ó ö.ÓIc@s4dZddlmZddlmZmZddlmZddlm Z d„Z d„Z ej d„Z ej d „Zd e fd „ƒYZed kr0d GHddlZddlmZd„ZxxejdƒD]gZejdƒ rejdƒ rqØneGHedeƒZee eƒƒZejƒeeƒqØWxgejdƒD]VZejdƒsnqSneGHedeƒZee eƒƒZejƒeeƒqSWddlmZxpejdƒD]\ZejdƒsèqÍneGHedeƒZeejedƒƒZejƒeeƒqÍWndS(síBio.SeqIO support for the "genbank" and "embl" file formats. You are expected to use this module via the Bio.SeqIO functions. Note that internally this module calls Bio.GenBank to do the actual parsing of both GenBank and EMBL files. See also: International Nucleotide Sequence Database Collaboration http://www.insdc.org/ GenBank http://www.ncbi.nlm.nih.gov/Genbank/ EMBL Nucleotide Sequence Database http://www.ebi.ac.uk/embl/ DDBJ (DNA Data Bank of Japan) http://www.ddbj.nig.ac.jp/ iÿÿÿÿ(t UnknownSeq(tGenBankScannert EmblScanner(tAlphabet(tSequentialSequenceWritercCstddƒj|ƒS(s Breaks up a Genbank file into SeqRecord objects. Every section from the LOCUS line to the terminating // becomes a single SeqRecord with associated annotation and features. Note that for genomes or chromosomes, there is typically only one record.tdebugi(Rt parse_records(thandle((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pytGenBankIterator's cCstddƒj|ƒS(s Breaks up an EMBL file into SeqRecord objects. Every section from the LOCUS line to the terminating // becomes a single SeqRecord with associated annotation and features. Note that for genomes or chromosomes, there is typically only one record.Ri(RR(R((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt EmblIterator2s cCstddƒj||ƒS(sBreaks up a Genbank file into SeqRecord objects for each CDS feature. Every section from the LOCUS line to the terminating // can contain many CDS features. These are returned as with the stated amino acid translation sequence (if given). Ri(Rtparse_cds_features(Rtalphabet((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pytGenBankCdsFeatureIterator=scCstddƒj||ƒS(sþBreaks up a EMBL file into SeqRecord objects for each CDS feature. Every section from the LOCUS line to the terminating // can contain many CDS features. These are returned as with the stated amino acid translation sequence (if given). Ri(RR (RR ((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pytEmblCdsFeatureIteratorGst GenBankWritercBsYeZdZdZd„Zd„Zd„Zded„Zd„Z d„Z d „Z RS( i iPcCs‰t|ƒ|jkst‚t|ƒ|j|jksStdt|ƒ|fƒ‚|jjd|j|jƒ|jddƒfƒdS(s0Used in the the 'header' of each GenBank record.s"Annotation %s too long for %s lines%s%s s t N( tlent HEADER_WIDTHtAssertionErrort MAX_WIDTHtreprRtwritetljusttreplace(tselfttagttext((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt_write_single_lineUs cCsÂ|j|j}t|ƒ|jks+t‚|jƒ}t|ƒ|kr]|j||ƒdS|jƒ}tg|D]}t|ƒ^qsƒ|ks tdƒ‚d}xQ|rùt|ƒdt|dƒ|krù|d|jdƒ7}|jƒ}q©Wt|ƒ|kst‚|j||ƒxŒ|r°d}xQ|r„t|ƒdt|dƒ|kr„|d|jdƒ7}|jƒ}q4Wt|ƒ|kst‚|jd|ƒq%W| s¾t‚dS(s0Used in the the 'header' of each GenBank record.Ns2Your description cannot be broken into nice lines!tiiR( RRRRtstripRtsplittmaxtpop(RRRtmax_lentwordstw((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt_write_multi_line]s.  . - -cCsa|j}| s|dkr(|j}n| s;|dkrV|j|ddtƒ}nt|ƒdkrtdt|ƒƒ‚nt|ƒdkr¢tdƒ‚ntj|j j ƒ}t |tjƒsØt d ƒ‚nBt |tj ƒród }n't |tjƒrd }n td ƒ‚t |tj ƒr5d }nBt |tjƒrPd}n't |tjƒrkd}n tdƒ‚y|jd}Wntk r¡d}nX|dSkr·d}nt|ƒd%ksÏt‚t|ƒd&ksçt‚d'|jdƒtt|ƒƒjd(ƒ||jd)ƒ|f}t|ƒdTksKtt|ƒƒ‚|d,d-!jƒ|ksttd.|ƒ‚|d-d/!d0kst‚|d/d1!jƒtt|ƒƒksÂtd2|ƒ‚|d1d3!dUksåtd6|ƒ‚|d3d7!dVkstd<|ƒ‚|d7d=!jƒd ksu|d7d=!jƒjdƒd>ksu|d7d=!jƒjdƒd>ksutd?