============= TODO: curator ============= - JavaScript shortcuts for next and previous would be amazingly nice! - support IPTC caption standard instead of the desc file - bug: the photos are not in the order that the dirindex shows it - don't forget to add EXIF tags - bug: if only --browse is available, curator doesn't pick them up anymore (e.g. use package_pics and try just running curator in the given subdir) - curator: BIG BUG 'description' and 'title' don't come out in default template!!! - curator: update man page, answer bugs. - make sure to test new curator with animated gif, there is a bug:: It seems that when curator hits an animated gif, it creates a thumbnail for each frame of the gif and then hangs. Maybe it goes into some infinite loop? This is the one that I hit: http://members.rogers.com/rcrosbie/images/svganimated.gif - Is it possible to clean up the help output? It seems the formatting in the code messes up the help output. .. end :: Greg Baker I have recently put my gallery into the Subversion version control system. It leaves .svn directories everywhere; these were mucking up my gallery. Anyway, I fiddled the curator code to ignore those (they do have image files, so it wasn't covered by the existing code). This patch is from Curator 2.1 (Debian package 2.1-0.2): 697,700c697,703 < subdir = Dir( af, self, root ) < # ignore directories which do not have images under them. < if subdir.hasImages(): < self._subdirs.append( subdir ) --- > if string.find(paf, ".svn")==-1: > subdir = Dir( af, self, root ) > # ignore directories which do not have images under them. > if subdir.hasImages(): > self._subdirs.append( subdir ) > elif opts.quiet: print "ignoring subversion directory '%s'" % \ > ( af ) If that's interesting, feel free to use it. Other Project: photo organizer program to process images before curator ----------------------------------------------------------------------- Basically, something that would do all I need to do to unpack and present my pictures: - categorize pictures for directories - write descriptions and titles, use previous easily (offer in list) - assign locations for organizing into directories - tag for publishing, not nice shouldn't need be published - delete - rotate - set attributes for a bunch of pictures at the same time (w/ multi-select list) - automatically generate directory names that will sort automatically in curator. Also: other tool that can be used to select photos in a file hierarchy from particular attributes and tar them up with the description files * we must make the new curator able to deal with AVI or MPG files nicely. Reply sent to Fernando Perez who contributed a good summary of the work to be done on curator:: On Thursday 25 September 2003 03:18, Fernando Perez wrote: > Hi Martin, > > I have a few comments/suggestions concerning your great curator program. > First, I've included here a patch so that option information is properly > formatted. I suspect when you switched to optparse you didn't reformat > your option help strings, so I did it in the least intrusive way I could. > I also made 2.3 a requirement, since optparse effectively makes it so. > Finally, I removed the distutils imports you were using for option parsing, > since you seem to have