<?php
/**
 * Fortune Journals EPMC Data Extractor — Secure UI
 * Location: /home/cpe0003707/public_html/epmc-ui/index.php
 *
 * Requires (outside public_html):
 *  - /home/cpe0003707/epmc/.ui_auth.php  (login user + password hash)
 *  - /home/cpe0003707/epmc/.db.php       (DB config)
 *
 * Features:
 *  - Login (session)
 *  - Run pipeline (month-by-month + 10-day chunks inside each month)
 *  - Live progress bar + live output (AJAX)
 *  - Deletes raw files immediately after successful import (month folder cleanup)
 *  - Download CSV from DB (latest / all / filtered by firstPublicationDate)
 *
 * SECURITY:
 *  - Also enable cPanel -> Directory Privacy on /public_html/epmc-ui/
 */

declare(strict_types=1);
session_start();
ini_set('display_errors', '1');
error_reporting(E_ALL);
set_time_limit(0);

$ROOT = "/home/cpe0003707/epmc";

$AUTH_FILE = $ROOT . "/.ui_auth.php";
$DB_CONF   = $ROOT . "/.db.php";
$STATE_DIR = $ROOT . "/state";
$LOCK_FILE = $STATE_DIR . "/ui_run.lock";

@mkdir($ROOT . "/logs", 0755, true);
@mkdir($STATE_DIR, 0755, true);

$PHP  = "/usr/local/bin/php";
$PERL = "/usr/bin/perl";

$CMD_FETCH   = "$PHP -q $ROOT/scripts/epmc_daily_fetch.php";
$CMD_MERGE   = "$PHP -q $ROOT/scripts/merge_xml.php";
$CMD_EXTRACT = "$PERL $ROOT/scripts/extract_pubmed_emails.pl";
$CMD_IMPORT  = "$PHP -q $ROOT/scripts/import_email_csv.php";

function h(string $s): string { return htmlspecialchars($s, ENT_QUOTES, 'UTF-8'); }

function safe_date(?string $d): string {
  if (!$d) return "";
  return preg_match('/^\d{4}-\d{2}-\d{2}$/', $d) ? $d : "";
}

function dt_utc(string $ymd): DateTime {
  return new DateTime($ymd, new DateTimeZone("UTC"));
}

function ymd(DateTime $dt): string { return $dt->format("Y-m-d"); }

function require_auth(): void {
  if (empty($_SESSION['epmc_ui_authed'])) {
    http_response_code(403);
    echo "Forbidden";
    exit;
  }
}

function load_auth(string $path): array {
  if (!file_exists($path)) return [];
  $a = require $path;
  return is_array($a) ? $a : [];
}

/**
 * Execute a command and return [output, exitCode]
 */
function run_cmd_code(string $cmd): array {
  $lines = [];
  $code = 0;
  // redirect stderr to stdout
  exec($cmd . " 2>&1", $lines, $code);
  return [implode("\n", $lines), $code];
}

function db_connect(string $dbConfPath): PDO {
  if (!file_exists($dbConfPath)) throw new RuntimeException("Missing DB config: $dbConfPath");
  $db = require $dbConfPath;
  if (!is_array($db) || empty($db['host']) || empty($db['db']) || empty($db['user'])) {
    throw new RuntimeException("Invalid DB config format in .db.php");
  }
  $charset = $db['charset'] ?? 'utf8mb4';
  $dsn = "mysql:host={$db['host']};dbname={$db['db']};charset={$charset}";
  $pdo = new PDO($dsn, $db['user'], $db['pass'] ?? '', [
    PDO::ATTR_ERRMODE => PDO::ERRMODE_EXCEPTION,
    PDO::ATTR_DEFAULT_FETCH_MODE => PDO::FETCH_ASSOC,
  ]);
  $pdo->exec("SET NAMES $charset");
  return $pdo;
}

function csv_download(string $filename, array $header, iterable $rows): void {
  header('Content-Type: text/csv; charset=utf-8');
  header('Content-Disposition: attachment; filename="'.$filename.'"');
  $out = fopen('php://output', 'w');
  fputcsv($out, $header);
  foreach ($rows as $r) fputcsv($out, $r);
  fclose($out);
  exit;
}

