gnu-social/classes/File.php

<?php
/*
 * StatusNet - the distributed open-source microblogging tool
 * Copyright (C) 2008, 2009, StatusNet, Inc.
 *
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Affero General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.     See the
 * GNU Affero General Public License for more details.
 *
 * You should have received a copy of the GNU Affero General Public License
 * along with this program.     If not, see <http://www.gnu.org/licenses/>.
 */

if (!defined('GNUSOCIAL')) { exit(1); }

/**
 * Table Definition for file
 */
class File extends Managed_DataObject
{
    ###START_AUTOCODE
    /* the code below is auto generated do not remove the above tag */

    public $__table = 'file';                            // table name
    public $id;                              // int(4)  primary_key not_null
    public $url;                             // varchar(255)  unique_key
    public $mimetype;                        // varchar(50)
    public $size;                            // int(4)
    public $title;                           // varchar(255)
    public $date;                            // int(4)
    public $protected;                       // int(4)
    public $filename;                        // varchar(255)
    public $modified;                        // timestamp()   not_null default_CURRENT_TIMESTAMP

    /* the code above is auto generated do not remove the tag below */
    ###END_AUTOCODE

    public static function schemaDef()
    {
        return array(
            'fields' => array(
                'id' => array('type' => 'serial', 'not null' => true),
                'url' => array('type' => 'varchar', 'length' => 255, 'description' => 'destination URL after following redirections'),
                'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'),
                'size' => array('type' => 'int', 'description' => 'size of resource when available'),
                'title' => array('type' => 'varchar', 'length' => 255, 'description' => 'title of resource when available'),
                'date' => array('type' => 'int', 'description' => 'date of resource according to http query'),
                'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'),
                'filename' => array('type' => 'varchar', 'length' => 255, 'description' => 'if a local file, name of the file'),

                'modified' => array('type' => 'timestamp', 'not null' => true, 'description' => 'date this record was modified'),
            ),
            'primary key' => array('id'),
            'unique keys' => array(
                'file_url_key' => array('url'),
            ),
        );
    }

    function isProtected($url) {
        return 'http://www.facebook.com/login.php' === $url;
    }

    /**
     * Save a new file record.
     *
     * @param array $redir_data lookup data eg from File_redirection::where()
     * @param string $given_url
     * @return File
     */
    public static function saveNew(array $redir_data, $given_url) {

        // I don't know why we have to keep doing this but I'm adding this last check to avoid
        // uniqueness bugs.

        $x = File::getKV('url', $given_url);
        
        if (!$x instanceof File) {
            $x = new File;
            $x->url = $given_url;
            if (!empty($redir_data['protected'])) $x->protected = $redir_data['protected'];
            if (!empty($redir_data['title'])) $x->title = $redir_data['title'];
            if (!empty($redir_data['type'])) $x->mimetype = $redir_data['type'];
            if (!empty($redir_data['size'])) $x->size = intval($redir_data['size']);
            if (isset($redir_data['time']) && $redir_data['time'] > 0) $x->date = intval($redir_data['time']);
            $file_id = $x->insert();
        }

        $x->saveOembed($redir_data, $given_url);
        return $x;
    }

    /**
     * Save embedding information for this file, if applicable.
     *
     * Normally this won't need to be called manually, as File::saveNew()
     * takes care of it.
     *
     * @param array $redir_data lookup data eg from File_redirection::where()
     * @param string $given_url
     * @return boolean success
     */
    public function saveOembed(array $redir_data, $given_url)
    {
        if (isset($redir_data['type'])
                && (('text/html' === substr($redir_data['type'], 0, 9)
                || 'application/xhtml+xml' === substr($redir_data['type'], 0, 21)))) {
            try {
                $oembed_data = File_oembed::_getOembed($given_url);
            } catch (Exception $e) {
                return false;
            }
            if ($oembed_data === false) {
                return false;
            }
            $fo = File_oembed::getKV('file_id', $this->id);

            if ($fo instanceof File_oembed) {
                common_log(LOG_WARNING, "Strangely, a File_oembed object exists for new file $file_id", __FILE__);
            } else {
                File_oembed::saveNew($oembed_data, $this->id);
                return true;
            }
        }
        return false;
    }