fully gone to optparse now and they aren't needed > anymore. yes 2.3 is now a requirement. i didn't fully port it yet, i just needed to fix it for a friend who runs windows, i'm scraping for time to clean this up > > Those were my little contributions. Now come the 'requests' :) I've been > long looking for a program to replace a highly hacked gallery creator I > use, and curator _almost_ fits the bill. I really like some of the > concepts you implemented in curator, like the idea of tracks for images. > But there are a couple of things which prevent me from using it in full. > I'll mention them in case you find the time to implement them for a future > release, which would be great. I actually considered doing it myself, but > the logic of the program is delicate enough that it would take me more than > the limited time I could devote to this as a side hack. thanks for the patch, i thought optparse did paragraph filling automatically > The key issue is not having an option for making the final gallery with > copies of the original images at a different size. With 4-6 megapixel > images as originals, it's not practical for the thumbnails to point to the > originals for web browsing. But I don't want to have to manually run a > resizing script over whole directories, especially if I want to generate > albums on CD at different sizes. With my hacked igal (image gallery) > script, typical usage is something like: > > igal -size 800 --no-originals .... # make an album with max.dim 600pixels > igal -size 1000 --no-originals .... # make an album with max.dim 1000pixels > > and I can build out of the same original images quickly two albums with > small/large resolution images (for uploading to a space-limited website) > but still smaller than the huge originals. If I don't pass the > --no-originals option, each medium-sized image becomes a link to the > full-size original (useful for burning on CD where space is less of an > issue). well, about this one, this must be the most requested feature, but everyone wants a different thing. so my attitude towards (and my own usage of it) is to leave that functionality to an external script. curator support multiple "representations" of the same image. i'd like to implement it at some point anyway, but this will only happen when i do the rewrite, and splitting it up in librarires and all, with specific objects for methods for image reduction generation, etc. (which you desribe a bit below) > These options make it very easy to produce, out of a set of very high > resolution orignals, standalone directories for different purposes. If you > want to play with this functionality to see if you'd like curator to have > it, feel free to grab my hacked igal from > http://windom.colorado.edu/~fperez/igal. I don't suggest you read the code > (it's heinous perl), just that it can illustrate useful functionality which > I find curator currently lacks. > Having that would definitely make (IMHO) curator a better tool for those > handling high-res images. My next comments are comparatively minor, just > consider them suggestions for the future: > > - split curator into a set of one or more modules and curator itself. Then > curator should simply read: > > #!