/**
 * Safe recursive delete with guard (only deletes inside /home/cpe0003707/epmc/raw/)
 */
function rrmdir_safe(string $dir): bool {
  $dir = rtrim($dir, "/");
  $allowedRoot = "/home/cpe0003707/epmc/raw";
  if ($dir === "" || !is_dir($dir)) return false;
  if (strpos($dir, $allowedRoot) !== 0) return false; // guard

  $items = scandir($dir);
  if ($items === false) return false;

  foreach ($items as $item) {
    if ($item === "." || $item === "..") continue;
    $path = $dir . "/" . $item;
    if (is_dir($path)) rrmdir_safe($path);
    else @unlink($path);
  }
  return @rmdir($dir);
}

/* -----------------------
   LOGIN / LOGOUT
------------------------ */
$auth = load_auth($AUTH_FILE);

if (isset($_GET['logout'])) {
  session_destroy();
  header("Location: ./");
  exit;
}

$login_error = "";
if (isset($_POST['login_user'], $_POST['login_pass'])) {
  $u = trim((string)$_POST['login_user']);
  $p = (string)$_POST['login_pass'];

  if (!$auth || empty($auth['user']) || empty($auth['pass_hash'])) {
    $login_error = "Auth config missing. Create /home/cpe0003707/epmc/.ui_auth.php";
  } elseif (hash_equals((string)$auth['user'], $u) && password_verify($p, (string)$auth['pass_hash'])) {
    session_regenerate_id(true);
    $_SESSION['epmc_ui_authed'] = 1;
    header("Location: ./");
    exit;
  } else {
    $login_error = "Invalid username or password.";
  }
}

/* -----------------------
   DOWNLOAD ROUTES
------------------------ */
$action = $_GET['action'] ?? "";

if ($action === "download_all") {
  require_auth();
  $pdo = db_connect($DB_CONF);

  $stmt = $pdo->query("SELECT email, author_name, article_title, doi, first_publication_date, affiliation, grant_ids, grant_agencies, created_at, updated_at
                       FROM epmc_emails ORDER BY id DESC");

  $rows = (function() use ($stmt) {
    while ($r = $stmt->fetch()) {
      yield [
        $r['email'], $r['author_name'], $r['article_title'], $r['doi'],
        $r['first_publication_date'], $r['affiliation'],
        $r['grant_ids'], $r['grant_agencies'],
        $r['created_at'], $r['updated_at']
      ];
    }
  })();

  csv_download("epmc_emails_all.csv",
    ["email","author_name","article_title","doi","first_publication_date","affiliation","grant_ids","grant_agencies","created_at","updated_at"],
    $rows
  );
}

if ($action === "download_latest") {
  require_auth();
  $pdo = db_connect($DB_CONF);

  $stmt = $pdo->query("SELECT email, author_name, article_title, doi, first_publication_date, affiliation, grant_ids, grant_agencies, created_at, updated_at
                       FROM epmc_emails ORDER BY id DESC LIMIT 2000");

  $rows = (function() use ($stmt) {
    while ($r = $stmt->fetch()) {
      yield [
        $r['email'], $r['author_name'], $r['article_title'], $r['doi'],
        $r['first_publication_date'], $r['affiliation'],
        $r['grant_ids'], $r['grant_agencies'],
        $r['created_at'], $r['updated_at']
      ];
    }
  })();

  csv_download("epmc_emails_latest_2000.csv",
    ["email","author_name","article_title","doi","first_publication_date","affiliation","grant_ids","grant_agencies","created_at","updated_at"],
    $rows
  );
}

if ($action === "download_filtered") {
  require_auth();
  $from = safe_date($_GET['from'] ?? "");
  $to   = safe_date($_GET['to'] ?? "");
  if ($from === "" || $to === "") { http_response_code(400); echo "Invalid dates"; exit; }

  $pdo = db_connect($DB_CONF);

  $sql = "SELECT email, author_name, article_title, doi, first_publication_date, affiliation, grant_ids, grant_agencies, created_at, updated_at
          FROM epmc_emails
          WHERE first_publication_date IS NOT NULL
            AND first_publication_date BETWEEN :f AND :t
          ORDER BY id DESC";
  $stmt = $pdo->prepare($sql);
  $stmt->execute([':f'=>$from, ':t'=>$to]);

  $rows = (function() use ($stmt) {
    while ($r = $stmt->fetch()) {
      yield [
        $r['email'], $r['author_name'], $r['article_title'], $r['doi'],
        $r['first_publication_date'], $r['affiliation'],
        $r['grant_ids'], $r['grant_agencies'],
        $r['created_at'], $r['updated_at']
      ];
    }
  })();

  csv_download("epmc_emails_{$from}_to_{$to}.csv",
    ["email","author_name","article_title","doi","first_publication_date","affiliation","grant_ids","grant_agencies","created_at","updated_at"],
    $rows
  );
}