|ƒ‚|d=d@!d0ks˜tdA|ƒ‚|d@dB!jƒdWksÁtdE|ƒ‚|dBdF!d0ksätdG|ƒ‚|dHdI!d0kstdJ|ƒ‚|dKdL!dMks*tdN|ƒ‚|dOdP!dMksMtdQ|ƒ‚|jj|ƒdRS(XsWrite the LOCUS line.ss t accessiont just_firstisLocus identifier %s is too longIÿçvHsSequence too long!sInvalid alphabettbptaas%Need a Nucleotide or Protein alphabetRtDNAtRNAs#Need a DNA, RNA or Protein alphabettdata_file_divisiontUNKtPRItRODtMAMtVRTtINVtPLNtBCTtVRLtPHGtSYNtUNAtESTtPATtSTStGSStHTGtHTCtENViis4LOCUS %s %s %s %s %s 01-JAN-1980 i iiOii is@LOCUS line does not contain the locus at the expected position: iRi(sALOCUS line does not contain the length at the expected position: i,s bp s aa s=LOCUS line does not contain size units at expected position: i/s sss-sds-sms-sCLOCUS line does not have valid strand type (Single stranded, ...): i6iÿÿÿÿsALOCUS line does not contain valid sequence type (DNA, RNA, ...): i7s2LOCUS line does not contain space at position 55: i?tlineartcircularsALOCUS line does not contain valid entry (linear, circular, ...): i@s2LOCUS line does not contain space at position 64: iCiDs2LOCUS line does not contain space at position 68: iFiGt-s6LOCUS line does not contain - at position 71 in date: iJiKs6LOCUS line does not contain - at position 75 in date: N(R-R.R/R0R1R2R3R4R5R6R7R8R9R:R;R<R=R>iP(s bp s aa (s sss-sds-sms-(RR?R@(tnametidt_get_annotation_strtTrueRt ValueErrorRRt_get_base_alphabettseqR t isinstancet TypeErrortProteinAlphabettNucleotideAlphabett DNAAlphabett RNAAlphabett annotationstKeyErrorRRtstrtrjusttrstriptlstripRtfindRR(Rtrecordtlocustatunitstmol_typetdivisiontline((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt_write_the_first_lineys†                $ (   "%       t.cCsry|j|}Wntk r%|SXt|tƒrd|sVt|ƒdksVt‚nt|dƒSt|ƒSdS(súGet an annotation dictionary entry (as a string). Some entries are lists, in which case if just_first=True the first entry is returned. If just_first=False (default) this verifies there is only one entry before returning it.iiN(RORPRItlistRRRQ(RRVtkeytdefaultR&tanswer((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyRDÖs cCsÚd}d}t|jtƒr"dS|j|ƒ}t|ƒ}x–td||ƒD]‚}|jjt|dƒj |ƒƒxFt|t |||ƒdƒD]%}|jjd|||d!ƒq™W|jjdƒqPWdS(Ni<i iii s %ss ( RIRHRt_get_seq_stringRtrangeRRRQRRtmin(RRVtLETTERS_PER_LINEtSEQUENCE_INDENTtdatatseq_lent line_numberR"((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt_write_sequenceæs #&#c Cs‘|j}|j|ƒ|j|d|jjddƒddtƒ}|}|jj|dƒr¤y-d|t|jjddƒdƒf}Wq¤tk r q¤Xn|j|ddtƒ}|j }|dkrÚd}n|j d |ƒ|j d |ƒ|dkr#|j d d ||fƒn|j d d |ƒydj |j dƒ}Wntk rjd}nX|j d|ƒ|j d|j|dƒƒ|j|dƒ}t|ƒ|j|jkrä||j|jd d}n|j d|ƒydj |j dƒ} Wntk r'd} nX|j d| ƒ|jdƒx|jD]} |j| ƒqOW|jdƒ|j|ƒ|jdƒdS(s)Write a single record to the output file.R%R^iiR&s%s.%itgist DEFINITIONt ACCESSIONtVERSIONs %s GI:%ss%ss; tkeywordstKEYWORDStSOURCEtsourcetorganismis...s ORGANISMttaxonomyRs)FEATURES Location/Qualifiers sORIGIN s// N(RR]RDRCRREt startswithtintRFt descriptionR$RtjoinRORPRRRRtfeaturest_write_featureRk( RRVRR%tacc_with_versionRltdescrRptorgRutfeature((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyt write_recordùsV    *            cCsdS(sÁWrite a single SeqFeature object to features table. Not implemented yet, but this stub exists in the short term to facilitate working on writing GenBank files with a sub-class.N((RR((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyR{1s( t__name__t __module__RRRR$R]tFalseRDRkR€R{(((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pyRQs   ]  8t__main__sQuick self testN(tStringIOcCsètƒ}t|ƒj|ƒ|jdƒtt|ƒƒ}t|ƒt|ƒksYt‚xˆt||ƒD]w\}}|j j ddƒ|j ks™t‚|j |j ks±t‚|j |j ksÉt‚t |jƒt |jƒksít‚xRddddgD]>}||jkr|j||j|ks>t|ƒ‚qqWx›dgD]}||jkrL|j|}|j|}t|t ƒr™t|t ƒsŸt‚||ksÜ|jd ƒrÍ|j|d ƒsÜt|ƒ‚qLqLWqiWdS( Nis RRlRpRsRuRts...iýÿÿÿ(R…Rt write_filetseekR_RRRtzipRxRRCRBRQRHRORItendswithRv(trecordsRtrecords2tr1tr2R`tv1tv2((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pytcheck_genbank_writer>s(  $$-  $ s../../Tests/GenBanks.gbks.gbs../../Tests/GenBank/%ss../../Tests/EMBLs.embls../../Tests/EMBL/%s(tSeqIOs../../Tests/SwissProttsps../../Tests/SwissProt/%stswiss(t__doc__tBio.SeqRtBio.GenBank.ScannerRRtBioRt InterfacesRRR tgeneric_proteinR R RRtosR…RtlistdirtfilenameR‰topenRR_RŠtcloseR‘Rvtparse(((s„/oak/stanford/groups/akundaje/marinovg/programs/biopython-1.50.tar.gz/biopython-1.50/build/lib.linux-x86_64-2.7/Bio/SeqIO/InsdcIO.pytsN   è