/usr/bin/env python > import curator > curator.main() > > or something similar. curator/ would be a package and 'bin/curator' would > be the above 3 line script. The advantage of this is drastically reduced > startup times (besides code organization), because everything but the 3 > line startup script is precompiled. If you are not familiar with distutils > package organization, I'd be happy to help you with this breakup (if you > are interested, of course :) the original idea behind curator was driven by my frustration in not being able to find a simple program with little dependencies. imagemagick and python (now PIL and python), in a single file script was the best i found. i did not even intend to publish this thing originally, but i just created a SF project anyway, and now i find lots of people are using it... i need to rewrite this thing (especially now since i've gained a lot of experience with python since then) > - A package organization would also make it easier to handle the different > parts of the script as standalone units for a bit of an architectural > cleanup (less reliance on globals, for example). yes! > - Embedding text into the image itself. Useful for pasting copyright > notices into images for web display. ImageMagick can do this, I don't know > about PIL. again, this is often part of the image reduction phase, and everyone wants something different > - Have the ability to store all the per-image text info into a single file. > When captioning an entire gallery, it feels a bit clunky to have to create > a caption as a separate file for each image. It would be a lot easier to > keep emacs open and do something like: > > [img001.jpg] > text.... > [img002.jpg] > text... > > etc. yes, and i would probably move to a simple XML format parsing it with the effbot's elementtree. > In fact, igal can be asked to create such an empty captions file for you > automatically (with text... defaulting to the filename). In this way, one > can then just go into the file and fill in the the details for some images > with minimal work. > > Furthermore, igal will build the gallery with the images in the order they > appear in the captions file. This makes it extremely easy to reorder > images in each gallery so they appear in an order different than the > default one given by 'ls'. The same concept is applied to a gallery-wide > index, with which one can reorder and caption each subgallery. in the TODO file this is my next feature that i wanted to implement, the "tracks" should indeed act like slideshows and the order should be specifiable yes, it's not very useful otherwise. also, instead of generating the tracks in the same HTML page (not easy to browse), i would generate different sets of HTML pages for each track, so you can use the imagemap navigation > > Well, in summary: I really like curator and I think it has lots of > potential. A clean codebase, written in python, and some very interesting > ideas. But I feel it lacks some truly useful functionality which I've > grown to rely on from igal. On the other hand, I simply won't touch igal > any further, so I'd love to be able to move over to curator. If you find > any of these suggestions useful, great. If you end up implementing some of > them, even better :) these are great