/* -----------------------
   API: RUN ONE MONTH (fetch in 10-day chunks, then merge/extract/import)
   Deletes month folder immediately after successful import.
------------------------ */
if (isset($_GET['api']) && $_GET['api'] === 'run_month') {
  require_auth();
  header('Content-Type: application/json; charset=utf-8');
  set_time_limit(0);

  $from = safe_date($_POST['from'] ?? '');
  $to   = safe_date($_POST['to'] ?? '');
  $mode = strtoupper($_POST['mode'] ?? 'UPDATE_DATE');
  if (!in_array($mode, ['UPDATE_DATE','FIRST_PDATE'], true)) $mode = 'UPDATE_DATE';

  $do_fetch   = !empty($_POST['do_fetch']);
  $do_merge   = !empty($_POST['do_merge']);
  $do_extract = !empty($_POST['do_extract']);
  $do_import  = !empty($_POST['do_import']);

  if ($from === '' || $to === '') {
    echo json_encode(['ok'=>false,'error'=>'Invalid dates. Use YYYY-MM-DD.']); exit;
  }

  // lock to prevent parallel runs
  $lockFp = fopen($LOCK_FILE, "c");
  if (!$lockFp || !flock($lockFp, LOCK_EX | LOCK_NB)) {
    echo json_encode(['ok'=>false,'error'=>'Another run is already in progress.']); exit;
  }

  $out = "";
  $cleanupDone = false;

  try {
    $monthDir = "/home/cpe0003707/epmc/raw/" . $mode . "_" . $from . "_to_" . $to;
    @mkdir($monthDir, 0755, true);

    $out .= "MONTH ($from → $to)\n";
    $out .= "RAW FOLDER: $monthDir\n\n";

    // 1) FETCH as 10-day chunks
    if ($do_fetch) {
      $out .= "▶ FETCH (10-day chunks)\n";
      $cur = dt_utc($from);
      $end = dt_utc($to);
      $chunkNo = 1;

      while ($cur <= $end) {
        $chunkStart = clone $cur;
        $chunkEnd   = clone $cur;
        $chunkEnd->modify("+9 days");
        if ($chunkEnd > $end) $chunkEnd = clone $end;

        $cs = ymd($chunkStart);
        $ce = ymd($chunkEnd);

        $out .= "  - Chunk $chunkNo: $cs → $ce\n";
        [$o, $code] = run_cmd_code($GLOBALS['CMD_FETCH']." ".escapeshellarg($cs)." ".escapeshellarg($ce)." ".escapeshellarg($mode)." ".escapeshellarg($monthDir));
        $out .= $o . "\n";
        if ($code !== 0) {
          echo json_encode(['ok'=>false,'error'=>"FETCH failed (exit $code)", 'output'=>$out]); exit;
        }

        $cur = clone $chunkEnd;
        $cur->modify("+1 day");
        $chunkNo++;
      }
      $out .= "\n";
    }

    // 2) MERGE month folder
    if ($do_merge) {
      $out .= "▶ MERGE (month folder)\n";
      [$o, $code] = run_cmd_code($GLOBALS['CMD_MERGE']." ".escapeshellarg($monthDir));
      $out .= $o . "\n";
      if ($code !== 0) {
        echo json_encode(['ok'=>false,'error'=>"MERGE failed (exit $code)", 'output'=>$out]); exit;
      }
      $out .= "\n";
    }

    // 3) EXTRACT
    if ($do_extract) {
      $out .= "▶ EXTRACT\n";
      [$o, $code] = run_cmd_code($GLOBALS['CMD_EXTRACT']);
      $out .= $o . "\n";
      if ($code !== 0) {
        echo json_encode(['ok'=>false,'error'=>"EXTRACT failed (exit $code)", 'output'=>$out]); exit;
      }
      $out .= "\n";
    }

    // 4) IMPORT + CLEANUP raw folder immediately if success
    if ($do_import) {
      $out .= "▶ IMPORT\n";
      [$o, $code] = run_cmd_code($GLOBALS['CMD_IMPORT']);
      $out .= $o . "\n";
      if ($code !== 0) {
        echo json_encode(['ok'=>false,'error'=>"IMPORT failed (exit $code). Raw folder kept.", 'output'=>$out]); exit;
      }

