+    if (!Number.isNaN(parsed)) {
+      return parsed;
+    }
+  }
+
+  const bodyText = $.text();
+  const match = bodyText.match(/Total\s*[:：]\s*([\d,]+)건/i);
+  if (match) {
+    const parsed = Number(match[1].replace(/,/g, ""));
+    if (!Number.isNaN(parsed)) {
+      return parsed;
+    }
+  }
+
+  return 0;
+}
+
+function parseSearchForm(html: string): Record<string, string> {
+  const $ = cheerio.load(html);
+  const form = $('form[name="search_form"]');
+  if (!form.length) {
+    throw new Error("search_form was not found in the customer list page");
+  }
+
+  const fields: Record<string, string> = {};
+
+  form.find("input[name], select[name], textarea[name]").each((_, element) => {
+    const input = $(element);
+    const name = input.attr("name");
+    if (!name) {
+      return;
+    }
+
+    const tagName = element.tagName.toLowerCase();
+    const type = (input.attr("type") ?? "").toLowerCase();
+
+    if (type === "checkbox" || type === "radio") {
+      if (!input.is(":checked")) {
+        return;
+      }
+      fields[name] = input.val()?.toString() ?? "on";
+      return;
+    }
+
+    if (tagName === "select") {
+      fields[name] = input.find("option[selected]").val()?.toString() ?? input.val()?.toString() ?? "";
+      return;
+    }
+
+    fields[name] = input.val()?.toString() ?? "";
+  });
+
+  return fields;
+}
+
+function buildFormBody(fields: Record<string, string>): URLSearchParams {
+  const params = new URLSearchParams();
+  for (const [name, value] of Object.entries(fields)) {
+    params.set(name, value);
+  }
+  return params;
+}
+
+function getFileNameFromDisposition(contentDisposition: string | undefined): string | null {
+  if (!contentDisposition) {
+    return null;
+  }
+
+  const decodeFileName = (value: string): string => {
+    try {
+      return decodeURIComponent(value);
+    } catch {
+      return value;
+    }
+  };
+
+  const utf8Match = contentDisposition.match(/filename\*=UTF-8''([^;]+)/i);
+  if (utf8Match) {
+    return decodeFileName(utf8Match[1]);
+  }
+
+  const basicMatch = contentDisposition.match(/filename="?([^"]+)"?/i);
+  if (basicMatch) {
+    return decodeFileName(basicMatch[1]);
+  }
+
+  return null;
+}
+
+function toYmd(date: Date): string {
+  return date.toISOString().slice(0, 10);
+}
+
+function addMonths(date: Date, months: number): Date {
+  return new Date(Date.UTC(date.getUTCFullYear(), date.getUTCMonth() + months, date.getUTCDate()));
+}
+
+function addDays(date: Date, days: number): Date {
+  return new Date(date.getTime() + days * 24 * 60 * 60 * 1000);
+}
+
+function minDate(left: Date, right: Date): Date {
+  return left.getTime() <= right.getTime() ? left : right;
+}
+
+function buildQuarterRanges(startYmd: string, batchMonths: number): Array<{ start: string; end: string }> {
+  const startDate = new Date(`${startYmd}T00:00:00Z`);
+  const today = new Date();
+  const endDate = new Date(Date.UTC(today.getUTCFullYear(), today.getUTCMonth(), today.getUTCDate()));
+
+  if (Number.isNaN(startDate.getTime())) {
+    throw new Error(`Invalid BIZMAX_RANGE_START: ${startYmd}`);
+  }
+
+  if (!Number.isInteger(batchMonths) || batchMonths <= 0) {
+    throw new Error(`Invalid BIZMAX_BATCH_MONTHS: ${batchMonths}`);
+  }
+
+  const ranges: Array<{ start: string; end: string }> = [];
+  let rangeStart = startDate;
+
+  while (rangeStart.getTime() <= endDate.getTime()) {
+    const nextRangeStart = addMonths(rangeStart, batchMonths);
+    const rangeEnd = minDate(addDays(nextRangeStart, -1), endDate);
+    ranges.push({ start: toYmd(rangeStart), end: toYmd(rangeEnd) });
+    rangeStart = nextRangeStart;
+  }
+
+  return ranges;
+}
+
+async function main(): Promise<void> {
+  const loginId = getRequiredEnv("BIZMAX_ID", LOGIN_ID);
+  const loginPassword = getRequiredEnv("BIZMAX_PASSWORD", LOGIN_PASSWORD);
+  await mkdir(OUTPUT_DIR, { recursive: true });
+  const ranges = buildQuarterRanges(RANGE_START, BATCH_MONTHS);
+
+  for (const baseUrl of BASE_URLS) {
+    const jar = new SimpleCookieJar();
+    const siteSlug = new URL(baseUrl).hostname.replace(/\./g, "_");
+
+    const loginPageResponse = await request<string>(jar, {
+      method: "GET",
+      url: absoluteUrl(baseUrl, "/"),
+      responseType: "text",
+    });
+
+    const { appName, authStep } = parseLoginPage(loginPageResponse.data);
+
+    const loginBody = new URLSearchParams({
+      from: "pc",
+      nhoj: loginId,
+      mluap: md5(loginPassword),