suggestions, thanks for taking the time to organize your thoughts, i will use this email to plan for the rewrite! > But I know that open source development happens when time allows, so please > consider this lengthy message just a friendly set of comments. I maintain > a reasonably large python program myself (http://ipython.scipy.org, which > you may find useful), so I know how it goes :) indeed, i have a zillion other projects that are written and i'm dying for time to finish them, i've got a travel site generator script to move to docutils input, an advanced jukebox music collection tools to publish and write a more flexible backend for, some tiny contractor tools to finish, like being able to generate my invoices with an XML file as input instead of fiddling aimlessly with gnumeric to do things right (http://www.furius.ca/contract-tools/) and xxdiff to maintain, and moving website, and building my company website, and i'm going crazy, i'm working by the hour now, so it's hard to dedicate time to oss, this might happen when i finish that contract. all the ideas you mention are cool. i'll check out your package, on my current assignment, i installed the alternative readline lib for python (on windows xp to make it feel more like at home on gentoo), i'll try your ipython instead later on today. cheers, * add an option to use the in the HEAD of the generated html to use paths all relative to the root instead of the current directory. Note that this is not a really good solution and should be optional only because it won't work when accessing the documents as file://, then again some people might prefer that kind of output. * super nice idea from fxt: make it able to create custom slide shows with a list of pictures in a certain order input a text file with the pictures in the desired order * bug: if only --browse is available, curator doesn't pick them up anymore (e.g. use package_pics and try just running curator in the given subdir) * curator --out-onedir picks up its own html output directory when run twice in a row. * update the Debian man file curator.sgml (fetch from the web). * something that is really annoying is that new directories are not stored in order of date in the directory index, find a solution for that. - for tracks, make curator able to use the basenames without the directories for the picture names BIG BUG! no-html is not the opposite of force-html, but they're implemented the same, do this! - implement lowsrc in images using thumbnails as a source (tdemarge) - write a script to copy the exif info from one jpeg file to another - write a script to extract thumbnails from jpegs which have them, this should make generating thumbnails much faster - include "source: " field for archival retrieval - instructions for making a CDROM under windows on web page #------------------------------------------------------------------------------- # long term # - table generator: align image to top of cells, not center (vertically) - global definitions file to be executed for each environment for all templates # merge getprop() and getattr() ? Eventually: - the images don't necessarily reside where the HTML pages are, and can be scattered/organized over many different directories; - implement user-specified