o Dfi@sndZddlZddlZddlZddlZddlZddlmZmZm Z m Z m Z m Z m Z ddlmZmZmZddlZddlmZddlmZmZmZddlmZddlmZmZmZmZm Z m!Z!m"Z"m#Z#m$Z$m%Z%m&Z&m'Z'm(Z(m)Z)m*Z*m+Z+m,Z,dd l-m.Z.m/Z/m0Z0dd l1m2Z2m3Z3m4Z4m5Z5m6Z6m7Z7m8Z8m9Z9dd l:m;Z;mZ>dd l?m@Z@mAZAeBZCGd ddeZDGdddeEZFdefddZGdeHdeeIeIffddZJdeHdeeeIdffddZKde>de eHde eHde6fddZLdd ZMd!d"ZNde eHeOffd#d$ZPdeOfd%d&ZQd'e eHd(eOd)eOdeeHeeHffd*d+ZRd(eOddfd,d-ZSde2fd.d/ZTdee ee effd0d1ZUd2e2d3e eId4e eId(eOfd5d6ZVd(eOd7eOd8eeHd9eId:eOd;eOdZWde efd?d@ZXdAdBZYdCdDZZej[j\fde0fdEdFZ]d2e2dGe5dHe6dIede7fdLdMZ^dNdOZ_dPdQZ`dJeIdeeIfdRdSZadTdUZbecdVkredeZdSdS)Wa cutadapt version {version} Copyright (C) 2010-2021 Marcel Martin cutadapt removes adapter sequences from high-throughput sequencing reads. Usage: cutadapt -a ADAPTER [options] [-o output.fastq] input.fastq For paired-end reads: cutadapt -a ADAPT1 -A ADAPT2 [options] -o out1.fastq -p out2.fastq in1.fastq in2.fastq Replace "ADAPTER" with the actual sequence of your 3' adapter. IUPAC wildcard characters are supported. All reads from input.fastq will be written to output.fastq with the adapter sequence removed. Adapter matching is error-tolerant. Multiple adapter sequences can be given (use further -a options), but only the best-matching adapter will be removed. Input may also be in FASTA format. Compressed input and output is supported and auto-detected from the file name (.gz, .xz, .bz2). Use the file name '-' for standard input/output. Without the -o option, output is sent to standard output. Citation: Marcel Martin. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.Journal, 17(1):10-12, May 2011. http://dx.doi.org/10.14806/ej.17.1.200 Run "cutadapt --help" to see all command-line options. See https://cutadapt.readthedocs.io/ for full documentation. N)TupleOptionalSequenceListAnyIteratorUnion)ArgumentParserSUPPRESS HelpFormatter) __version__)warn_duplicate_adaptersAdapterInvalidCharacter) AdapterParser)SingleEndModifierLengthTagModifier SuffixRemoverPrefixSuffixAdder ZeroCapperQualityTrimmerUnconditionalCutter NEndTrimmer AdapterCutterPairedAdapterCutterErrorPairedAdapterCutterNextseqQualityTrimmer ShortenerReverseComplementerPairedEndRenamerRenamerInvalidTemplate) full_reportminimal_report Statistics)PipelineSingleEndPipelinePairedEndPipeline InputPaths OutputFilesPipelineRunnerSerialPipelineRunnerParallelPipelineRunner)available_cpu_countProgress DummyProgress FileOpener) setup_loggingREPORTcs8eZdZdZGdddeZfddZddZZS)CutadaptArgumentParserz This ArgumentParser customizes two things: - The usage message is not prefixed with 'usage:' - A brief message is shown on errors, not full usage cs&eZdZfddZdddZZS)z/CutadaptArgumentParser.CustomUsageHelpFormattercsJzddl}td|j|d<Wn tyYnwtj|i|dS)Nrhwidth)shutilminget_terminal_sizecolumns ImportErrorsuper__init__)selfargskwargsr6 __class__Z/var/www/html/software/conda/envs/catlas/lib/python3.10/site-packages/cutadapt/__main__.pyr<Zs z8CutadaptArgumentParser.CustomUsageHelpFormatter.__init__NcCs*|tur|||df}||j|dSdS)N)r _add_item _format_usage)r=usageactionsgroupsprefixr>rBrBrC add_usagebs z9CutadaptArgumentParser.CustomUsageHelpFormatter.add_usageN)__name__ __module__ __qualname__r<rK __classcell__rBrBr@rCCustomUsageHelpFormatterYs rQcs4|j|d<|ddt|d<tj|i|dS)Nformatter_classrGz {version})rQreplacer r;r<)r=r>r?r@rBrCr<gs zCutadaptArgumentParser.