+      app_name: appName,
+      captcha_str: "",
+      auth_step: authStep,
+      force_sms: "n",
+    });
+
+    const loginResponse = await request<LoginResponse>(jar, {
+      method: "POST",
+      url: absoluteUrl(baseUrl, "/login_pcs_jr.php"),
+      data: loginBody.toString(),
+      responseType: "json",
+      headers: {
+        "Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
+        Accept: "application/json, text/javascript, */*; q=0.01",
+        Referer: absoluteUrl(baseUrl, "/"),
+        Origin: baseUrl,
+        "X-Requested-With": "XMLHttpRequest",
+      },
+    });
+
+    const loginData = loginResponse.data;
+    if (loginData.u_sms_a === "y") {
+      throw new Error(`[${baseUrl}] This account requires SMS verification. Axios-only automation cannot finish that step.`);
+    }
+
+    if (`${loginData.login_succ ?? ""}` !== "1" || !loginData.f_url) {
+      throw new Error(`[${baseUrl}] Login failed: ${loginData.ret_msg ?? "unknown error"}`);
+    }
+
+    const mainPageUrl = absoluteUrl(baseUrl, loginData.f_url);
+    await request<string>(jar, {
+      method: "GET",
+      url: mainPageUrl,
+      responseType: "text",
+      headers: {
+        Referer: absoluteUrl(baseUrl, "/"),
+      },
+    });
+
+    const customerListResponse = await request<string>(jar, {
+      method: "POST",
+      url: absoluteUrl(baseUrl, CUSTOMER_LIST_PATH),
+      data: new URLSearchParams({
+        _ihr: "n",
+        _ihl: "n",
+        _content_only: "n",
+        site_gubun: "pc",
+      }).toString(),
+      responseType: "text",
+      headers: {
+        "Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
+        Referer: mainPageUrl,
+        Origin: baseUrl,
+        "X-Requested-With": "XMLHttpRequest",
+      },
+    });
+
+    const baseSearchFields = parseSearchForm(customerListResponse.data);
+
+    for (const range of ranges) {
+      const searchFields = {
+        ...baseSearchFields,
+        _ihr: "n",
+        _ihl: "n",
+        _content_only: "n",
+        site_gubun: "pc",
+        mode: "search",
+        page: "",
+        use_date_1: "a.reg_date",
+        use_date_1_src: "a.reg_date",
+        use_date_1_dsp: "- 접수일자",
+        start_date_1: range.start,
+        end_date_1: range.end,
+      };
+
+      const rangedCustomerListResponse = await request<string>(jar, {
+        method: "POST",
+        url: absoluteUrl(baseUrl, CUSTOMER_LIST_PATH),
+        data: buildFormBody(searchFields).toString(),
+        responseType: "text",
+        headers: {
+          "Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
+          Referer: mainPageUrl,
+          Origin: baseUrl,
+          "X-Requested-With": "XMLHttpRequest",
+        },
+      });
+
+      const totalCount = parseTotalCount(rangedCustomerListResponse.data);
+      if (totalCount === 0) {
+        process.stdout.write(`[${siteSlug}] SKIP ${range.start}~${range.end} (0건)\n`);
+        continue;
+      }
+
+      const { action, fields } = parseExcelForm(rangedCustomerListResponse.data);
+      fields.e_type = "1";
+
+      const excelResponse = await request<ArrayBuffer>(jar, {
+        method: "POST",
+        url: absoluteUrl(baseUrl, action),
+        data: buildFormBody(fields).toString(),
+        responseType: "arraybuffer",
+        headers: {
+          "Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
+          Referer: absoluteUrl(baseUrl, CUSTOMER_LIST_PATH),
+          Origin: baseUrl,
+        },
+      });
+
+      const contentDisposition =
+        typeof excelResponse.headers["content-disposition"] === "string"
+          ? excelResponse.headers["content-disposition"]
+          : undefined;
+      const fileName =
+        getFileNameFromDisposition(contentDisposition) ??
+        `bizmax-customer-list-${range.start}-${range.end}.xlsx`;
+
+      const ext = path.extname(fileName);
+      const baseName = ext ? fileName.slice(0, -ext.length) : fileName;
+      const rangedFileName = `${siteSlug}_${baseName}_${range.start}_${range.end}${ext || ".xlsx"}`;
+      const outputPath = path.join(OUTPUT_DIR, rangedFileName);
+      await writeFile(outputPath, Buffer.from(excelResponse.data));
+
+      process.stdout.write(`${outputPath}\n`);
+    }
+  }
+}
+
+main().catch((error: unknown) => {
+  const message = error instanceof Error ? error.message : String(error);
+  process.stderr.write(`${message}\n`);
+  process.exitCode = 1;
+});