conversins # add 'next directory' functionality to script env # running surrogate programs, grab errors and print to stderr of this. # e.g. convert program # add reporting line or error in template # try accessing Image class... it works... but it shouldn't what can we do about # it? # --urls mode # - only a global file, thumbnails generated locally # - dependency on wget # - look for desc file at URL location, then locally # support generating an HTML global index with a list of URLs away from output # dir. Should generate global index and image pages only. This would be a # special mode. # perhaps vars( class(envir) ) would solve the ugly hack for methods otherwise # we could bin by hand # allow placing thumbnails in an arbitrary directory. # you'll want to support adding a copyright to individual images, and other # types of conversions, processing # implement field to automatically make conversions (optional) the # code should be the same as thumbnails, the thumbnails is only a particular # case of this. Commands: for f in `find . -name \*.jpg ! -name \*thumb* `; do echo $f ; j=`echo $f | sed -e 's/jpg/desc/'`; echo "settings:" > $j ; ~/curator/exif995.py $f >> $j ; done ------------------------------ Currently the following files are produces: in root directory: allindex.html sortindex.html trackindes.html in every directory including root: dirindex.html image.html It is good practice to have an index.html in each directory, to give the web server a default file to display. True, one could change the web server default to "allindex.html, dirindex.html", but only if you run your own web server. With this in mind, it would be nice if there were an option to tell curator the following: - name the dirindex.html file in all none root directories 'index.html' - name the allindex.html file 'index.html' or name the sortindex.hml file 'index.hml' or name the trackindex.hml file 'index.hml' or name the dirindex.hml file in the root directory 'index.hml' References from the templates to the files should take this into account and point to the right file, of course. Maybe I am the only one who ran into this problem, maybe not. Hope you like the idea. Regards, Thomas ------------------------------ From: Jean-Luc Duprat To: Martin Blais Subject: RE: copyright notice > dis, avec quel logiciel es-tu ajout=E9 le copyright notice sur > tes photos? > c'est joli. Un petit shell script qui invoke ImageMagick: 1) resize photo 2) Composite the copyright notice 3) Add a copyright comment to the image the jpg. The last step is nice if you "file" the image. L'image que je composite sur la photo a ete cree avec gimp: 800x15, fill blanc avec 50% opacite, texte par dessus, saved as= png. L'avantage de faire un composite, plutot que la ligne= commentee du script, c'est que cela m'evite d'utiliser "gs" pour= dessiner le texte, ce qui est plus rapide -- et gs n'est pas= installe sur le serveur que j'utilise. De plus c'est un peu= plus rapide... JL --------- #!