[Biojava-dev] "No match found" in MSFFormatReader

[Thomas Down]

Hi...

Once upon a time, Bradford Powell wrote:
> 
> When trying to read an MSF-formatted alignment (either using
> SeqIOTools.fileToBiojava or using MSFAlignmentFormat directly) I get an
> error message saying "MSFFormatReader No match found".
> 
> I think what needs to happen is for matcher.find() to be called before
> trying to get the sequence name out of the match. I suggest changing the
> code that reads: (starting with line 29 of MSFAlignmentFormat.java)
> 
> ------------------
>                 Matcher matcher = mtc.matcher(line);
>                 sequenceName = matcher.group(1).trim();
>                 if (sequenceName == null) {            
>                     break;
>                 }               //end of sequence names
>                 //sequenceName = line.substring(rem.getSubStartIndex(1),
>                 //                              rem.getSubEndIndex(1));        
>                 if ((line.trim()).length() == 0) {                     
>                     break;
>                 }
>                 sequenceNames.add(sequenceName);
> -------------------
> so that it becomes:
> -------------------
>                 Matcher matcher = mtc.matcher(line);
>                 if (!matcher.find()) {            
>                     break;
>                 }               //end of sequence names
>                 //sequenceName = line.substring(rem.getSubStartIndex(1),
>                 //                              rem.getSubEndIndex(1));        
>                 if ((line.trim()).length() == 0) {                     
>                     break;
>                 }
>                 sequenceName = matcher.group(1).trim();
>                 sequenceNames.add(sequenceName);
> -------------------
> (i.e. change the break condition to not finding a match and move the
> matcher.group(1).trim() call to after matcher.find() has been called).

Okay, that seems to work -- I've just applied this and
checked it in.  Thanks.

I've just added a simple test for MSFAlignmentFormat to the
test suite, so hopefully we'll spot any future regressions.

     Thomas.