__init__cCs8tdtjdtdtjd|ddj|j|ddS)z If you override this in a subclass, it should not return -- it should either exit or raise an exception. z2Run "cutadapt --help" to see command-line options.filez|jd?d"d@d!dAdB|jdCdDtdEddFd>|jdGdHtdIdJdKd>|jdLdddMd|jdNdOd"d!dPdQdR|jdSdTdUdVdW|jdXdYdZddd[d#|jd\d]d^d_td`|jdad]d^dbtd`|dc}|jddded+gtdfdgdh|jditddjdkdl|jdmdnddodpdq|jdrtdsdtdudl|jdvdwtddfdxdl|jdydddzd|jd{d|d}d~|jdd+gdd|jddddd|jddddd|jdddd~|jdddddd|jddd&}|jddddddq|jddddddq|jdtddEddl|jddtddddl|jdddddd|jdddddd|jddddd|d}|jddddd|jdddddW|jddddd~|jddddd|jdd^dddd|jdddd~|jddddd~|jdddd~|jdddd~|jdddd~|jdddddq|jddd&}|jddd*dd+gd,ddʍ|jddd*dd+gd,ddʍ|jddd*dd+gd,ddʍ|jddd+gtdfddԍ|jddddd~|jdddd|jddddddލ|jddddd|jdddd~|jddddd|jddddd|jddtd|S)NF)rGadd_helpOptionsz-hz--helphelpzShow this help message and exit)actionrdz --versionversionzShow version number and exit)rerdrfz--debugcountrz4Print debug log. Use twice to also print DP matrices)redefaultrdz --profile store_truez-jz--coreszFNumber of CPU cores to use. Use 0 to auto-detect. Default: %(default)s)typerhrdz --gc-content2z --buffer-sizei =z--compression-levelz --no-indexindexT store_false)destrhrerdzFinding adaptersaCParameters -a, -g, -b specify adapters to be removed from each read (or from the first read in a pair if data is paired). If specified multiple times, only the best matching adapter is trimmed (but see the --times option). When the special notation 'file:FILE' is used, adapter sequences are read from the given FASTA file.) descriptionz-az --adaptercSd|fSNbackrBxrBrBrCz%get_argument_parser..appendZADAPTERadapterszSequence of an adapter ligated to the 3' end (paired data: of the first read). The adapter and subsequent bases are trimmed. If a '$' character is appended ('anchoring'), the adapter is only found if it is a suffix of the read.)rkrerhmetavarrprdz-gz--frontcSrrNZfrontrBrurBrBrCrwrxaSequence of an adapter ligated to the 5' end (paired data: of the first read). The adapter and any preceding bases are trimmed. Partial matches at the 5' end are allowed. If a '^' character is prepended ('anchoring'), the adapter is only found if it is a prefix of the read.z-bz --anywherecSrrNZanywhererBrurBrBrCrwrxaSequence of an adapter that may be ligated to the 5' or 3' end (paired data: of the first read). Both types of matches as described under -a and -g are allowed. If the first base of the read is part of the match, the behavior is as with -g, otherwise as with -a. This option is mostly for rescuing failed library preparations - do not use if you know which end your adapter was ligated to!z-ez --error-ratez--errorsEg?zMaximum allowed error rate (if 0 <= E < 1), or absolute number of errors for full-length adapter match (if E is an integer >= 1). Error rate = no. of errors divided by length of matching region. Default: %(default)s (10%%))rkr{rhrdz --no-indelsindelszNAllow only mismatches in alignments. Default: allow both mismatches and indels)rerprhrdz-nz--timesZCOUNTz@Remove up to COUNT adapters from each read. Default: %(default)sz-Oz --overlapZ MINLENGTHzcRequire MINLENGTH overlap between read and adapter for an adapter to be found. Default: %(default)sz--match-read-wildcardsz8Interpret IUPAC wildcards in reads. Default: %(default)sz-Nz--no-match-adapter-wildcardsmatch_adapter_wildcardsz-Do not interpret IUPAC wildcards in adapters.)rerhrprdz--action)trimretainmaskZ lowercasenonerzWhat to do if a match was found. trim: trim adapter and