README

README for CDDB.py and DiscID.py, version 1.4

Copyright (C) 1999-2003 Ben Gertzfield <che@debian.org>
-------------------------------------------------------

The dynamic duo of CDDB.py and DiscID.py, along with their side-kick C
module cdrommodule.so, provide an easy way for Python programs to
fetch information on audio CDs from CDDB (http://www.cddb.com/) -- a
very large online database of track listings and other information on
audio CDs.  UNIX platforms and Windows are both supported.

This work is released under the GNU GPL; see the file COPYING for more
information.

Note that for CDDB.py version 1.3 and later, the default CDDB servers
have changed to FreeDB, as GraceNote, the owners of the
previously-free CDDB archives, have become unreasonably restrictive
with access to their database.  See http://www.freedb.org/ for more
information on FreeDB.

The old CDDB servers are still usable, but GraceNote has banned all
"non-registered" programs -- including CDDB.py -- from accessing their
databases.  I'm not particularly interested in bending over backwards
for them, but CDDB.query() can still access any CDDB host you specify,
and from version 1.3 on, also allows you to 'fake' the identification
string to anything you wish for accessing their servers.

-----

If you are using Python 1.6 or 2.0 or later on Linux, FreeBSD,
OpenBSD, Mac OS X, or Solaris, installing is simple.  As root, run:

# python3 setup.py install

(This also works if you have installed Distutils on an older
Python. See http://www.python.org/sigs/distutils-sig/download.html for
more information.)

-----

If you have an older version of Python installed on Linux, FreeBSD,
OpenBSD, Mac OS X, or Solaris:

% make -f Makefile.pre.in boot
% make

Then copy CDDB.py, DiscID,py, and cdrommodule.so to your local
Python packages directory (usually /usr/lib/python1.5/site-packages).

-----

If you are using Python 2.0 on Windows, you can simply extract the
following files to your C:\Python20\Lib\ directory:

CDDB.py DiscID.py win32\cdrom.py win32\mci.dll

Older revisions of Python will require a by-hand recompile of
win32\mci.c; win32\mci.dsp has been provided for Visual C++ users.

I do not have access to a Windows compiler, so I cannot support the
module very well.  However, I have tested that it works on Windows
2000 with Python 2.0.  Thanks to Mike Roberts <zoarre@yahoo.com> for
recompiling the mci.dll library for Python 2.0!

There is an experimental setup-win32.py Distutils, which (in theory)
works the same as the *nix method (python setup-win32.py install) on
Windows.  Completely untested.

-----

The distribution is split up into three parts:

* CDDB.py, for querying the online CDDB servers.
* cdrommodule.so, a C extension module, for actually querying the
  CD-ROM drive for the track start times and other data.
* DiscID.py, a wrapper around cdrommodule.so, which formats the
  track data the way the CDDB protocol requires it.

Also included is cddb-info.py, a quick example of how to use CDDB.py
and DiscID.py together to fetch the track info of an audio CD.  Feel
free to use this to understand how to make your own Python CDDB
applications!

For another example as to how CDDB.py works, see Russ Nelson's
MP3 renamer program: http://russnelson.com/rename-tracks

DiscID.py
---------

The DiscID.py module is a platform-independant wrapper around the
cdrommodule.so C extension, and is used to get the track information
of an audio CD into the format the CDDB protocol requires.

There are two methods of the DiscID module that are of interest:

device = DiscID.open()

  This is the default, cross-platform way to open the audio
  CD-ROM device.  The device defaults to /dev/cdrom on Linux and BSD,
  /dev/disk1 on Mac OS X, /dev/vol/aliases/cdrom0 on Solaris, and
  cdaudio under Windows.  You do not have to use this function to open
  the audio CD-ROM device, but it's a sane default.

  If you want to specify a different device name or flags for opening,
  use the following:

device = DiscID.open(devicename, flags)

  The optional flags parameter should be an int under *nix, and a
  string under Windows.  Flags come from the fcntl module under *nix.

disc_id = DiscID.disc_id(device)

  Given an opened file object corresponding to a CD-ROM device (either
  opened by hand, or returned from DiscID.open above), return an array
  of the format:

  [cddb_checksum, num_tracks, num_frames 1, ... num_frames n, num_seconds]

  cddb_checksum is the 128-bit checksum of the track information,
  as specified by CDDB

  num_tracks is the number of tracks on the disc

  num_frames x is the offset of the start of track x, in terms of
  frames. A frame is simply 1/75th of a second.  x ranges from
  1 to num_tracks.

  num_seconds is the total number of seconds on the disc, as measured
  by the start of the special 'lead-out' track.

The disc_id returned from DiscID.disc_id may look complex, but it's
exactly in the format CDDB wants, and is the format needed for
CDDB.query(disc_id), as described below.

CDDB.py
-------

The CDDB.py module is platform-independant; its job is to query
the online CDDB servers via HTTP and retrieve track listings and
any other data relevant to the specified audio CD.

There are two methods of the CDDB module that are of interest:

(status, info) = CDDB.query(disc_id, [server_url], [user], [host],
                            [client_name], [client_version])

  Given a disc_id, as returned from DiscID.disc_id() above,
  return a tuple of two items. The first item returned is an integer
  indicating the status returned from the server; it will be one of
  the following:

  * 200: Success
  * 211: Multiple inexact matches were found
  * 210: Multiple exact matches were found
  * 202: No match found
  * 403: Error; database entry is corrupt
  * 409: Error; no handshake. (client-side error?)

  Any other return code is failure due to a server or client error.