      // ✅ Cleanup raw month folder immediately after successful import
      $out .= "\n▶ CLEANUP RAW FILES\n";
      $ok = rrmdir_safe($monthDir);
      $cleanupDone = $ok;
      $out .= $ok ? "Deleted raw folder: $monthDir\n" : "Cleanup failed (folder kept): $monthDir\n";
    }

    echo json_encode(['ok'=>true, 'output'=>$out, 'cleanup'=>$cleanupDone]);
    exit;

  } finally {
    flock($lockFp, LOCK_UN);
    fclose($lockFp);
  }
}

/* -----------------------
   If not authed -> show login page
------------------------ */
$authed = !empty($_SESSION['epmc_ui_authed']);
if (!$authed) {
?>
<!doctype html>
<html>
<head>
  <meta charset="utf-8">
  <title>Login — Fortune Journals EPMC Data Extractor</title>
  <meta name="viewport" content="width=device-width, initial-scale=1" />
  <style>
    body{margin:0;font-family:system-ui,-apple-system,Segoe UI,Roboto,Arial;background:linear-gradient(135deg,#0f172a,#0b3b8a,#0f766e);min-height:100vh;display:flex;align-items:center;justify-content:center;padding:24px;}
    .card{width:100%;max-width:520px;background:rgba(255,255,255,.95);border-radius:20px;box-shadow:0 25px 60px rgba(0,0,0,.25);padding:22px 22px 18px;}
    h1{margin:0 0 6px;font-size:22px;}
    p{margin:0 0 14px;color:#475569;font-size:13px;line-height:1.35}
    label{display:block;margin:10px 0 6px;color:#0f172a;font-weight:700;font-size:13px;}
    input{width:100%;padding:12px;border:1px solid #cbd5e1;border-radius:12px;font-size:14px;outline:none}
    input:focus{border-color:#2563eb;box-shadow:0 0 0 4px rgba(37,99,235,.15)}
    button{width:100%;margin-top:14px;padding:12px;border:0;border-radius:12px;background:#2563eb;color:#fff;font-weight:800;font-size:14px;cursor:pointer}
    button:hover{background:#1d4ed8}
    .err{margin-top:10px;background:#fee2e2;color:#991b1b;border:1px solid #fecaca;padding:10px;border-radius:12px;font-size:13px}
    .brand{display:flex;gap:10px;align-items:center;margin-bottom:12px}
    .dot{width:12px;height:12px;border-radius:999px;background:#22c55e;box-shadow:0 0 0 8px rgba(34,197,94,.15)}
    .footer{margin-top:12px;color:#64748b;font-size:12px;line-height:1.35}
    code{background:#f1f5f9;padding:2px 6px;border-radius:8px}
  </style>
</head>
<body>
  <div class="card">
    <div class="brand">
      <div class="dot"></div>
      <div>
        <h1>Fortune Journals EPMC Data Extractor</h1>
        <p>Secure access required to run searches and downloads.</p>
      </div>
    </div>

    <form method="post">
      <label>Username</label>
      <input name="login_user" autocomplete="username" required>
      <label>Password</label>
      <input name="login_pass" type="password" autocomplete="current-password" required>
      <button type="submit">Login</button>
    </form>

    <?php if ($login_error): ?>
      <div class="err"><?= h($login_error) ?></div>
    <?php else: ?>
      <?php if (!file_exists($AUTH_FILE)): ?>
        <div class="err">Auth config missing. Create <code>/home/cpe0003707/epmc/.ui_auth.php</code></div>
      <?php endif; ?>
    <?php endif; ?>

    <div class="footer">
      Tip: also enable <b>cPanel → Directory Privacy</b> for <code>/public_html/epmc-ui/</code>.
    </div>
  </div>
</body>
</html>
<?php
exit;
}

/* -----------------------
   MAIN PAGE (authed)
------------------------ */
$tab = $_GET['tab'] ?? 'run';
if (!in_array($tab, ['run','download'], true)) $tab = 'run';

$defaultFrom = gmdate("Y-m-d", strtotime("-1 day"));
$defaultTo   = gmdate("Y-m-d");

$from = safe_date($_POST['from'] ?? $defaultFrom) ?: $defaultFrom;
$to   = safe_date($_POST['to'] ?? $defaultTo) ?: $defaultTo;

