<?php
// tts_synthesize.php
// Place in public_html/tts_synthesize.php
// This script fetches MP3 from Google Translate TTS in chunks and returns combined MP3.
// NOTE: This uses an unofficial Google Translate endpoint for small/personal usage.

// --- Basic headers and CORS ---
header("Access-Control-Allow-Origin: *");
header("Access-Control-Allow-Methods: POST, OPTIONS");
header("Access-Control-Allow-Headers: Content-Type");

// Preflight
if ($_SERVER['REQUEST_METHOD'] === 'OPTIONS') {
    http_response_code(200);
    exit;
}

// disable PHP output buffering + errors to avoid corrupting MP3 output
ini_set('display_errors', '0');
error_reporting(0);

// read JSON
$raw = file_get_contents('php://input');
$input = json_decode($raw, true);
$text = isset($input['text']) ? trim($input['text']) : '';

if ($text === '') {
    http_response_code(400);
    header('Content-Type: text/plain; charset=utf-8');
    echo "No text provided";
    exit;
}

// logging helper
function log_error($msg) {
    $f = __DIR__ . '/tts_error.log';
    $now = date('Y-m-d H:i:s');
    @file_put_contents($f, "[$now] $msg\n", FILE_APPEND);
}

// split text into chunks (max ~180 characters) on word boundaries
function split_text_chunks($text, $max = 180) {
    $text = preg_replace('/\s+/', ' ', trim($text));
    $words = preg_split('/\s+/', $text);
    $chunks = [];
    $current = '';
    foreach ($words as $w) {
        // mb_strlen for Unicode
        $len = mb_strlen($current . ' ' . $w, 'UTF-8');
        if ($current === '') {
            $current = $w;
        } elseif ($len <= $max) {
            $current .= ' ' . $w;
        } else {
            $chunks[] = $current;
            $current = $w;
        }
    }
    if ($current !== '') $chunks[] = $current;
    return $chunks;
}

$chunks = split_text_chunks($text);
$totalAudio = ''; // binary concatenation (we'll build using PHP strings)

// cURL fetch function (returns binary string or false)
function fetch_tts_chunk($chunkText, $idx, $total) {
    // prepare URL - include idx and total to be safe
    $q = rawurlencode($chunkText);
    $url = "https://translate.google.com/translate_tts?ie=UTF-8&q={$q}&tl=hi&client=tw-ob&idx={$idx}&total={$total}";

    $ch = curl_init();
    curl_setopt($ch, CURLOPT_URL, $url);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
    // user agent is required
    curl_setopt($ch, CURLOPT_USERAGENT, "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko)");
    // follow redirects
    curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
    // timeout
    curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 10);
    curl_setopt($ch, CURLOPT_TIMEOUT, 20);
    // Accept audio
    curl_setopt($ch, CURLOPT_HTTPHEADER, [
        "Accept: audio/mpeg, */*"
    ]);
    $data = curl_exec($ch);
    $code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
    $err = curl_error($ch);
    curl_close($ch);

    if ($data === false || $code !== 200) {
        $errMsg = "TTS fetch failed (code={$code}) idx={$idx} err=" . $err;
        log_error($errMsg);
        return false;
    }
    return $data;
}

// iterate chunks and append
$totalChunks = count($chunks);
if ($totalChunks === 0) {
    http_response_code(400);
    header('Content-Type: text/plain; charset=utf-8');
    echo "No valid text chunks";
    exit;
}

foreach ($chunks as $i => $chunkText) {
    $idx = $i;
    $bin = fetch_tts_chunk($chunkText, $idx, $totalChunks);
    if ($bin === false) {
        http_response_code(502);
        header('Content-Type: text/plain; charset=utf-8');
        echo "Failed to fetch audio from TTS provider. Check tts_error.log on server.";
        exit;
    }
    // append binary data
    $totalAudio .= $bin;
}

// output combined MP3
// Clean any output buffer (to avoid corrupting audio)
while (ob_get_level()) ob_end_clean();

header('Content-Type: audio/mpeg');
header('Content-Disposition: attachment; filename="hindi-tts.mp3"');
echo $totalAudio;
exit;