up- or downstream sequence; retain: trim, but retain adapter; mask: replace with 'N' characters; lowercase: convert to lowercase; none: leave unchanged. Default: %(default)s)choicesrhrdz--rcz --revcompreverse_complementzCheck both the read and its reverse complement for adapter matches. If match is on reverse-complemented version, output that one. Default: check only readz --no-trimre store_constr)rpreconstrdz--mask-adapterrzAdditional read modificationsz-uz--cutZLENGTHaRemove bases from each read (first read only if paired). If LENGTH is positive, remove bases from the beginning. If LENGTH is negative, remove bases from the end. Can be used twice if LENGTHs have different signs. This is applied *before* adapter trimming.)rerhrkr{rdz--nextseq-trimz3'CUTOFFzhNextSeq-specific quality trimming (each read). Trims also dark cycles appearing as high-quality G bases.)rkrhr{rdz-qz--quality-cutoffz[5'CUTOFF,]3'CUTOFFa!Trim low-quality bases from 5' and/or 3' ends of each read before adapter removal. Applied to both reads if data is paired. If one value is given, only the 3' end is trimmed. If two comma-separated cutoffs are given, the 5' end is trimmed with the first cutoff, the 3' end with the second.)rhr{rdz--quality-base!NzAssume that quality values in FASTQ are encoded as ascii(quality + N). This needs to be set to 64 for some old Illumina FASTQ files. Default: %(default)sz--lengthz-lzShorten reads to LENGTH. Positive values remove bases at the end while negative ones remove bases at the beginning. This and the following modifications are applied after adapter trimming.z--trim-nzTrim N's on ends of reads.z --length-tagZTAGzSearch for TAG followed by a decimal number in the description field of the read. Replace the decimal number with the correct length of the trimmed read. For example, use --length-tag 'length=' to correct fields like 'length=123'.)r{rdz--strip-suffixzKRemove this suffix from read names if present. Can be given multiple times.z-xz--prefixrDzUAdd this prefix to read names. Use {name} to insert the name of the matching adapter.)rhrdz-yz--suffixz6Add this suffix to read names; can also include {name}z--renameTEMPLATEzfRename reads using TEMPLATE containing variables such as {id}, {adapter_name} etc. (see documentation)z --zero-capz-zz'Change negative quality values to zero.zFiltering of processed readsz|Filters are applied after above read modifications. Paired-end reads are always discarded pairwise (see also --pair-filter).z-mz--minimum-lengthz LEN[:LEN2]z*Discard reads shorter than LEN. Default: 0z-Mz--maximum-lengthz0Discard reads longer than LEN. Default: no limitz--max-nzDiscard reads with more than COUNT 'N' bases. If COUNT is a number between 0 and 1, it is interpreted as a fraction of the read length.z--max-expected-errorsz--max-eeZERRORSz\Discard reads whose expected number of errors (computed from quality values) exceeds ERRORS.z--discard-trimmedz --discardzhDiscard reads that contain an adapter. Use also -O to avoid discarding too many randomly matching reads.z--discard-untrimmedz--trimmed-onlyz-Discard reads that do not contain an adapter.z--discard-casavazBDiscard reads that did not pass CASAVA filtering (header has :Y:).Outputz--quietzPrint only error messages.)rhrerdz--report)fullminimalzAWhich type of report to print: 'full' or 'minimal'. Default: fullz-oz--outputFILEzWrite trimmed