export using httrack et.al.

We could export our G1 albums using httrack/wget. This does not seem to be possible in G2 - will this be possible in the production release? Will there be other tools?

----

Gallery version: 2Beta1
Webserver apache Mac Apache/1.3.33 (Darwin) configured
Datatabase (with version): mySql
PHP version (eg 4.2.1):
phpinfo URL (optional):
Graphics Toolkit(s):
Operating system: darwin
Web browser/version: firefox+safari
G1 version (for migration bugs): n/a

probably JoergSchulz just wants to do a wget export of publically visible albums/images... shouldn't this be possible?

if you'd like to help us implement this please do the following:
1) try it.. export your G2 using wget or similar and test it out.. find out what links don't work (for example, login won't work too well...)
2) file a feature request on sourceforge; include the information you find in your tests, as this will help the person who implements the feature for G2.

offline mode? and how would that help only fetching a part of your albums? you know that you can tell wget which links to follow, and which not. wget is very powerful. don't think you need an offline mode. maybe you better explain what you mean by it.

@second G2 instance: you can have a second G2 instance with its own g2data directory and its own database tables (its own database or its own database prefix), but using the same gallery2 application files. just run the installer again and choose "multisite".

right, the G1 "offline mode" disables certain links so that the pages wget retrieves are nicely browsable.. based on the results of your tests we can decide if G2 needs such a mode.. for example, "offline mode" in G2 could using GalleryCapabilities to remove login links. Also possibly disable breadcrumb links for higher level albums to enable retrieving only a subtree of the gallery..

<OT> Ok; my mac apache setup is flawy, no 2nd instance available. No problem - this will be solved later. Lets turn to linux; g2setup worked flawlessly again as expected. Compliments. My mini G2 has 3 albums (2 main, 1sub) with 5 images</OT>

HTTRACK:
httrack http://regentag/gallery2/main.php/v/Barcelona/
starts to copy ALL photos, not only the selected sub-album.
In the process, an endless loop collects the non-desirable Action Items repeatedly in an endless loop. Bad. Had to interrupt that.
The resulting website copy looks surprisingly well - the pictures are completely available and reachable locally (well - more than I wanted: I only wanted one sub-album. We will come to the sub-album tests later).
The interactive links like "login, search, AddToCart, module:link to other directories, ViewCart, MemberList" point to - surprise - the original site.

Obviously, what we need is a set_offline mode that eliminates these buttons and that forbids recursive gets of other directories.

ad 1) fine..
ad 2) wget has the same problems; difference: wget starts with these URLs and never gbets to the photos.
ad 3) sounds like a bigger change: eache module should be supposed to support an offline mode (otherwise it should be excluded by a master-offline mode).

ad sourceforge: I add the above findings as comment to the existing request.

would this allow to keep the look-and-feel of G2?

i was thinking a particular user-agent string could have a little block in GallerySession that makes GalleryCapabilities calls to restrict what gets shown.. for example, disabling the sidebar and login links would remove many of the problematic things mentioned so far.

in wget, one can use the --no-parent option to prevent crawling upwards. This won't work for non-short-urls

Shoot. Um. Unless you can find something in the man pages for wget, I don't know.

nope: The images are not fetched; there are only links to the images in the original gallery.

ahem - not for me:

cmd Line:
wget --recursive --convert-links --include-directories='/gallery2/v/spaceoddities,/gallery2/d,/gallery2/themes,/gallery2/layouts,/gallery2/templates,/gallery2/images' http://regentag/gallery2/main.php/v/spaceoddities/

similar effect with httrack:
httrack -D http://regentag/gallery2/main.php/v/spaceoddities/
I get NO thumbnails (only the link to the original URLs instead), but the images in intermediary size, but not the images in full size. The navigation structure is included; the interactive links as well.

Obviously we are on a wrong track. We definitely need an option for g2 that allows for image downloads w/out the interactive links. the wget/httrack options alone don't help.

no - the whole gallery gets loaded then. And the mentioned problem with the recursive fetching of all interactive links reappears. No Way.

My gallery is currently only available in an intranet.
I Will try and get a grip on httrack/wget->G2 tomorrow night. The idea is to get back to the best option and possibility I had (that was: look-and-feel available, thumbnails available, fullsize pictures not) and find a possibility to add the urls to the midsize/fullsize images.
No - I won't only complain. I want to and will contribute to the great G2 project.
Lets do it the other way round: do you have a real-small test site I can connect to my wget/httrack? We can then see the differences between our results.

you can try this patch of modules/core/classes/GallerySession.class which removes the sidebar and login links when you access the gallery with wget:

diff -u -r1.65 GallerySession.class
--- modules/core/classes/GallerySession.class   3 Apr 2005 08:42:05 -0000       1.65
+++ modules/core/classes/GallerySession.class   8 Apr 2005 20:50:08 -0000
@@ -177,7 +177,7 @@
        $forceResendCookie = false;

        /* Verify the remote address to avoid casual session hijacking */
-       $currentRemoteIdentifier = $this->_getRemoteIdentifier();
+       list ($currentRemoteIdentifier, $userAgent) = $this->_getRemoteIdentifier();

        if (!isset($this->_remoteIdentifier) ||
            $this->_compareIdentifiers($this->_remoteIdentifier,
@@ -205,6 +205,13 @@
            }
        }

+       if (!strncasecmp($userAgent, 'wget', 4)) {
+           /* Offline mode */
+           $this->_isUsingCookies = true;
+           GalleryCapabilities::set('login', false);
+           GalleryCapabilities::set('showSidebar', false);
+       }
+
        if (!isset($_COOKIE[SESSION_ID_PARAMETER]) ||
                $_COOKIE[SESSION_ID_PARAMETER] != $this->_sessionId ||
                $forceResendCookie) {
@@ -569,7 +576,7 @@
        $this->_sessionData = array();
        $this->_loadedSessionData = array();
        $this->_creationTime = time();
-       $this->_remoteIdentifier = $this->_getRemoteIdentifier();
+       list ($this->_remoteIdentifier) = $this->_getRemoteIdentifier();
     }

     /**
@@ -583,8 +590,9 @@
      */
     function _getRemoteIdentifier() {
        $httpUserAgent = GalleryUtilities::getServerVar('HTTP_USER_AGENT');
-       return array(GalleryUtilities::getRemoteHostAddress(),
-                    isset($httpUserAgent) ? md5($httpUserAgent) : null);
+       return array(array(GalleryUtilities::getRemoteHostAddress(),
+                          isset($httpUserAgent) ? md5($httpUserAgent) : null),
+                    $httpUserAgent);
     }

     /**

I will probably put this or something similar into cvs after beta-2 is released next week.

dunno if that should apply with 'patch' or not.. it's small enough i thought it would be a manual job.

Has any more progress been made here? I would really like to be able to quickly and easily make an offline copy of gallery2 albums.

Thanks,
-Jon