/bin/sh function process_image () { echo Processing $1... convert $1 -size 800x600 -geometry 800x600 +profile '*' $1.1 # convert $1.1 -fill gray98 -draw "rectangle 0,585 800,600"= -fill black -draw 'text 200,596 " Copyright (c) 2003 Jean-Luc= Duprat All rights reserved."' $1.2 composite -geometry +0+585= ~/src/adventures/share/atom/resources/copyright2003.png $1.1= $1.2 convert $1.2 -comment "Copyright 2003 Jean-Luc Duprat= All rights reserved." $1.3 if test -e $1.3; then rm $1 $1.1 $1.2 mv $1.3 $1 else rm $1.1 $1.2 fi } function extract_exif () { if test ! -e $f.desc; then ~/src/adventures/tools/exiftool -f= ~/src/adventures/etc/exif_template -i desc $1 fi } echo Extracting EXIF data... for f in *.jpg; do if test -e $f; then extract_exif $f fi done for f in *.JPG; do if test -e $f; then extract_exif $f fi done for f in *.jpg; do if test -e $f; then process_image $f fi done for f in *.JPG; do if test -e $f; then process_image $f fi done --46523270 Content-Type: image/png; name="copyright2003.png" Content-Transfer-Encoding: Base64 Content-Disposition: attachment; filename="copyright2003.png" iVBORw0KGgoAAAANSUhEUgAAAyAAAAAPCAYAAAAcYqsWAAAABmJLR0QA/wD/AP+gvaeTAAAACXBI WXMAAAsSAAALEgHS3X78AAAAB3RJTUUH0woVDBYVDVbTdQAADzJJREFUeJztnU1r28rXwI8f/msp IRC4oIAC3tRZZCER6uzawDUO8arxpsgE4kWgeOF8AKMR+QDVIhSysCFckY2TVUxECkl2cTCZRRZx NoYKLCgUQq/0BfIs7ONOpvJbXvpyOz8obSXNzNHMkXzOzDmj2N3dHQGBQCAQCAQCgUAg+AH8388W QCD4E6CUSrquZ2OxmBmLxcyJiYnN5eXlv3+2XAAAKBchJDFqGbyPx7ZNKZVisZiZz+dfPsV1CCEk QQhJUEqlqLrwfFRZx3EUQkjCcRxl3LoHyROLxUxd17OjlnlqfoYMrM6zf34V3X9qJiYmNmOxmDk7 O5vnz/HP2UOeu4eM4UPaEXR4qvecQCD4HuGACATPTNf52KSU9gyAIAgk13WTP9MgfU5GNZRWV1ez siyH5XL5YtB1mqaF6XS6XqlUUoMMf2zXsqysZVlZXdc3WRkIIQld1zfx/MTExCbraOi6ns3lcnnL srK5XC7PGpKEkMTExMRmv7oFo/Or6P5TOmWO4yhBEEgAAJ7nKeM4qD8bYWgLBIIfzf9+tgACwX+d 1dXVLACApmnNnZ2dY03TQsdxlEKhkKWUJroz6k28Hmcq4/F4aBiGj8cdx1FarZaUyWR8AIDDw0Ml k8n4mqaFAB1H5/DwUGHL8cewbkJI03Ec5cWLF+Hl5WU1Sm5CSCKqHH8dKxfKMgqO4yie5ynr6+vH 7HGUmW9va2vrwnXd5OrqavbTp09lvj5KqWRZVhYAYH19/XhmZia0bTtFKU3k8/mXr1698i3Lysqy HBaLxeN2uy1VKpVUoVDIGobxfnl5+W9KaULTtObKysp1rVabo5QmlpeX/z46Ovpo23YqCAIJ5T04 OEhSShOO4yjsOI0D9h0/1nw/8Of5cXxI/0f13zD9GUXuKEzTrOJYEkISlmVl2b6L0i/2WJQe821H XYP1obwAAI/tp37s7e3dW2EolUovj46OPj60Pv55v7m5kQghTUKI1e/aYfrQbzz7wa6aRD37g2TF +h+q4/x5tv1++sDfd1RfDNPdQe86gUDwdAgHRCB4Riilkud5CgAAa+gbhuGfnZ3Vr66uZvAYGmZs edu2m1jOtu0kpTRRq9WauJpiWRak0+n60dHRx5ubGwkNbMMw3gN0nBTLsrLpdLpuGIaP9Tcajbrr uknTNKsbGxtzlNIEGomUUmlpaSmPs7m2bffa442f5eXlv13XTXavC7e3t6utVqvnCFBKE7quZ6Oc nFKplAIAePfuXe+HPp/Pv6xUKin8v2VZPeNV07RQVVUfZ5d54woNFU3TmuyKimVZWbaf37x5U0fj 4vT0dM7zPMVxHOXm5mYGAACdxEwm4+u6njg/P58DgI/xeNyfn59vY91XV1czlNIEGrbjout6ll0V Y8c6ShdKpZKPjhc/jthX//zzT/mhzhDqiqZpTcMwqv2Ozc7O5lGnAQAKhUJ4cnJSHtWoJ4Q0Ue69 vb0Eq5esfrHHUA5VVf1CoSChbpZKJX9/f7+qaVrIylooFJQgCCRCiMX3s2VZPQd1FD0dla6egGma Vcuysqg3D60Pn3fsK03Tmq1W6xrvEWVl78+27XBycjLsjk9V07Tes9VoNBRWp87Ozo7L5fIFu/IR i8XMu7s7i38HYN39xjlKVsMwquPq+KDzbPtRY51Op+uu6yYbjUYdHb9SqZTyPE+Jx+NlTdPCQfLw 91yr1YQDIhA8IyIESyB4RlijmD9XLpcvLi8vqzhryc7em6ZZlWU5xBl4thylNLG+vn6MM/Gu6yZx JlmW5TAIAgnDimq12hwAwNu3b++1f35+PqdpWjMej39nTOCPsKZpTdM0q61WKzIXgq0HoBNWZtt2 MpPJ+Ol0ug4AoKqqXywW61FlPc9TZFkO2ZnJSqWSkmU5NE2zivdn23bPIXnx4kUbAODDhw/fxbNn MhnfNM0q21673ZYAAKanpwN0QmZmZnr3PDU1FQIAtFqtnqOI8uDfaJBcXl5Wy+XyheM4Sj6ff4mr JQ+ZKWVXW0zTrKqq6rNjHaULnucpfBw/2//dvkqOK8u4cnuep6DcmqY1gyCQSqXSSLk5yMLCgg8A 8OXLF3mccl1j0mf7hG+bUpqYnJwMu2OToJQmeJ2qVCqpUfV0FDD8SlVVnxDS5J/Dx4AG/fz8fJs/ h3qE9xePx33WOYyqR5blEKDTBwAdhwmvwX+XSqWX7DsAx3ljYyMVVXeUrMN0HJ9r/n2H+Vc4mWKa ZjWdTteDIJCWlpbu5dawY43vOJxIAPj2jjEMwx8mzzjvPYFA8HjECohA8AuA4RvpdLqOM+zxeDzM 