coulomb-social/scripts/export-bubble-inventory.mjs

#!/usr/bin/env node
/**
 * Authenticated Bubble inventory export for CSOC-WP-0001-T02.
 *
 * Uses Playwright storageState (never prints secrets). Captures routes, page
 * text/HTML samples, and network JSON-ish responses for offline mapping.
 *
 * Usage:
 *   ./scripts/capture-auth-state.sh   # once
 *   node ./scripts/export-bubble-inventory.mjs
 *
 * Env:
 *   COULOMB_SOCIAL_STORAGE_STATE  default ~/.config/coulomb-social/auth/storage-state.json
 *   COULOMB_SOCIAL_URL            default https://coulomb.social
 *   COULOMB_SOCIAL_EXPORT_DIR     default ~/.config/coulomb-social/auth/exports/<stamp>
 */
import { chromium } from "playwright";
import {
  mkdirSync,
  writeFileSync,
  chmodSync,
  existsSync,
  appendFileSync,
} from "node:fs";
import { homedir } from "node:os";
import { join } from "node:path";

const BASE_URL = process.env.COULOMB_SOCIAL_URL || "https://coulomb.social";
const AUTH_DIR =
  process.env.COULOMB_SOCIAL_AUTH_DIR ||
  join(homedir(), ".config", "coulomb-social", "auth");
const STATE_PATH =
  process.env.COULOMB_SOCIAL_STORAGE_STATE ||
  join(AUTH_DIR, "storage-state.json");

const ROUTES = [
  "/",
  "/vw_pages",
  "/vw_spaces",
  "/vw_chunks",
  "/vw_search",
  "/vw_feed",
  "/vw_invites",
  "/vw_likes",
  "/vw_new",
  "/reset_pw",
];

function stamp() {
  const d = new Date();
  const p = (n) => String(n).padStart(2, "0");
  return (
    d.getUTCFullYear() +
    p(d.getUTCMonth() + 1) +
    p(d.getUTCDate()) +
    "-" +
    p(d.getUTCHours()) +
    p(d.getUTCMinutes()) +
    p(d.getUTCSeconds())
  );
}

function ensureDir(path, mode = 0o700) {
  mkdirSync(path, { recursive: true, mode });
  try {
    chmodSync(path, mode);
  } catch {
    /* best effort */
  }
}

function writeText(path, text) {
  writeFileSync(path, text, { mode: 0o600 });
  try {
    chmodSync(path, 0o600);
  } catch {
    /* best effort */
  }
}

async function main() {
  if (!existsSync(STATE_PATH)) {
    console.error(`ERROR: No session file at ${STATE_PATH}`);
    console.error("Capture a session first (do not paste passwords into chat):");
    console.error("  ./scripts/capture-auth-state.sh");
    console.error("Docs: docs/migration/bubble-auth-and-export.md");
    process.exit(1);
  }

  const outRoot =
    process.env.COULOMB_SOCIAL_EXPORT_DIR ||
    join(AUTH_DIR, "exports", stamp());
  const routesDir = join(outRoot, "routes");
  const netDir = join(outRoot, "network");
  ensureDir(outRoot);
  ensureDir(routesDir);
  ensureDir(netDir);

  console.log(`Session: ${STATE_PATH} (present; not printed)`);
  console.log(`Export → ${outRoot}`);

  const networkLog = [];
  const browser = await chromium.launch({ headless: true });
  try {
    const context = await browser.newContext({ storageState: STATE_PATH });
    const page = await context.newPage();

    page.on("response", async (response) => {
      try {
        const url = response.url();
        const ct = (response.headers()["content-type"] || "").toLowerCase();
        const status = response.status();
        if (status < 200 || status >= 400) return;
        const interesting =
          ct.includes("json") ||
          url.includes("/api/") ||
          url.includes("elasticsearch") ||
          url.includes("userdata") ||
          url.includes("obj");
        if (!interesting) return;
        let body = "";
        try {
          body = await response.text();
        } catch {
          return;
        }
        if (!body || body.length > 2_000_000) return;
        const entry = {
          url: url.slice(0, 500),
          status,
          contentType: ct.slice(0, 120),
          bytes: body.length,
          at: new Date().toISOString(),
        };
        networkLog.push(entry);
        const safe = String(networkLog.length).padStart(3, "0");
        writeText(join(netDir, `${safe}.meta.json`), JSON.stringify(entry, null, 2));
        writeText(join(netDir, `${safe}.body.txt`), body);
      } catch {
        /* ignore single response failures */
      }
    });