reads to FILE. FASTQ or FASTA format is chosen depending on input. Summary report is sent to standard output. Use '{name}' for demultiplexing (see docs). Default: write to standard outputz--fastaz4Output FASTA to standard output even on FASTQ input.z-Zcompression_levelzNUse compression level 1 for gzipped output files (faster, but uses more space))rerrprdz --info-filezoWrite information about each read and its adapter matches into FILE. See the documentation for the file format.z-rz --rest-filez]When the adapter matches in the middle of a read, write the rest (after the adapter) to FILE.z--wildcard-filezyWhen the adapter has N wildcard bases, write adapter bases matching wildcard positions to FILE. (Inaccurate with indels.)z--too-short-outputzdWrite reads that are too short (according to length specified by -m) to FILE. Default: discard readsz--too-long-outputzcWrite reads that are too long (according to length specified by -M) to FILE. Default: discard readsz--untrimmed-outputzbWrite reads that do not contain any adapter to FILE. Default: output to same file as trimmed readszPaired-end optionszrThe -A/-G/-B/-U options work like their -a/-b/-g/-u counterparts, but are applied to the second read in each pair.z-AcSrrrsrBrurBrBrCrw<rx adapters2z43' adapter to be removed from second read in a pair.)rkrprerhr{rdz-GcSrrr|rBrurBrBrCrw?rxz45' adapter to be removed from second read in a pair.z-BcSrrr}rBrurBrBrCrwBrxz65'/3 adapter to be removed from second read in a pair.z-Ucut2z/Remove LENGTH bases from second read in a pair.)rprerhrkr{rdz-pz--paired-outputz$Write second read in a pair to FILE.z--pair-adapterszcTreat adapters given with -a/-A etc. as pairs. Either both or none are removed from each read pair.z --pair-filterz(any|both|first))anyZbothfirstzWhich of the reads in a paired-end read have to match the filtering criterion in order for the pair to be filtered. Default: any)r{rhrrdz --interleavedz/Read and/or write interleaved paired-end reads.z--untrimmed-paired-outputzWrite second read in a pair to this FILE when no adapter was found. Use with --untrimmed-output. Default: output to same file as trimmed readsz--too-short-paired-outputz>Write second read in a pair to this file if pair is too short.)r{rhrdz--too-long-paired-outputz=Write second read in a pair to this file if pair is too long.inputs*)nargsrd)r3r_add_argument_group add_argumentr r intfloat)parsergrouprBrBrCget_argument_parserzs                                            rsc Cs~z dd|dD}Wnty}ztd|d}~wwt|dkr-d|dg}n t|dkr7td |d|dfS) zParse a string INT[,INT] into a pair of integers >>> parse_cutoffs("5") (0, 5) >>> parse_cutoffs("6,7") (6, 7) cSsg|]}t|qSrBr).0valuerBrBrC nz!parse_cutoffs..,z'Quality cutoff value not recognized: {}NrjrrVzJExpected one value or two values separated by comma for the quality cutoff)split ValueErrorr`r]len)rcutoffserBrBrC parse_cutoffses  r.c Cs|d}t|dvrtdz tdd|D}Wnty-}ztd|d}~wwt|dkrG|d durG|d durGtd |t|S) zParse [INT][:[INT]] into a pair of integers. If a value is omitted, use None >>> parse_lengths('25') (25,) >>> parse_lengths('17:25') (17, 25) >>> parse_lengths('25:') (25, None) >>> parse_lengths(':25') (None, 25) :)rjrVz!Only at most one colon is allowedcss$|] }|dkr t|ndVqdS)rDNr)rfrBrBrC s"z parse_lengths..zValue not recognized: {}NrVrrjz8Cannot parse {!r}: At least one length needs to be given)rrr`tuplerr])rfieldsvaluesrrBrBrC