5XL57gx3byaVvebz588SO5O8uLh47bpu8uzsTDEMw0fjhJ8VZ2cybdvuHUdDSpblXmhWdyVgM0pu rIddudA0LVxYWPBd14WpqamBYQ7xeLx3DvuAXaEA6Kw0oIOF9XadiXt5I5qmhawxTghJVCqVlKZp zaOjo49PlXNwdnamHBwc9Az9qNWYYeBsObfasomz5l1j8l74WxRR/Q/QmRVnr3vMzP4guQkhTV3X s+M6Eg+F1Ut8PviVBlVVeytF3TyMKoYGsY7rKHo6Kqi7r1+/vgYAwOcQn8uH1gvQMdDxeY9yQAEA tre3q11dafIrVHw9mJOGxwkhTcuyev9my9ze3koAnfE+PDy8jpqw6CfrxMTEJpaN0nEs8/nzZ2lm ZiY8OTkpY1gpOpXMu6A5Ozs7g6uVWJYda4DOihiukOLkz+Li4jXbV1HyOI7THOe9JxAIHo9wQASC ZwR/sKNm09hYZDTgcGYYoBOmlcvlvquTvQYNKCz/9u3bpuu6ydPT0znHcXyAbz/ALP0MZgwnYh2D QcY1nmNXFR4K3gNbV9eYGZigzkMplUql0kvXdZMYnvZY2VjK5fJFuVy+QKO/VCoFfBsY6sHmPrDg qgpv4ODxdrst2badyuVykqqqfQ3Yfv3Phpk8JSgfqxMPcW5wZWpcWL3E54MNEwL4tqqFsKFHUSuR TwGGwZ2ens7puj6Dhvtjw7AABj9beO+skzM1NRV6nte3nlGc5a2trQuAjvzdTRdAVVV/bW1t4CoR K+swHd/e3q7atp10XTfpui5YlgWapjUzmYyP74JKpZJiHWuAb+8ovFf23OvXr68rlYpyeHio8Ku/ g+QZ970nEAgej3BABIJnxDAMv1AohN149HvJ5oVCIRsEgWSaZnV6ejoA6BhLANAE6M3efgdrvPGG HLbneZ6Cs7J8+NUosA7Tj97Nh72nfD7/8urqagZnLYeVxThuAIDLy8v3bJn5+fl2N06918d4n/F4 vJdfgqstONMqy3LoOI6Sy+Xy/IwrwPhhRCxs+AuCoWiqqvr//vvve4Dv80UeUu+ooPEM0N9RYFd9 dF3PTk9Pf+eE9YNSKuEK0srKyjV/TtO0sJ/OsbKNopfoiLIz80+92xO7KuF5nsIa/xiG9dhVkGGw bbB99FAODw+VhYUF/+jo6COlVPrw4UOiUqmkLMvKRiXBDyJKFymlUqvVkorFYt0wjKrjOArmkWxs bPSuS6fTdXbCBaCzMoGrGzzv3r1rViqVVK1Wm2u1WkrU6m+UPFETRb/TLmYCwe+IcEAEgmfmzZs3 dfzx3t3d9aempsJWq9ULdermgITdmcBkPp8PcQcngM6PMFvfwcFBEhO3oww5bM913WTUD/AgCCFN 27bDIAgkXdezKysr12wOxrjc3t5GGmAY9sH+4BeLxXoul0uw4U2YE4LGbtd5gKh4eIBvcdyqqvps vPr8/Hz71atXfqVSwS1g5dvbWwnHwDAMf29vr+15nlIoFLK2bfso2+Li4jXr2M3OzuanpqZCdAii ZEFnx7btnjEE0MlFAeiEjniep9RqtbmVlZXrRqOhuK6bZGeZv379KhFCEu12Wxp3RWNQXkqr1VL4 EK35+fk2zl5H3SOCib6rq6vZtbW1Ou4Uxusoz+7ubhJnpHndZ69bWlrKx+Nxv1/8PebBEEKaOL6D VjXQOby6upohhIQoAw+rp+j0do3joc8O1skby9g3TxGG1Q8cj0KhkG21Wse1Wm2uXw7IKGDf7u7u Jj3PUxqNxr17wvyRURik45lMpoo5HmdnZ3V25WR6ejr466+/Qkpp4vz8fG5hYcFvt9vSwcFBMggC KZPJvO/XJm5UgXrL6uUgeT59+lTGiaJB7z18bp4qpFEg+JMRDohA8MyUy+ULzNVgZ0hVVfX39/er AJ2Vi1arVbUsK8uGHGD+Al8nG0bAJ0LjLCBAxxkZV95isXiM26Ri0ua4BjDOKHqep9i2ncQdlFj4 Ha26TkDddd0kyo9b5mIZTDBld85CCCEJ9jsMfBhKuVy+wD7G+5FlOTw5OSkDAHTzRGS8b4D7/V8s Fo9t206xdfM7brFt4S5ZWJeqqj6Gtuzv71eXlpby/Hnc0QmdQDTSsK/Y1ZuHEgRBpENTLperuMMa 3iMauHjN1tbWxc3NzYzneb2wJva++sGPB9vvAN92jkLZ+HbZchgShP/f2dk55q9D0KlldRnPOY6j ROkps7vZd6GLUWBfbm1tXbArbt0clcRThGH1Y2tr6+L8/HwuSlfGqQeT5nGFY21trW7bdgrDoxD2 WRzGMB3HMWbfd7Ish9iPOA6oZ5iQ3t0Fq2+7GIYFcH/1d5A8eG+4bXe/995zhTYKBH8isbu7O/Kz hRAI/hT6feMj6hp+H382r6DRaChfvnyR2W1hWTARlQ9DitrjPmqvfH6ffQxbubu7s6LqGecbEgjm ULDhMWw5Xk5Mno0Kg+LL8UTJxdfP98ew7wSM8j2JYXUNOj/oewSj9D/PqP0zyndFRv0OCPv9jai2 ouTr1xe47Sp+o2V6ejpgjf5+5Yf1DX9+nO+qDNOlQd9rGdbOKM9lVDt8/lFUPVG6NehYv/sbJCt/ 7jm/AzJKmVHk6dcPo5wTCATjIRwQgeA3YVhiM4J5EJOTk2GUoT4MdDbwmxL4rQ9ZlkPMSXgqZmdn 81+/fpVGqRfl4J0qwX8f1gER4S/fwHcCbraAuUoA8KhvwggEAsFzI74DIhD8hyCEJHRd3wyCQBq2 Y00/MEkzl8vlY7GYic4HGzLzVOzv71eDIJDy+fzA70hQSiVMJhbOh0DQYWdn51iW5dB13WQsFjPR +TBNsyqcD4FA8CsjVkAEgt+EUUJDMPxglPCRYYwSLiYQ/AhGCTH7k2FD3UR4kEAg+B34f2skgcnp F36hAAAAAElFTkSuQmCCDQotLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tLTdkMzMxMmMyMDE1 Ng0KQ29udGVudC1EaXNwb3NpdGlvbjogZm9ybS1kYXRhOyBuYW1lPSJ4Ig0KDQo0MA0KLS0tLS0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tLS03ZDMzMTJjMjAxNTYNCkNvbnRlbnQtRGlzcG9zaXRpb246 IGZvcm0tZGF0YTsgbmFtZT0ieSINCg0KMTg= --46523270--