[Biopython-dev] Last chance to clam PIR parser

[Andrew Dalke]

Cayte:
>  The PIR format described in
>www.sander.embl-ebi.ac.uk/Services/webin/help/webin-align/align_format_help
.html#pir

> seems different and more fastalike than the way you describe PIR.  Are
> there different renditions of pir?

Oh, right.  Yes, there are.  The PIR format I have is for the CODATA
format, which includes a lot more data than the NBRF format you're
probably thinking of.

PIR releases data in three formats:
  CODATA -- human readable (meaning it has 2D formatting to make
      it easier to find the different sections), hard to parse --
      even with Martel
  NBRF -- machine readable, hard for humans to read
  XML (new) -- somewhere in the middle
These are linked from
  http://pir.georgetown.edu/pirwww/dbinfo/pirpsd.html

It seems most people want to read the NBRF format, not the CODATA
one.  But I did the CODATA one because I was thinking about the
"convert to HTML" ability of Martel.  Also, there were more fields
in the CODATA format than the other two -- at least, there were
fields there which were undocumented.  I sent some email to the
PIR people on them but never got a response about what they meant.

                    Andrew
                    dalke at dalkescientific.com