ó ²“Ic@s>dZddlmZddlmZddlmZddlmZed d„Z defd„ƒYZ e d kr:d GHdd l Z dd lmZmZd ZdZd„Zd„Ze jjeƒrGdGHdGHe eedƒdedeƒZdZx"eD]ZedZeeƒqWedks6t‚eejƒGHne jjeƒrÓdGHdGHe eedƒdedeƒZdZx#eD]ZedZeeƒPq‘WedksÂt‚eejƒGHnddlmZdGHdGHe edƒƒZdZxeD]ZedZq Wedks2t‚dGHnd S(s‡Bio.SeqIO support for the "fasta" (aka FastA or Pearson) file format. You are expected to use this module via the Bio.SeqIO functions.iÿÿÿÿ(tsingle_letter_alphabet(tSeq(t SeqRecord(tSequentialSequenceWriterccszx:tr<|jƒ}|dkr%dS|ddkrPqqWx$trc|ddkretdƒ‚n|r||djƒƒ\}}}n&|djƒ}|jƒd}|}g}|jƒ}x_tr&|sØPn|ddkrìPn|j|jƒjddƒjddƒƒ|jƒ}qÈWttdj |ƒ|ƒd |d |d |ƒV|s@dSq@Wt svt d ƒ‚dS( sIGenerator function to iterate over Fasta records (as SeqRecord objects). handle - input file alphabet - optional alphabet title2ids - A function that, when given the title of the FASTA file (without the beginning >), will return the id, name and description (in that order) for the record as a tuple of strings. If this is not given, then the entire title line will be used as the description, and the first word as the id and name. Note that use of title2ids matches that of Bio.Fasta.SequenceParser but the defaults are slightly different. tNit>s6Records in Fasta files should start with '>' characterit s tidtnamet descriptionsShould not reach this line( tTruetreadlinet ValueErrortrstriptsplittappendtreplaceRRtjointFalsetAssertionError(thandletalphabett title2idstlineRRtdescrtlines((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyt FastaIterators8    "  +t FastaWritercBs&eZdZddd„Zd„ZRS(s"Class to write Fasta format files.i<cCsMtj||ƒd|_|r7|dkr7t‚q7n||_||_dS(s¿Create a Fasta writer. handle - Handle to an output file, e.g. as returned by open(filename, "w") wrap - Optional line length used to wrap sequence lines. Defaults to wrapping the sequence at 60 characters Use zero (or None) for no wrapping, giving a single long line for the sequence. record2title - Optional function to return the text to be used for the title line of each record. By default the a combination of the record.id and record.description is used. If the record.description starts with the record.id, then just the record.description is used. You can either use: myWriter = FastaWriter(open(filename,"w")) writer.write_file(myRecords) Or, follow the sequential file writer system, for example: myWriter = FastaWriter(open(filename,"w")) writer.write_header() # does nothing for Fasta files ... Multiple calls to writer.write_record() and/or writer.write_records() ... writer.write_footer() # does nothing for Fasta files writer.close() iN(Rt__init__tNonetwrapR t record2title(tselfRRR((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyRHs    cCs‚|jst‚|j st‚t|_|jrL|j|j|ƒƒ}n_|j|jƒ}|j|jƒ}|r›|j ddƒd|kr›|}nd||f}d|ks½t‚d|ksÏt‚|j j d|ƒ|j |ƒ}d|kst‚d|kst‚|jrjx\tdt|ƒ|jƒD](}|j j ||||j!dƒq;Wn|j j |dƒdS(s(Write a single Fasta record to the file.iis%s %ss s s>%s N(t_header_writtenRt_footer_writtenR t_record_writtenRtcleanRR RRRtwritet_get_seq_stringRtrangetlen(R trecordttitleRR tdatati((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyt write_recordos(  "  ")N(t__name__t __module__t__doc__RRR-(((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyRFs't__main__sRunning quick self testN(tgeneric_proteintgeneric_nucleotides NC_005213.fnas NC_005213.faacCsN|jddƒ\}}|jdƒd}|jddƒd}|||fS(Nit|it.i(RR(ttextRRR((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pytgenbank_name_functionœscCsˆd|jGHd|jGHd|jGH|jGHx'|jD]}d||j|fGHq6W|jr„dGHx|jD]}d|GHqnWndS(NsID:sName:sDescr:s/%s=%ssDatabase cross references:s - %s(RRR tseqt annotationstdbxrefs(R)tfeaturetx((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyt print_record¢s    s--------sFastaIterator (single sequence)trRRiis"FastaIterator (multiple sequences)(tStringIOs FastaIterator (empty input file)RtDone(R0t Bio.AlphabetRtBio.SeqRt Bio.SeqRecordRt InterfacesRRRRR.tosR2R3t fna_filenamet faa_filenameR7R=tpathtisfiletopentiteratortcountR)Rtstrt __class__t cStringIOR?(((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/FastaIO.pyt sX2K    !  !