    /**
     * Go look at a URL and possibly save data about it if it's new:
     * - follow redirect chains and store them in file_redirection
     * - look up oEmbed data and save it in file_oembed
     * - if a thumbnail is available, save it in file_thumbnail
     * - save file record with basic info
     * - optionally save a file_to_post record
     * - return the File object with the full reference
     *
     * @fixme refactor this mess, it's gotten pretty scary.
     * @param string $given_url the URL we're looking at
     * @param int $notice_id (optional)
     * @param bool $followRedirects defaults to true
     *
     * @return mixed File on success, -1 on some errors
     *
     * @throws ServerException on some errors
     */
    public function processNew($given_url, $notice_id=null, $followRedirects=true) {
        if (empty($given_url)) return -1;   // error, no url to process
        $given_url = File_redirection::_canonUrl($given_url);
        if (empty($given_url)) return -1;   // error, no url to process
        $file = File::getKV('url', $given_url);
        if (empty($file)) {
            $file_redir = File_redirection::getKV('url', $given_url);
            if (empty($file_redir)) {
                // @fixme for new URLs this also looks up non-redirect data
                // such as target content type, size, etc, which we need
                // for File::saveNew(); so we call it even if not following
                // new redirects.
                $redir_data = File_redirection::where($given_url);
                if (is_array($redir_data)) {
                    $redir_url = $redir_data['url'];
                } elseif (is_string($redir_data)) {
                    $redir_url = $redir_data;
                    $redir_data = array();
                } else {
                    // TRANS: Server exception thrown when a URL cannot be processed.
                    throw new ServerException(sprintf(_("Cannot process URL '%s'"), $given_url));
                }
                // TODO: max field length
                if ($redir_url === $given_url || strlen($redir_url) > 255 || !$followRedirects) {
                    $x = File::saveNew($redir_data, $given_url);
                    $file_id = $x->id;
                } else {
                    // This seems kind of messed up... for now skipping this part
                    // if we're already under a redirect, so we don't go into
                    // horrible infinite loops if we've been given an unstable
                    // redirect (where the final destination of the first request
                    // doesn't match what we get when we ask for it again).
                    //
                    // Seen in the wild with clojure.org, which redirects through
                    // wikispaces for auth and appends session data in the URL params.
                    $x = File::processNew($redir_url, $notice_id, /*followRedirects*/false);
                    $file_id = $x->id;
                    File_redirection::saveNew($redir_data, $file_id, $given_url);
                }
            } else {
                $file_id = $file_redir->file_id;
            }
        } else {
            $file_id = $file->id;
            $x = $file;
        }

        if (empty($x)) {
            $x = File::getKV('id', $file_id);
            if (empty($x)) {
                // @todo FIXME: This could possibly be a clearer message :)
                // TRANS: Server exception thrown when... Robin thinks something is impossible!
                throw new ServerException(_('Robin thinks something is impossible.'));
            }
        }

        if (!empty($notice_id)) {
            File_to_post::processNew($file_id, $notice_id);
        }
        return $x;
    }

    public static function respectsQuota(Profile $scoped, $fileSize) {
        if ($fileSize > common_config('attachments', 'file_quota')) {
            // TRANS: Message used to be inserted as %2$s in  the text "No file may
            // TRANS: be larger than %1$d byte and the file you sent was %2$s.".
            // TRANS: %1$d is the number of bytes of an uploaded file.
            $fileSizeText = sprintf(_m('%1$d byte','%1$d bytes',$fileSize),$fileSize);

            $fileQuota = common_config('attachments', 'file_quota');
            // TRANS: Message given if an upload is larger than the configured maximum.
            // TRANS: %1$d (used for plural) is the byte limit for uploads,
            // TRANS: %2$s is the proper form of "n bytes". This is the only ways to have
            // TRANS: gettext support multiple plurals in the same message, unfortunately...
            throw new ClientException(
                    sprintf(_m('No file may be larger than %1$d byte and the file you sent was %2$s. Try to upload a smaller version.',
                              'No file may be larger than %1$d bytes and the file you sent was %2$s. Try to upload a smaller version.',
                              $fileQuota),
                    $fileQuota, $fileSizeText));
        }

        $file = new File;