  The actual contents of the second item returned depend on the
  status returned by the server. On 200, a dictionary containing
  three items is returned:

  info['category']: The category the audio CD belongs in
  info['disc_id']:  The CDDB checksum disc ID of the given disc
  info['title']:    The title of the audio CD

  On 211 or 210, a list will be returned as the second item.  Each
  element of the list will be a dictionary containing three items,
  exactly the same as a single 200 success return.

  On any other status, None will be returned as the second item.

  By default, the server http://freedb.freedb.org/~cddb/cddb.cgi will
  be used. You can specify a different server URL as the second
  argument to CDDB.query().

  The identification string used to tell the CDDB servers what
  program is accessing them defaults to CDDB.py, but you can
  change that by specifying different client_name and
  client_version arguments to CDDB.query().

  The optional user and host arguments are sent in the 'hello'
  portion of the CDDB query. They should correspond to the
  user's email address (if you believe in the CDDB docs, that
  is -- it may not be nice to send the user's email address
  without asking them!)

  By default, if the EMAIL environment variable is set, user
  and host will be set by parsing $EMAIL into user@host.

  Otherwise, user will default to os.geteuid(), falling back
  to os.environ['USER'], then just plain 'user', and host
  will default to socket.gethostname() or just plain 'host' if
  that fails.

(status, info) = CDDB.read(category, disc_id, [server_url], [user], [host],
                           [client_name], [client_version])

  Given a category and a disc ID checksum as returned from
  CDDB.query(), return a tuple of two items. The first item returned
  is an integer indicating the status, as above. It will be one of
  the following:

  * 210: Success
  * 401: Specified entry not found
  * 402: Server error
  * 403: Database entry is corrupt
  * 409: No handshake
  * 417: Access limit exceeded

  Any other return value is a server or client error.

  The second value returned from CDDB.read() depends on the status
  returned by the server. On 210, a dictionary will be returned,
  the keys of which follow the keywords returned by the CDDB read
  command.

  Dictionary entries that will be of interest (note: not all of
  these keys will always be available for all discs):

  info['TTITLE#'] (where # ranges from 1 to the last track on the disc):
    The title of track number #. Note that this is not padded on the
    left with zeroes.

  info['DTITLE']
    The title of the disc.

  info['EXTD']
    Extended data on the CD. (credits, etc.)

  info['EXTT#']
    Extended data on track number #.

  info['submitted_via']
    Optional; name of the client that submitted the data for this disc

  info['revision']
    Optional; revision of this CDDB entry

  info['DYEAR']
    Optional; year this disc was released

  info['DGENRE']
    Optional; Genre string describing this disc

  Note that aside from TTITLE#, most of these fields will be empty
  strings for most discs. Also, submitted_via and revision are
  technically in the comments returned from the server, so may or
  may not even be there. Test first!

  As with CDDB.query(), server_url, user, host, client_name, and
  client_version are all optional. See above for the defaults.

cdrommodule.so
--------------

  This is the platform-specific part of the trio. I have included
  unix/cdrommodule.c, which implements the necessary functions under
  Linux, FreeBSD, OpenBSD, Mac OS X, and Solaris, but the interface is
  intended to be easy enough to implement on any platform.

  If you want to write your own cdrommodule.so for your platform,
  here are the functions you'll need to export:

  device = cdrom.open([device], [flags])
    If given no arguments, perform the proper platform-specific
    steps to open the default CDROM audio device and return an opened
    Python file object corresponding to it.

    If called with one argument, override the default CDROM audio
    device with this argument.  If called with two arguments, open
    the device with the specified flags (integer on *nix, string on
    Windows.)

    On Windows, which does not use an opened file object to represent
    the CDROM audio device, returning a string describing the device
    is acceptable, as the other functions will just use that instead
    of the opened file.

  (first, last) = cdrom.toc_header(file)
    Given an open file object corresponding to a CD-ROM device,
    return a tuple of two integer values, (first, last). These are the
    track numbers of the first and last tracks, respectively.

  (min, sec, frame) = cdrom.toc_entry(file, track)
    Given an open file object and a track number, return a tuple
    of three integer values, (minute, second, frame). These correspond
    to the start time of the track, as given in MSF form. Note that
    a frame is simply 1/75th of a second.

  (min, sec, frame) = cdrom.leadout(file)
    Given an open file object, return a tuple of three integer values,
    (minute, second, frame). These correspond to the start time of
    the special 'leadout' track, generally track 0xAA, as given in
    MSF form.

    If your platform (i.e. Windows) cannot return this leadout
    track's time directly, you'll need to calculate it by taking the
    starting position of the last track and adding the length of the
    last track onto that. Watch out, because Windows' awful interface
    returns the length of the last track as one frame short of the
    real value!

  Note that I throw the exception 'cdrom.error' upon an error;
  you should stick to throwing that if your ioctls (or whatever
  your platform uses) fail, so we can stay consistant.

  It's relatively easy to write your own cdrommodule.so for whatever
  platform you're on. Look at unix/cdrommodule.c for hints. If you
  end up doing it and you'd like to have your module included in
  the source distribution, just contact me at the email address
  above.

  Thanks to Viktor Fougstedt <viktor@dtek.chalmers.se> for the porting
  information for Solaris, and Michael Yoon <michael@yoon.org> for the
  FreeBSD port!

  Also, thanks to Frank David <F.David@digitecgmbh.de> for the win32
  port.  Note that this email address seems to be gone; I have no more
  contact with Frank.  Mike Roberts <zoarre@yahoo.com> has been extremely
  helpful in providing an updated mci.dll for Python 2.0; thank you,
  Mike!

  And thanks to Alexander S . Guy <a7r@andern.org> for the OpenBSD
  patch.

  Finally, thanks to Andre Beckedorf and Jeffrey C. Jacobs for the
  Mac OS X cdrommodule.c port.