$mode = strtoupper($_POST['mode'] ?? "UPDATE_DATE");
if (!in_array($mode, ["UPDATE_DATE","FIRST_PDATE"], true)) $mode = "UPDATE_DATE";

$do_fetch   = isset($_POST['do_fetch']) ? true : true;
$do_merge   = isset($_POST['do_merge']) ? true : true;
$do_extract = isset($_POST['do_extract']) ? true : true;
$do_import  = isset($_POST['do_import']) ? true : true;
$run_monthly = isset($_POST['run_monthly']);

$stats = ['count'=>null,'updated'=>null];
try {
  $pdo = db_connect($DB_CONF);
  $stats['count'] = (int)$pdo->query("SELECT COUNT(*) c FROM epmc_emails")->fetch()['c'];
  $stats['updated'] = $pdo->query("SELECT MAX(updated_at) m FROM epmc_emails")->fetch()['m'] ?? null;
} catch (Throwable $e) {}

$dlFrom = $from;
$dlTo   = $to;

?>
<!doctype html>
<html>
<head>
  <meta charset="utf-8">
  <title>Fortune Journals EPMC Data Extractor</title>
  <meta name="viewport" content="width=device-width, initial-scale=1" />
  <style>
    :root{
      --bg1:#0f172a; --bg2:#0b3b8a; --bg3:#0f766e;
      --card:#ffffff; --muted:#64748b; --text:#0f172a; --line:#e2e8f0;
      --primary:#2563eb; --primary2:#1d4ed8; --good:#16a34a; --warn:#f59e0b;
    }
    *{box-sizing:border-box}
    body{margin:0;font-family:system-ui,-apple-system,Segoe UI,Roboto,Arial;background:#f1f5f9;color:var(--text)}
    .top{background:linear-gradient(135deg,var(--bg1),var(--bg2),var(--bg3));color:#fff;padding:22px 18px;}
    .wrap{max-width:1120px;margin:0 auto}
    .brand{display:flex;align-items:center;justify-content:space-between;gap:14px;flex-wrap:wrap}
    .title{display:flex;align-items:center;gap:12px}
    .logoDot{width:12px;height:12px;border-radius:999px;background:#22c55e;box-shadow:0 0 0 8px rgba(34,197,94,.18)}
    h1{margin:0;font-size:20px;letter-spacing:.2px}
    .sub{margin:4px 0 0;color:rgba(255,255,255,.85);font-size:13px}
    .right{display:flex;gap:10px;align-items:center}
    .pill{background:rgba(255,255,255,.12);border:1px solid rgba(255,255,255,.18);padding:8px 10px;border-radius:999px;font-size:12px}
    .logout{color:#fff;text-decoration:none;border:1px solid rgba(255,255,255,.28);padding:8px 12px;border-radius:10px}
    .logout:hover{background:rgba(255,255,255,.12)}
    .main{padding:18px}
    .grid{display:grid;grid-template-columns: 1fr;gap:14px}
    @media(min-width:980px){ .grid{grid-template-columns: 1.3fr .7fr} }
    .card{background:var(--card);border:1px solid var(--line);border-radius:16px;box-shadow:0 10px 30px rgba(2,6,23,.06);padding:16px}
    .tabs{display:flex;gap:10px;margin-bottom:12px}
    .tab{display:inline-flex;align-items:center;gap:8px;padding:10px 12px;border-radius:12px;border:1px solid var(--line);text-decoration:none;color:var(--text);background:#fff}
    .tab.active{background:rgba(37,99,235,.08);border-color:rgba(37,99,235,.25);color:#1d4ed8;font-weight:800}
    .row{display:flex;gap:12px;flex-wrap:wrap}
    label{font-weight:800;font-size:12px;color:#0f172a;display:block;margin-bottom:6px}
    input,select{padding:11px 12px;border:1px solid #cbd5e1;border-radius:12px;font-size:14px;outline:none;min-width:250px}
    input:focus,select:focus{border-color:var(--primary);box-shadow:0 0 0 4px rgba(37,99,235,.12)}
    .checks{margin-top:10px;display:grid;gap:8px}
    .check{display:flex;align-items:flex-start;gap:10px}
    .check input{min-width:auto;margin-top:3px}
    .hint{color:var(--muted);font-size:12px;line-height:1.35;margin-top:4px}
    .btnRow{display:flex;gap:10px;flex-wrap:wrap;margin-top:12px}
    .btn{border:0;border-radius:12px;padding:11px 14px;font-weight:900;cursor:pointer}
    .btn.primary{background:var(--primary);color:#fff}
    .btn.primary:hover{background:var(--primary2)}
    .btn.ghost{background:#fff;border:1px solid var(--line)}
    .btn.ghost:hover{background:#f8fafc}
    pre.terminal{margin-top:12px;background:#0b1220;color:#86efac;border-radius:14px;padding:12px;overflow:auto;max-height:520px;font-family:ui-monospace,SFMono-Regular,Menlo,Consolas,monospace;font-size:12.5px;line-height:1.35;border:1px solid rgba(255,255,255,.08)}
    .statBox{display:grid;gap:10px}
    .stat{display:flex;justify-content:space-between;gap:10px;padding:12px;border-radius:14px;border:1px solid var(--line);background:#fff}
    .stat b{font-size:14px}
    .tag{font-size:11px;padding:4px 8px;border-radius:999px;background:rgba(22,163,74,.1);color:var(--good);border:1px solid rgba(22,163,74,.25)}
    .warnTag{background:rgba(245,158,11,.12);color:#92400e;border:1px solid rgba(245,158,11,.3)}
    .dlBox a{display:block;text-decoration:none;border:1px solid var(--line);padding:12px;border-radius:14px;color:var(--text);background:#fff}
    .dlBox a:hover{background:#f8fafc}
    .dlTitle{font-weight:900}
    .dlSub{color:var(--muted);font-size:12px;margin-top:3px}
  </style>
</head>
<body>
  <div class="top">
    <div class="wrap brand">
      <div class="title">
        <div class="logoDot"></div>
        <div>
          <h1>Fortune Journals EPMC Data Extractor</h1>
          <div class="sub">Search Europe PMC (title keywords), extract emails + metadata, store into MySQL.</div>
        </div>
      </div>
      <div class="right">
        <div class="pill"><b>DB rows:</b> <?= $stats['count'] !== null ? h((string)$stats['count']) : '—' ?></div>
        <div class="pill"><b>Last update:</b> <?= $stats['updated'] ? h((string)$stats['updated']) : '—' ?></div>
        <a class="logout" href="?logout=1">Logout</a>
      </div>
    </div>
  </div>