        $query = "select sum(size) as total from file join file_to_post on file_to_post.file_id = file.id join notice on file_to_post.post_id = notice.id where profile_id = {$scoped->id} and file.url like '%/notice/%/file'";
        $file->query($query);
        $file->fetch();
        $total = $file->total + $fileSize;
        if ($total > common_config('attachments', 'user_quota')) {
            // TRANS: Message given if an upload would exceed user quota.
            // TRANS: %d (number) is the user quota in bytes and is used for plural.
            throw new ClientException(
                    sprintf(_m('A file this large would exceed your user quota of %d byte.',
                              'A file this large would exceed your user quota of %d bytes.',
                              common_config('attachments', 'user_quota')),
                    common_config('attachments', 'user_quota')));
        }
        $query .= ' AND EXTRACT(month FROM file.modified) = EXTRACT(month FROM now()) and EXTRACT(year FROM file.modified) = EXTRACT(year FROM now())';
        $file->query($query);
        $file->fetch();
        $total = $file->total + $fileSize;
        if ($total > common_config('attachments', 'monthly_quota')) {
            // TRANS: Message given id an upload would exceed a user's monthly quota.
            // TRANS: $d (number) is the monthly user quota in bytes and is used for plural.
            throw new ClientException(
                    sprintf(_m('A file this large would exceed your monthly quota of %d byte.',
                              'A file this large would exceed your monthly quota of %d bytes.',
                              common_config('attachments', 'monthly_quota')),
                    common_config('attachments', 'monthly_quota')));
        }
        return true;
    }

    // where should the file go?

    static function filename(Profile $profile, $origname, $mimetype)
    {
        try {
            $ext = common_supported_mime_to_ext($mimetype);
        } catch (Exception $e) {
            // We don't support this mimetype, but let's guess the extension
            $ext = substr(strrchr($mimetype, '/'), 1);
        }

        // Normalize and make the original filename more URL friendly.
        $origname = basename($origname);
        if (class_exists('Normalizer')) {
            // http://php.net/manual/en/class.normalizer.php
            // http://www.unicode.org/reports/tr15/
            $origname = Normalizer::normalize($origname, Normalizer::FORM_KC);
        }
        $origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname);

        $nickname = $profile->nickname;
        $datestamp = strftime('%Y%m%d', time());
        do {
            // generate new random strings until we don't run into a filename collision.
            $random = strtolower(common_confirmation_code(16));
            $filename = "$nickname-$datestamp-$origname-$random.$ext";
        } while (file_exists(self::path($filename)));
        return $filename;
    }

    /**
     * Validation for as-saved base filenames
     */
    static function validFilename($filename)
    {
        return preg_match('/^[A-Za-z0-9._-]+$/', $filename);
    }

    /**
     * @throws ClientException on invalid filename
     */
    static function path($filename)
    {
        if (!self::validFilename($filename)) {
            // TRANS: Client exception thrown if a file upload does not have a valid name.
            throw new ClientException(_("Invalid filename."));
        }
        $dir = common_config('attachments', 'dir');

        if ($dir[strlen($dir)-1] != '/') {
            $dir .= '/';
        }

        return $dir . $filename;
    }

    static function url($filename)
    {
        if (!self::validFilename($filename)) {
            // TRANS: Client exception thrown if a file upload does not have a valid name.
            throw new ClientException(_("Invalid filename."));
        }

        if (common_config('site','private')) {

            return common_local_url('getfile',
                                array('filename' => $filename));

        }

        if (StatusNet::isHTTPS()) {

            $sslserver = common_config('attachments', 'sslserver');

            if (empty($sslserver)) {
                // XXX: this assumes that background dir == site dir + /file/
                // not true if there's another server
                if (is_string(common_config('site', 'sslserver')) &&
                    mb_strlen(common_config('site', 'sslserver')) > 0) {
                    $server = common_config('site', 'sslserver');
                } else if (common_config('site', 'server')) {
                    $server = common_config('site', 'server');
                }
                $path = common_config('site', 'path') . '/file/';
            } else {
                $server = $sslserver;
                $path   = common_config('attachments', 'sslpath');
                if (empty($path)) {
                    $path = common_config('attachments', 'path');
                }
            }

