[Biophp-dev] parser news

biophp-dev@bioinformatics.org biophp-dev@bioinformatics.org
Mon, 02 Jun 2003 8:30:56 PDT


 
> >Just finished reworking the parser class.  Methods move_Next() is now
> >called next() and method move_Previous() is called prev().  I
> propose to use these names in all methods that 'walk' through records.
> 
> Duh?  It has always been named next() and prev()
> in the original GenePHP code.  I don't know why
> we have to rename things, esp. in the direction
> of short to long.

Ehmm, naming is simply very important,especially in this  development
phase. But...  in the heat of coding you often do not pay that   much
attention to names (at least,I don't).  These names (that came from me
originally) were simply too long.  The parser class was not present in
the original GenePHP code, so there never has been a renaming from short
to long as far as I am aware of.  It will simply be important to be
consistent in these matters, so, since you wrote most of the code, it
would be great if you could have a look at these things in the newly
added code.  

 
> >The Genbank parser has been reworked substantially to >deal with a
> couple of apparently non-standard genbank >like files.  
> 
> Hmmm... are you referring to the CODATA format?

What is the CODATA format?
I am refering to files produced by Vector NTI, and an old (1988) Genbank
file coming from a Mac. I simply want the thing to work on any sequence
file I can get my hands on. 

The main problems occurr in the first line.  Apparently not all files
adhere strictly to the placement of the various tags, and it becomes a
guessing game to know which tag is which. 

Also, the COMMENT section was implemented as a subsection (of FEATURES, I 
forgot), which does not work for Vector NTI files (it is now a section by 
itself)


> Btw, would you have any idea on how to best organize
> and store these sections in a PDB file?
> 
> REMARK N                    
> ANISOU                      SIGUIJ                      
> HETATM                      
> MODEL and ENDMDL  
> CONECT  
> MASTER                      

No.  Are you trying to write a PDB file?  I can try asking some
crystallographers,but would need to know what the exact question is...


Best,

Nico