[Biococoa-dev] read Fasta File with gap symbols

[Scott Christley]

On Apr 1, 2009, at 5:19 PM, Craig Bateman wrote:

> I was going to recommend (or at least ask permission) to omit that  
> very code while I'm working on the file-based vs memory-based  
> sequence classes.  Unfortunately, things have been very crazy at  
> work for the last 2 months, so I haven't had the time I'd like to  
> work on this.  I only have skeletons of the new BCSequence class  
> cluster together, but I was not finding an easy way to filter the  
> file-based sequences like the memory-based ones are at init time.   
> Omitting that code from the library is a breaking change...  So it's  
> probably not going to be well-received by any end-users.

Yes, it doesn't seem reasonable to filter file-based sequences.  It's  
a bit unclear to me why blindly removing symbols is considered  
acceptable behavior in the first place.

I can see from a contract standpoint, the SymbolSet guarantees to  
algorithms working on the sequence, that they will only see a  
particular set of symbols, this is useful.  However, the idea of  
"guessing" a symbol set then removing symbols which doesn't conform,  
doesn't make much sense to me, especially with how simplistic the  
"guessing" code is.  What seems more reasonable is that you don't  
guess, you scan the sequence and construct a symbol set, if that  
symbol set matches a standard one, great, otherwise you have a unique  
symbol set that goes with that sequence.



> I have not heard back from Peter Schols regarding repository  
> access.  I would really like to branch the source and check-in on a  
> regular basis. Is there anyone else that can grant this access?  If  
> not to the trunk, at least to a branched folder?

I have admin access; if you give me your bioinformatics.org account  
name then I can add you to the project.  Yes please, check in what you  
have done, a branch is fine too if your work is in an intermediate  
state.

Scott