  <div class="main">
    <div class="wrap">
      <div class="tabs">
        <a class="tab <?= $tab==='run'?'active':'' ?>" href="?tab=run">⚙️ Run Pipeline</a>
        <a class="tab <?= $tab==='download'?'active':'' ?>" href="?tab=download">⬇️ Download Data</a>
      </div>

      <div class="grid">
        <div class="card">
          <?php if ($tab === 'run'): ?>
            <h3 style="margin:0 0 10px;">Run Search + Extraction</h3>
            <div class="hint">
              <span class="tag">UPDATE_DATE</span> daily updates
              &nbsp;•&nbsp;
              <span class="tag warnTag">FIRST_PDATE</span> backfill by publication date
              &nbsp;•&nbsp;
              <b>Note:</b> month runs fetch every <b>10 days</b> and cleanup raw files after import.
            </div>

            <form method="post" action="?tab=run" id="runForm">
              <div class="row" style="margin-top:12px;">
                <div>
                  <label>From date (YYYY-MM-DD)</label>
                  <input name="from" value="<?= h($from) ?>">
                </div>
                <div>
                  <label>To date (YYYY-MM-DD)</label>
                  <input name="to" value="<?= h($to) ?>">
                </div>
                <div>
                  <label>Mode</label>
                  <select name="mode">
                    <option value="UPDATE_DATE" <?= $mode==='UPDATE_DATE'?'selected':'' ?>>UPDATE_DATE (daily)</option>
                    <option value="FIRST_PDATE" <?= $mode==='FIRST_PDATE'?'selected':'' ?>>FIRST_PDATE (backfill)</option>
                  </select>
                </div>
              </div>

              <div class="checks">
                <div class="check">
                  <input type="checkbox" name="do_fetch" checked>
                  <div><b>1) Download</b><div class="hint">Fetch in 10-day chunks inside each month</div></div>
                </div>
                <div class="check">
                  <input type="checkbox" name="do_merge" checked>
                  <div><b>2) Merge</b><div class="hint">Merge month folder → pubmed.txt</div></div>
                </div>
                <div class="check">
                  <input type="checkbox" name="do_extract" checked>
                  <div><b>3) Extract</b><div class="hint">Creates email.csv + Emails_not_found_doi.csv</div></div>
                </div>
                <div class="check">
                  <input type="checkbox" name="do_import" checked>
                  <div><b>4) Import to DB</b><div class="hint">Upsert to MySQL then delete raw month folder</div></div>
                </div>