            $protocol = 'https';
        } else {
            $path = common_config('attachments', 'path');
            $server = common_config('attachments', 'server');

            if (empty($server)) {
                $server = common_config('site', 'server');
            }

            $ssl = common_config('attachments', 'ssl');

            $protocol = ($ssl) ? 'https' : 'http';
        }

        if ($path[strlen($path)-1] != '/') {
            $path .= '/';
        }

        if ($path[0] != '/') {
            $path = '/'.$path;
        }

        return $protocol.'://'.$server.$path.$filename;
    }

    function getEnclosure(){
        $enclosure = (object) array();
        $enclosure->title=$this->title;
        $enclosure->url=$this->url;
        $enclosure->title=$this->title;
        $enclosure->date=$this->date;
        $enclosure->modified=$this->modified;
        $enclosure->size=$this->size;
        $enclosure->mimetype=$this->mimetype;

        if(! isset($this->filename)){
            $notEnclosureMimeTypes = array(null,'text/html','application/xhtml+xml');
            $mimetype = $this->mimetype;
            if($mimetype != null){
                $mimetype = strtolower($this->mimetype);
            }
            $semicolon = strpos($mimetype,';');
            if($semicolon){
                $mimetype = substr($mimetype,0,$semicolon);
            }
            if(in_array($mimetype,$notEnclosureMimeTypes)){
                // Never treat generic HTML links as an enclosure type!
                // But if we have oEmbed info, we'll consider it golden.
                $oembed = File_oembed::getKV('file_id',$this->id);
                if($oembed && in_array($oembed->type, array('photo', 'video'))){
                    $mimetype = strtolower($oembed->mimetype);
                    $semicolon = strpos($mimetype,';');
                    if($semicolon){
                        $mimetype = substr($mimetype,0,$semicolon);
                    }
                    // @fixme uncertain if this is right.
                    // we want to expose things like YouTube videos as
                    // viewable attachments, but don't expose them as
                    // downloadable enclosures.....?
                    //if (in_array($mimetype, $notEnclosureMimeTypes)) {
                    //    return false;
                    //} else {
                        if($oembed->mimetype) $enclosure->mimetype=$oembed->mimetype;
                        if($oembed->url) $enclosure->url=$oembed->url;
                        if($oembed->title) $enclosure->title=$oembed->title;
                        if($oembed->modified) $enclosure->modified=$oembed->modified;
                        unset($oembed->size);
                    //}
                } else {
                    return false;
                }
            }
        }
        return $enclosure;
    }

    // quick back-compat hack, since there's still code using this
    function isEnclosure()
    {
        $enclosure = $this->getEnclosure();
        return !empty($enclosure);
    }

    /**
     * Get the attachment's thumbnail record, if any.
     *
     * @return File_thumbnail
     */
    function getThumbnail()
    {
        return File_thumbnail::getKV('file_id', $this->id);
    }

    public function getPath()
    {
        return self::path($this->filename);
    }
    public function getUrl()
    {
        return $this->url;
    }

    /**
     * Blow the cache of notices that link to this URL
     *
     * @param boolean $last Whether to blow the "last" cache too
     *
     * @return void
     */

    function blowCache($last=false)
    {
        self::blow('file:notice-ids:%s', $this->url);
        if ($last) {
            self::blow('file:notice-ids:%s;last', $this->url);
        }
        self::blow('file:notice-count:%d', $this->id);
    }

    /**
     * Stream of notices linking to this URL
     *
     * @param integer $offset   Offset to show; default is 0
     * @param integer $limit    Limit of notices to show
     * @param integer $since_id Since this notice
     * @param integer $max_id   Before this notice
     *
     * @return array ids of notices that link to this file
     */

    function stream($offset=0, $limit=NOTICES_PER_PAGE, $since_id=0, $max_id=0)
    {
        $stream = new FileNoticeStream($this);
        return $stream->getNotices($offset, $limit, $since_id, $max_id);
    }

    function noticeCount()
    {
        $cacheKey = sprintf('file:notice-count:%d', $this->id);
        
        $count = self::cacheGet($cacheKey);

        if ($count === false) {

            $f2p = new File_to_post();

            $f2p->file_id = $this->id;

            $count = $f2p->count();

            self::cacheSet($cacheKey, $count);
        } 

        return $count;
    }
}