On 4 Feb 2004, Keith James wrote: > >>>>> "Dan" == Dan Bolser <[EMAIL PROTECTED]> writes: > > Dan> Hello, I am a new user to biojava (and almost new to java). > > Dan> The following code works fine reading a 'FASTA' format file, > Dan> but causes an error reading 'MSF' format... > > [...] > > Dan> --- Exception in thread "main" > Dan> java.lang.IllegalArgumentException: No alphabet was set in > Dan> the identifier at > Dan> org.biojava.bio.seq.io.SeqIOTools.fileToBiojava(SeqIOTools.java:801) > Dan> at > Dan> org.biojava.bio.seq.io.SeqIOTools.fileToBiojava(SeqIOTools.java:787) > Dan> at > Dan> ReadAlignMakeDistribution.main(ReadAlignMakeDistribution.java:60) > > The exception message here is referring to the integer identifier > which biojava has for every known combination of file-format (fasta, > genbank, embl) and alphabet-type (dna, rna, protein). The way these > are created/interpreted is documented in SeqIOConstants (for the > sequence formats) and AlignIOConstants (for the alignment > formats). All the common ones exist as static int fields so that you > can compare using == or use them in switches. > > The format guessing code (in SeqIOTools.identifyFormat) appears to be > missing "msf" and "clustal". This is a bug - I'll fix it today. The > result is that it guesses SeqIOConstants.UNKNOWN as the format > identifier (which has no alphabet set - hence the message). > > The public method fileToBiojava(int fileType, BufferedReader br) > should work if you pass it the value AlignIOConstants.MSF_AA
Just for the record, I see two forms of the fileToBiojava function... fileToBiojava( int fileType, java.io.BufferedReader br ); fileToBiojava( java.lang.String formatName, java.lang.String alphabetName, java.io.BufferedReader br ); I was using the second form which sould not have to guess the format (perhaps I misunderstand what you said above). Aditionaly, I explicitly pass an alphabet name... Why are formats linked to alphabets? "... SeqIOConstants.UNKNOWN as the format identifier (which has no alphabet set...". Please let me know if I am terminally confused Cheers, Dan. > > Keith > > _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l