                <div class="check" style="margin-top:6px;">
                  <input type="checkbox" name="run_monthly" id="runMonthly" checked>
                  <div>
                    <b>Run month-by-month automatically</b>
                    <div class="hint">Recommended for long ranges (years). Shows live progress.</div>
                  </div>
                </div>
              </div>

              <div class="btnRow">
                <button class="btn primary" type="submit" id="runBtn">Run Pipeline</button>
                <a class="btn ghost" href="?tab=download" style="text-decoration:none;display:inline-flex;align-items:center;">Go to Downloads</a>
              </div>

              <!-- Progress Bar -->
              <div id="progressBox" style="margin-top:14px; display:none;">
                <div style="display:flex; justify-content:space-between; gap:10px; align-items:center;">
                  <div id="progressText" style="font-weight:800; font-size:13px;">Preparing…</div>
                  <div id="progressPct" style="font-size:12px; color:#64748b;">0%</div>
                </div>
                <div style="height:10px; background:#e2e8f0; border-radius:999px; overflow:hidden; margin-top:8px;">
                  <div id="progressBar" style="height:100%; width:0%; background:#2563eb; border-radius:999px; transition:width .35s ease;"></div>
                </div>
              </div>

              <!-- Live Output -->
              <pre id="liveOutput" class="terminal" style="display:none;"></pre>
            </form>

          <?php else: ?>
            <h3 style="margin:0 0 10px;">Download Saved Data from Database</h3>
            <div class="hint">Exports come from <b>CCM_EPMC_Database.epmc_emails</b>.</div>

            <div class="dlBox" style="display:grid;gap:10px;margin-top:12px;">
              <a href="?action=download_latest">
                <div class="dlTitle">Download Latest 2000 Rows (CSV)</div>
                <div class="dlSub">Quick export for recent data</div>
              </a>

              <a href="?action=download_all">
                <div class="dlTitle">Download ALL Rows (CSV)</div>
                <div class="dlSub">Full export (may be large)</div>
              </a>

              <a href="?action=download_filtered&from=<?= h($dlFrom) ?>&to=<?= h($dlTo) ?>">
                <div class="dlTitle">Download by firstPublicationDate (<?= h($dlFrom) ?> → <?= h($dlTo) ?>)</div>
                <div class="dlSub">Uses your current From/To inputs as filter</div>
              </a>
            </div>

            <div class="hint" style="margin-top:12px;">
              Tip: For historical backfills, use <b>FIRST_PDATE</b> in Run Pipeline, then download filtered range here.
            </div>
          <?php endif; ?>
        </div>

        <div class="card">
          <h3 style="margin:0 0 10px;">Quick Info</h3>
          <div class="statBox">
            <div class="stat">
              <div>
                <b>Database rows</b>
                <div class="hint">Total rows in epmc_emails</div>
              </div>
              <div style="text-align:right">
                <b><?= $stats['count'] !== null ? h((string)$stats['count']) : '—' ?></b><br>
                <span class="tag">LIVE</span>
              </div>
            </div>
            <div class="stat">
              <div>
                <b>Last updated</b>
                <div class="hint">Max(updated_at)</div>
              </div>
              <div style="text-align:right">
                <b><?= $stats['updated'] ? h((string)$stats['updated']) : '—' ?></b><br>
                <span class="tag warnTag">TIME</span>
              </div>
            </div>
            <div class="stat">
              <div>
                <b>Storage policy</b>
                <div class="hint">Raw files cleanup</div>
              </div>
              <div style="text-align:right">
                <span class="tag">Auto-delete</span><br>
                <span class="hint">after import</span>
              </div>
            </div>
          </div>

          <div class="hint" style="margin-top:12px;">
            ✅ Keep <code>/home/cpe0003707/epmc/.db.php</code> and <code>.ui_auth.php</code> outside public_html.<br>
            ✅ Enable cPanel → Directory Privacy for <code>/public_html/epmc-ui/</code>.
          </div>
        </div>
      </div>
    </div>
  </div>

