o RŀgEV@sdZddlmZddlZddlmZddlmZddlm Z ddlm Z ddlm Z dd l m Z dd l mZdd l mZdd l mZdd lmZGdddejZGdddejZdS)aBio.Align support for the "psl" pairwise alignment format. The Pattern Space Layout (PSL) format, described by UCSC, stores a series of pairwise alignments in a single file. Typically they are used for transcript to genome alignments. PSL files store the alignment positions and alignment scores, but do not store the aligned sequences. See http://genome.ucsc.edu/FAQ/FAQformat.html#format2 You are expected to use this module via the Bio.Align functions. Coordinates in the PSL format are defined in terms of zero-based start positions (like Python) and aligning region sizes. A minimal aligned region of length one and starting at first position in the source sequence would have ``start == 0`` and ``size == 1``. As we can see in this example, ``start + size`` will give one more than the zero-based end position. We can therefore manipulate ``start`` and ``start + size`` as python list slice boundaries. )chainN) Alignment) interfaces)reverse_complement)Seq)UndefinedSequenceError)CompoundLocation) ExactPosition) SeqFeature)SimpleLocation) SeqRecordcs6eZdZdZdZd fdd Zdd Zd d ZZS) AlignmentWriterzEAlignment file writer for the Pattern Space Layout (PSL) file format.PSLTNNcsLt|||_|dur|dkrt|}nt|}||_||_dS)aCreate an AlignmentWriter object. Arguments: - target - output stream or file name - header - If True (default), write the PSL header consisting of five lines containing the PSL format version and a header for each column. If False, suppress the PSL header, resulting in a simple tab-delimited file. - mask - Specify if repeat regions in the target sequence are masked and should be reported in the `repMatches` field of the PSL file instead of in the `matches` field. Acceptable values are None : no masking (default); "lower": masking by lower-case characters; "upper": masking by upper-case characters. - wildcard - Report alignments to the wildcard character in the target or query sequence in the `nCount` field of the PSL file instead of in the `matches`, `misMatches`, or `repMatches` fields. Default value is 'N'. Nupper)super__init__headerordlowerrwildcardmask)selftargetrrr __class__A/var/www/html/myenv/lib/python3.10/site-packages/Bio/Align/psl.pyr2s   zAlignmentWriter.__init__cCsN|jsdSz|j}Wn tyd}Ynw|dd}|d|ddS)zWrite the PSL header.N3psLayout versionzpsLayout version a match mis- rep. N's Q gap Q gap T gap T gap strand Q Q Q Q T T T T block blockSizes qStarts tStarts match match count bases count bases name size start end name size start end count --------------------------------------------------------------------------------------------------------------------------------------------------------------- )rmetadataAttributeErrorgetwrite)rstream alignmentsr versionrrr write_headerTs   zAlignmentWriter.write_headerc(Cst|ts td|j}|jsdS|j\}}z|j}Wn ty&d}Ynwz|j}Wn ty5Ynwz|j}Wn tyFd}Ynwz|j}Wn tyUYnwt |}t |}t t |d} | ddk} | dd| fd\} } | | krd} |d |d kr|ddddd f}|d |d krd}t|}|}||dddf|dddf<n9d}n6| d| krd} |d |d krd}t|}|}||dddf|dddf<n d}ntd| | f|j}|j}d}d}d}d}d}d}d}d}g}g}g}|dddf\}}|ddddfD]\}}||krC|dkr?||kr?|d7}|||7}|}q!||kr`|dkr\||kr\|d7}|||7}|}q!||} ||} |||||| | | kr| dusJn| d| ksJ| dusJ|||} |||}!zt| } Wntyt| d} Yn tyd} Ynwzt|!}!Wntyt|!d}!Yn tyd}!Ynw| dus|!dur|| 7}n|dkr2t| |!| D]0\}"}#}$|"|ks|#|kr|d7}q|"|#kr+|"|$kr%|d7}q|d7}q|d7}qno|dkrst| |!| D]0\}"}#}$|"|ksP|#|krV|d7}qA|"|#krl|"|$krf|d7}qA|d7}qA|d7}qAn.t| |!D]$\}"}#|"|ks|#|kr|d7}q||"|#kr|d7}q||d7}q||}|}q!z|j}Wn tyYnwz|j}Wn tyYnwz|j}Wn tyYnwz|j}Wn tyYnw|d}|d}|d | }|d | }|dkr| dur||||}}n ||||}}t |}%dt t!|d}dt t!