[Biopython-dev] Martel in CVS

[Andrew Dalke]

Getting there.  I'm able to dialup and get my linux machine
talking to CVS on biopython.org.  I even have a little program
to set up the SSH connection so once I'm connected I can
do CVS over it without it asking for a password every time.

Want to make sure about the configuration before I do the
final commit.  As I recall we said that
  - Martel is a sibling to Bio
  - Martel should be able to be distributed w/ or w/o the read of
biopython
  - Biopython should allow distributions w/ or w/o Martel, as determined
      by a flag Brad added to setup.py

There are some things I need to do to make things integrate nicely.

  Martel needs its own setup.py for independent distribution.  I
  assume it should be independent from the main setup.py (that is,
  there will be a Martel/setup.py).

  I'm planning to move the format definitions to biopython proper,
  and pull most (all?) of them out of Martel/formats.  Where
  should they go?  It looks like they will be scattered about,
  so they go into the directory associated with that datatype.

  Where do the tests go?  There are two types of tests.  Some are
  integral to Martel and should stay with a Martel distribution.
  Others are related to the formats.  I'm thinking of putting
  the format tests under Tests, which would only contain tests
  related to syntax matching like Martel already has.  The
  problem there would be in upping the namespace pollution of
  putting all the tests in one spot.  But we've only a couple
  dozen so far so it isn't a serious problem.

  I'm going to make all the core Martel tests use the Python 2.1
  unittest and the other biopython tests (in Test) use
  Test/unittest.py.  I hope they aren't too different.

  Where should the documentation go?  I think it's okay to
  leave it in Martel/doc since nearly everything that is there
  is specific to Martel the parser, and not the format definitions
  for biopython.

  I will scatter the files in the Martel/builders/ directory
  to the appropriate parts of biopython.

  This all means there will be two different ways to parse
  the same data files.  How should that changeover be managed,
  since the two styles of processing are somewhat different?

					Andrew
					dalke at acm.org