<script>
(function(){
  const form = document.getElementById('runForm');
  const out  = document.getElementById('liveOutput');
  const pbx  = document.getElementById('progressBox');
  const ptxt = document.getElementById('progressText');
  const ppct = document.getElementById('progressPct');
  const pbar = document.getElementById('progressBar');
  const runMonthly = document.getElementById('runMonthly');
  const runBtn = document.getElementById('runBtn');

  if (!form || !out || !pbx || !ptxt || !ppct || !pbar || !runMonthly) return;

  function ymd(d){
    const mm = String(d.getUTCMonth()+1).padStart(2,'0');
    const dd = String(d.getUTCDate()).padStart(2,'0');
    return `${d.getUTCFullYear()}-${mm}-${dd}`;
  }

  function monthChunks(fromStr, toStr){
    const from = new Date(fromStr + "T00:00:00Z");
    const to   = new Date(toStr   + "T00:00:00Z");
    const chunks = [];
    let cur = new Date(Date.UTC(from.getUTCFullYear(), from.getUTCMonth(), from.getUTCDate()));
    while(cur <= to){
      const start = new Date(cur);
      const end = new Date(Date.UTC(cur.getUTCFullYear(), cur.getUTCMonth()+1, 0)); // last day of month
      if(end > to) end.setTime(to.getTime());
      chunks.push([ymd(start), ymd(end)]);
      cur = new Date(Date.UTC(end.getUTCFullYear(), end.getUTCMonth(), end.getUTCDate()+1));
    }
    return chunks;
  }

  async function postMonth(payload){
    const fd = new FormData();
    for (const k in payload) fd.append(k, payload[k]);
    const res = await fetch("?tab=run&api=run_month", { method:"POST", body: fd });
    return await res.json();
  }

  function setProgress(done, total, label){
    pbx.style.display = "block";
    const pct = total > 0 ? Math.round((done/total) * 100) : 0;
    ptxt.textContent = label;
    ppct.textContent = pct + "%";
    pbar.style.width = pct + "%";
  }

  function append(txt){
    out.style.display = "block";
    out.textContent += txt;
    out.scrollTop = out.scrollHeight;
  }

  form.addEventListener('submit', async (e) => {
    // If not month-by-month, allow normal submit (will block page). For best UX keep it checked.
    if (!runMonthly.checked) return;

    e.preventDefault();

    out.textContent = "";
    out.style.display = "block";
    pbx.style.display = "block";

    const from = form.querySelector('input[name="from"]').value.trim();
    const to   = form.querySelector('input[name="to"]').value.trim();
    const mode = form.querySelector('select[name="mode"]').value;

    const chunks = monthChunks(from, to);
    const total = chunks.length;

    // steps
    const steps = {
      do_fetch: form.querySelector('input[name="do_fetch"]').checked ? "1" : "",
      do_merge: form.querySelector('input[name="do_merge"]').checked ? "1" : "",
      do_extract: form.querySelector('input[name="do_extract"]').checked ? "1" : "",
      do_import: form.querySelector('input[name="do_import"]').checked ? "1" : ""
    };

    runBtn.disabled = true;
    runBtn.textContent = "Running…";

    append(`=== RUN MONTH-BY-MONTH (${mode}: ${from} → ${to}) ===\n`);
    append(`Each month is fetched in 10-day chunks and raw files are deleted after import.\n`);
    append(`Total months: ${total}\n\n`);

    setProgress(0, total, "Starting…");

    for (let i=0; i<chunks.length; i++){
      const [ms, me] = chunks[i];
      setProgress(i, total, `Running month ${i+1}/${total} (${ms} → ${me})`);

      append(`\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n`);
      append(`MONTH ${i+1}/${total} (${ms} → ${me})\n`);
      append(`━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n`);

      const json = await postMonth({ from: ms, to: me, mode, ...steps });

      if (!json.ok){
        append(`\n❌ ERROR: ${json.error}\n`);
        if (json.output) append("\n" + json.output + "\n");
        setProgress(i, total, "Stopped with error");
        break;
      }

      append("\n" + (json.output || "") + "\n");
      setProgress(i+1, total, `Completed ${i+1}/${total} months`);
    }

    setProgress(total, total, "All done ✅");
    append(`\n=== DONE ===\n`);

    runBtn.disabled = false;
    runBtn.textContent = "Run Pipeline";
  });
})();
</script>

</body>
</html>