    // Warm session
    await page.goto(BASE_URL, { waitUntil: "domcontentloaded", timeout: 60_000 });
    await page.waitForTimeout(2500);
    const probeUrl = page.url();
    const probeTitle = await page.title().catch(() => "");
    writeText(
      join(outRoot, "probe.txt"),
      `url=${probeUrl}\ntitle=${probeTitle}\n`,
    );
    console.log(`Probe: ${probeUrl} — ${probeTitle}`);
    try {
      const path = new URL(probeUrl).pathname;
      if (path === "/" || path === "") {
        console.warn(
          "WARNING: Still on site root — session may be guest/expired. Re-run capture-auth-state.sh",
        );
      }
    } catch {
      /* ignore */
    }

    const summary = [];
    summary.push(`# Bubble export ${stamp()}`);
    summary.push("");
    summary.push(`- base: ${BASE_URL}`);
    summary.push(`- probe: ${probeUrl}`);
    summary.push(`- title: ${probeTitle}`);
    summary.push(`- session file: present (path not echoed in summary for safety)`);
    summary.push("");
    summary.push("## Routes");
    summary.push("");

    for (const route of ROUTES) {
      const url = route.startsWith("http") ? route : `${BASE_URL.replace(/\/$/, "")}${route}`;
      const slug = route.replace(/\W+/g, "_").replace(/^_|_$/g, "") || "root";
      console.log(`  GET ${route}`);
      try {
        const resp = await page.goto(url, {
          waitUntil: "domcontentloaded",
          timeout: 60_000,
        });
        await page.waitForTimeout(3500);
        const finalUrl = page.url();
        const title = await page.title().catch(() => "");
        const status = resp?.status() ?? 0;
        const html = await page.content();
        const text = await page.evaluate(() => document.body?.innerText || "");
        writeText(join(routesDir, `${slug}.html`), html);
        writeText(join(routesDir, `${slug}.txt`), text);
        writeText(
          join(routesDir, `${slug}.meta.json`),
          JSON.stringify(
            { route, url, finalUrl, title, status, htmlBytes: html.length, textBytes: text.length },
            null,
            2,
          ),
        );
        summary.push(
          `- \`${route}\` → ${status} ${title} (${text.length} chars text, final ${finalUrl})`,
        );
      } catch (err) {
        summary.push(`- \`${route}\` → ERROR ${err.message}`);
        console.warn(`  failed: ${err.message}`);
      }
    }

    // Public meta (no auth needed, still useful)
    try {
      const metaResp = await page.request.get(`${BASE_URL.replace(/\/$/, "")}/api/1.1/meta`);
      const metaText = await metaResp.text();
      writeText(join(outRoot, "api-meta.json"), metaText);
      summary.push("");
      summary.push("## API meta");
      summary.push(`- status ${metaResp.status()} (${metaText.length} bytes)`);
    } catch (err) {
      summary.push(`- api meta ERROR ${err.message}`);
    }

    writeText(
      join(outRoot, "network-index.json"),
      JSON.stringify(networkLog, null, 2),
    );
    summary.push("");
    summary.push("## Network captures");
    summary.push(`- ${networkLog.length} JSON/API-ish responses under network/`);
    summary.push("");
    summary.push("## Next");
    summary.push("- Review routes/*.txt for Title/Abstractor/Visual/copy cues");
    summary.push("- Map into docs/migration/schema-mapping (redacted notes only in git)");
    summary.push("- Do not commit this export directory");
    summary.push("");

    writeText(join(outRoot, "SUMMARY.md"), summary.join("\n"));
    console.log("");
    console.log(`Done. Summary: ${join(outRoot, "SUMMARY.md")}`);
    console.log(`Network hits: ${networkLog.length}`);
  } finally {
    await browser.close();
  }
}

main().catch((err) => {
  console.error(err.message || err);
  process.exit(1);
});