pymeds/README

This code is based upon code by Tim Potter.

It is licensed under the MIT license at:
http://opensource.org/licenses/mit-license.php

I got a D-Link DSM-520 but I needed a UPnP Media Server to stream data
with.  I tried one, but it had issues running under FreeBSD's Linux
emulation.  Since I know Python, I went looking for a python server
and found this code.  The code was a good framework, so I expanded upon
it.

Tested devices and/or programs:
	Intel's Media Control Point and Media Renderer
	D-Link DSM-520

The Intel tools are good for testing and are available at:
http://www.intel.com/cd/ids/developer/asmo-na/eng/downloads/upnp/index.htm

Either make a directory media and put the files there, or make a symlink
named media to your media files.  Either will work.  Run it as:
	./pymediaserv <localip> [ <http server port> ]

Good Luck!

John-Mark Gurney <gurney_j@resnet.uoregon.edu>

Ideas for future improvements:
	Add res to the DVDTitle container to play the title, and for other
	  containers that it makes sense for.
	Make the directory persistant so that the server's UUID can not
	  change each run and we don't have to recreate the objects.
	Figure out how to rearchitect ContentDirectoryControl so I don't
	  need to use doRecall.  This may be helped by not necessarily
	  figuring out all the children of a member just to fetch it.
	  childCount isn't a required attribute.
	Autodetect IP address.
	Support sorting by other attributes.

v0.x:
	Look inside MPEG-TS for TVCT and/or PAT and if there is more
	  than one program, make it a container w/ the different
	  programs.
	Add support for multiple res elements and automatic transcoding
	  to either avi/xvid or mpeg2 using ffmpeg.
	Look inside DVDs and handle titles and chapters.
	Empty dirs w/ no content would disappear, and cause a short
	  response to BrowseDirectChildren.  The DSM-520 askes for one
	  more than displayed, and uses the existant of the extra item
	  as indication if there are more items.
	Understands zip and tar files now.  It will dynamicly extract
	  items, so you can zip up your jpegs and view them w/o having
	  to unzip them.  tar files can be gzip'd or bzip2'd.
	Each item now has an optinal content attribute, which if set is
	  installed in the web server.
	Don't send 'foobar' when exiting, stops a traceback on another
	  instance of PyMedS.
v0.2:
	No longer require restarting to see new files/dirs in hierarchy.
	Add FSStorage which is a set of classes for handling filesystem
	  objects, also handles updateID's.
	Make the root container customizable, so you don't end up with
	  a single entry (like media) at the root.  This lets us use a
	  FSDirectory as the root and get auto-enumeration
	Support returning custom error codes, so that we can return 701
	  No such object among others.
	Support deleting items, so we can remove them when no longer on
	  the file system.
	Make Containers a subclass of list.  This lets use populate
	  childCount properly.
	Add required attributes to Storage* classes.
	Support custom container classes to addContainer.
	Add a few more custom mime-types.
	Sort listings by name for now.

v0.1:
	Don't bind the UDP socket to the multicast address, so replies
	  go out on our local IP.
	Send out notify requests when we register the services with SSDP.
	Send out byebye notifications when we are shutting down.
	Randomize the UUID for the server (this should be part of the
	  saved state).
	Randomize the port, or optionally set it on the command line.
	Teach ContentDirectory.py the basics on handling Containers and
	  generic browse support.  You can addItem and addContainer, which
	  each return either respective ObjectID.
	We already support partial chunking of responses, but we don't yet
	  support filtering or sorting.