crm-ms1/v4_ci4/app/Libraries/Radar/GmailReader.php

<?php

namespace App\Libraries\Radar;

/**
 * Lecture Gmail (readonly) pour une mailbox impersonnée.
 */
class GmailReader
{
    public function __construct(
        private GoogleWorkspaceToken $tokenClient,
        private string $scopes,
    ) {
    }

    /**
     * @return list<array{
     *   gmail_thread_id:string,
     *   mailbox:string,
     *   subject:string,
     *   from_email:string,
     *   from_name:string,
     *   snippet:string,
     *   gmail_url:string,
     *   received_at:?string
     * }>
     */
    public function listRecentThreads(string $mailbox, int $lookbackDays, int $maxMessages): array
    {
        $access = $this->tokenClient->getAccessToken($mailbox, $this->scopes);
        $q      = 'newer_than:' . max(1, $lookbackDays) . 'd';
        $url    = 'https://gmail.googleapis.com/gmail/v1/users/me/messages?'
            . \http_build_query([
                'q'          => $q,
                'maxResults' => $maxMessages,
            ]);

        [$code, $body] = $this->tokenClient->http(
            'GET',
            $url,
            ['Authorization: Bearer ' . $access]
        );
        if ($code !== 200) {
            throw new \RuntimeException("Gmail messages.list ({$code}): {$body}");
        }

        $data = \json_decode($body, true) ?: [];
        $ids  = $data['messages'] ?? [];
        if ($ids === []) {
            return [];
        }

        $byThread = [];
        foreach ($ids as $m) {
            $msgId = $m['id'] ?? null;
            if (! $msgId) {
                continue;
            }
            $detail = $this->getMessageMeta($access, $msgId);
            if ($detail === null) {
                continue;
            }
            $tid = $detail['gmail_thread_id'];
            // Garder le message le plus récent du fil dans ce batch
            if (! isset($byThread[$tid]) || ($detail['internal_ts'] ?? 0) > ($byThread[$tid]['internal_ts'] ?? 0)) {
                $detail['mailbox'] = $mailbox;
                $byThread[$tid]    = $detail;
            }
        }

        $out = array_values($byThread);
        usort($out, static fn ($a, $b) => ($b['internal_ts'] ?? 0) <=> ($a['internal_ts'] ?? 0));

        foreach ($out as &$row) {
            unset($row['internal_ts']);
        }
        unset($row);

        return $out;
    }

    /**
     * @return ?array<string, mixed>
     */
    private function getMessageMeta(string $access, string $messageId): ?array
    {
        $url = 'https://gmail.googleapis.com/gmail/v1/users/me/messages/' . \rawurlencode($messageId)
            . '?format=metadata&metadataHeaders=From&metadataHeaders=Subject&metadataHeaders=Date';

        [$code, $body] = $this->tokenClient->http(
            'GET',
            $url,
            ['Authorization: Bearer ' . $access]
        );
        if ($code !== 200) {
            return null;
        }

        $msg = \json_decode($body, true);
        if (! \is_array($msg)) {
            return null;
        }

        $headers = [];
        foreach ($msg['payload']['headers'] ?? [] as $h) {
            $headers[\strtolower($h['name'] ?? '')] = $h['value'] ?? '';
        }

        $from = $this->parseFrom($headers['from'] ?? '');
        $ts   = isset($msg['internalDate']) ? (int) \floor(((int) $msg['internalDate']) / 1000) : 0;

        return [
            'gmail_message_id' => $messageId,
            'gmail_thread_id'  => (string) ($msg['threadId'] ?? $messageId),
            'subject'          => self::decodeText((string) ($headers['subject'] ?? '(sans sujet)')),
            'from_email'       => $from['email'],
            'from_name'        => self::decodeText($from['name']),
            'snippet'          => self::decodeText((string) ($msg['snippet'] ?? '')),
            'gmail_url'        => 'https://mail.google.com/mail/u/0/#all/' . \rawurlencode((string) ($msg['threadId'] ?? $messageId)),
            'received_at'      => $ts > 0 ? \date('Y-m-d H:i:s', $ts) : null,
            'internal_ts'      => $ts,
        ];
    }

    /**
     * Corps texte du message (pour analyse IA). format=full.
     */
    public function getPlainBody(string $mailbox, string $messageId): string
    {
        $access = $this->tokenClient->getAccessToken($mailbox, $this->scopes);
        $url    = 'https://gmail.googleapis.com/gmail/v1/users/me/messages/' . \rawurlencode($messageId)
            . '?format=full';

        [$code, $body] = $this->tokenClient->http(
            'GET',
            $url,
            ['Authorization: Bearer ' . $access]
        );
        if ($code !== 200) {
            throw new \RuntimeException("Gmail messages.get ({$code}): {$body}");
        }

        $msg = \json_decode($body, true);
        if (! \is_array($msg)) {
            return '';
        }

        $text = $this->extractPlainFromPayload($msg['payload'] ?? []);
        if ($text === '' && ! empty($msg['snippet'])) {
            $text = (string) $msg['snippet'];
        }

        return self::decodeText(\mb_substr($text, 0, 12000));
    }

    /**
     * @param array<string, mixed> $payload
     */
    private function extractPlainFromPayload(array $payload): string
    {
        $mime = \strtolower((string) ($payload['mimeType'] ?? ''));
        if ($mime === 'text/plain' && ! empty($payload['body']['data'])) {
            return $this->b64urlDecode((string) $payload['body']['data']);
        }

        foreach ($payload['parts'] ?? [] as $part) {
            if (! \is_array($part)) {
                continue;
            }
            $found = $this->extractPlainFromPayload($part);
            if ($found !== '') {
                return $found;
            }
        }

        if ($mime === 'text/html' && ! empty($payload['body']['data'])) {
            $html = $this->b64urlDecode((string) $payload['body']['data']);

            return \trim(\html_entity_decode(\strip_tags($html), \ENT_QUOTES | \ENT_HTML5, 'UTF-8'));
        }

        return '';
    }

    private function b64urlDecode(string $data): string
    {
        $raw = \strtr($data, '-_', '+/');
        $pad = \strlen($raw) % 4;
        if ($pad > 0) {
            $raw .= \str_repeat('=', 4 - $pad);
        }
        $out = \base64_decode($raw, true);

        return $out === false ? '' : $out;
    }

    /** Gmail renvoie souvent des entités HTML dans snippet / headers. */
    public static function decodeText(string $text): string
    {
        if ($text === '') {
            return '';
        }

        return \html_entity_decode($text, \ENT_QUOTES | \ENT_HTML5, 'UTF-8');
    }

    /**
     * @return array{email:string,name:string}
     */
    private function parseFrom(string $from): array
    {
        if (\preg_match('/^(.*)<([^>]+)>$/', \trim($from), $m)) {
            return [
                'name'  => \trim($m[1], " \t\"'"),
                'email' => \strtolower(\trim($m[2])),
            ];
        }

        $email = \strtolower(\trim($from));

        return ['email' => $email, 'name' => ''];
    }
}