|d}dt t!|d}| rAd|}t!|t!|t!|t!|t!|t!|t!|t!|||t!|t!|t!||t!|t!|t!|t!|%|||g}&d|&d}'|'S)zBReturn a string with a single alignment formatted as one PSL line.zExpected an Alignment objectqueryrrNFrrrr-r*rr*r--+Tzinconsistent step sizes %d, %dASCIIrr,  )" isinstancer TypeError coordinatessize sequencesidr!seqlenabsnpdiffminsumrcopy ValueErrorrr transposeappendbytesrziprrmatches misMatches repMatchesnCountjoinmapstr)(r alignmentr9rr)qNametNametSizeqSizestepsalignedtCountqCountdnaxstrandrrrJrKrLrM qNumInsert qBaseInsert tNumInsert tBaseInsert blockSizesqStartstStartstStartqStarttEndqEndtSeqqSequ1u2c1 blockCountwordslinerrrformat_alignmentjsx          " ""                                        z AlignmentWriter.format_alignment)TNr) __name__ __module__ __qualname____doc__fmtrr'ro __classcell__rrrrr -s "r c@s$eZdZdZdZddZddZdS)AlignmentIteratora$Alignment iterator for Pattern Space Layout (PSL) files. Each line in the file contains one pairwise alignment, which are loaded and returned incrementally. Alignment score information such as the number of matches and mismatches are stored as attributes of each alignment. rcCst|}|drA|}|ddkrtd|dd|di|_t|}t|}t|}t|}|ddkr?td dS||_dS) Nz psLayout r*r&z#Unexpected word '%s' in header linerr0r(zEnd of header not found)next startswithsplitrEr lstripstrip_line)rr$rnrmrrr _read_header@s   zAlignmentIterator._read_headerc.Cs|z|j}Wn ty|}Yn w|`t|g|}|D]}|}t|dkr+d}nt|dkr4d}ntdt||d}|d}t|d}|d } t|d } t|d } d d |dddD} dd |dddD} dd |dddD}t| | krtdt| | ft| | krtdt| | ft|| krtdt|| ft | } t |}t | }|dvrd|}n|}| d}|d}||gg}t |||| D]0\}}}}||kr| ||g|}||kr| ||g|}||7}||7}| ||gqt | }d}d}d}d}|dddf\}}|ddddf D]G\}}||}||}|dkrb|dkr^||kr^|d7}||7}|}q;|dkr}|dkry|| kry|d7}||7}|}q;|}|}q;|t|dkrtd|d|f|t|dkrtd|d|f|t|dkrtd |d|f|t|d!krtd"|d!|ft|d#}t|d$}t|d%}t|d&}|d'kr||}}||dddf|dddf<n|d(kr||}}| |dddf|dddf<||d)kr-td*||d)f||d+kr>td,||d+f||d-krOtd.||d-f||d/kr`td0||d/fd} |durn|ddd}!|d1dd}"t t | |!}#|dvrQtd| d2}$t|#|d2}%|d3kr|dddf\}}g}&|ddddf D]#\}}||kr||krtt|t|dd4}'|& |'|}|}qt|&dkrt|&d5}'d6|"}(d7|(gi})t|'d8|)d9} n|d(krP|dddf\}}g}&|ddddf D]#\}}||kr-||kr-tt|t|d:d4}'|& |'|}|}qt|&dkr?t|&d5}'d6|"}(d7|(gi})t|'d8|)d9} n)t t ||"}(t|(| d2}$t|#|d2}%|d'krm|%}%n td| d2}$td|d2}%t|$| d6d;}*t|%|d6d;}+| dur|*j | |*|+g},t|,|}-t|d|-_t|d|-_t|d<|-_t|d|-_|-SdS)=NTFz&line has %d columns; expected 21 or 23 cSg|]}t|qSrint).0 blockSizerrr fsz:AlignmentIterator._read_next_alignment..r4cSrrrrstartrrrricSrrrrrrrrjrz5Inconsistent number of blocks (%d found, expected %d)zDInconsistent number of query start positions (%d found, expected %d)zEInconsistent number of target start positions (%d found, expected %d))+++-r2rr*z/Inconsistent qNumInsert found (%s, expected %d)z0Inconsistent qBaseInsert found (%s, expected %d)z/Inconsistent tNumInsert found (%s, expected %d)z0Inconsistent tBaseInsert found (%s, expected %d) r0rr+z+Inconsistent tStart found (%d, expected %d)r,z)Inconsistent tEnd found (%d, expected %d)r.z+Inconsistent qStart found (%d, expected %d)r/z)Inconsistent qEnd found (%d, expected %d))lengthr)r[rNr( translationCDS)type qualifiersr-)r< descriptionrw)r}r!rrzr>rErrstripr@arrayrIrGrFdictrr r rrNr rr featuresrrJrKrLrM).rr$rnlinesrmpslxr[rRrUrSrTrlr`rarb qBlockSizes tBlockSizes qPosition tPositionr9 tBlockSize qBlockSizercrdr\r]r^r_rerfrXrYfeatureqSeqstSeqsrhtarget_sequencequery_sequence locationslocationrgr target_record query_recordrecordsrQrrr_read_next_alignmentPs                                  "                               z&AlignmentIterator._read_next_alignmentN)rprqrrrsrtr~rrrrrrv6s  rv)rs itertoolsrnumpyr@ Bio.AlignrrBio.SeqrrrBio.SeqFeaturerr r r Bio.SeqRecordr r rvrrrrs"