parse_lengths{s $r file_opener adapter_namesadapter_names2cCs||jd}||jd}||jd}d}} |jdur)||j|jd\}} d} } |jdur=||j |j d\} } t |j t |j t |jdudkrTtdt|} | ra|j ratd| dkrxd} }d}}t|||\}}}}nT| dkrd|jvrd |jvsJd|jvrd |jvsJd} }d}}t||||\}}}}n$d}}d}}||j|jd\}}||j|jd\} }| dur|} tdid |d |d |d |d| d| d| d|d|d| d|d|d|d|d|d|jS)z Return an OutputFiles instance. If demultiplex is True, the untrimmed, untrimmed2, out and out2 attributes are not opened files, but paths (out and out2 with the '{name}' template). wbNrjzsOnly one of the --discard-trimmed, --discard-untrimmed and --untrimmed-output options can be used at the same time.z1Do not use --discard-trimmed when demultiplexing.normal combinatorial{name1}{name2}restinfowildcard too_short too_short2too_long too_long2 untrimmed untrimmed2outout2demultiplex_outdemultiplex_out2combinatorial_outcombinatorial_out2Z force_fastarB)Z xopen_or_none rest_file info_file wildcard_fileminimum_lengthZ xopen_pairtoo_short_outputtoo_short_paired_outputmaximum_lengthtoo_long_outputtoo_long_paired_outputrdiscard_trimmeddiscard_untrimmeduntrimmed_outputr`determine_demultiplex_modeopen_demultiplex_outoutput paired_outputopen_combinatorial_outuntrimmed_paired_outputr)Zfasta)r>default_outfilerrrrrrrrrrZdemultiplex_moderrrrrrrrrBrBrCopen_output_filess           rcCst}t}|jr g}ndg}|dd|D7}|dd|D7}tt|||D]>\}}|dur5|nd} |dur=|nd} |jd| d| } |jd| d| } || d|||f<|| d|||f<q+|j sp|j rtt d d} }||| |fS) NNNcSsg|]}d|fqSrLrB)rname2rBrBrCrrz*open_combinatorial_out..cSsg|]}|dfqSrLrB)rname1rBrBrCrrunknownrrrzCombinatorial demultiplexing (with {name1} and {name2}) cannot be combined with --untrimmed-output or --untrimmed-paired-output) dictrlist itertoolsproductrrSrxopenrrr`)rrr>rrrextrarrZfname1Zfname2path1path2rrrBrBrCrs(  rc Cst}|jdur tnd}|D]%}|jd|}||d||<|jdur4|jd|}||d||<q|jdd}|jrB|j}|jrHd} n||d} |jdurm|jdd} |jr`|j} |jrfd} n || d} nd} ||| | fS)N{name}rr)rrrrSrrrr) rr>rrrnamerrZuntrimmed_pathrZuntrimmed2_pathrrBrBrCrs0    rcCs|jduo d|jv}|jdur|d|jvkrtd|jduo7|jduo7d|jvo7d|jvo7d|jvo7d|jv}|r@|r@td|rDdS|rHdSd S) z0Return one of "normal", "combinatorial" or FalseNrz_When demultiplexing paired-end data, "{name}" must appear in both output file names (-o and -p)rrz2You cannot combine {name} with {name1} and {name2}rrF)rrr`)r>Z demultiplexZdemultiplex_combinatorialrBrBrCrs* rcCs4t|jp|jp|jp|jp|jp|jp|jp|jS)z> Determine whether we should work in paired-end mode. ) boolr interleavedrr pair_filterrrrr>rBrBrCdetermine_paired,s rrpairedrcCst|dkr tdt|dkr'tdt|dddd|Dd |d}|rA|sAt|d kr9td |d }||fSt|dkrKtd d }||fS)z> Return tuple (input_filename, input_paired_filename) rzIYou did not provide any input file names. Please give me something to do!rVzFYou provided {} input file names, but either one or two are expected. zThe file names were: - z - css|]}d|VqdS)z'{}'N)r])rprBrBrCrJsz$setup_input_files..zE Hint: If your path contains spaces, you need to enclose it in quotesrjzYou used an option that enabled paired-end mode (such as -p, -A, -G, -B, -U), but then you also need to provide two input files (you provided one) or use --interleaved.zIt appears you want to trim paired-end data because you provided two input files, but then you also need to provide two output files (with -o and -p) or use the --interleaved option.N)rr`r]join)rrrinput_filenameinput_paired_filenamerBrBrCsetup_input_files<s<    rcCs|s|jr td|jrtd|rM|jsM|jstd|js#td|j|jdf|j|jdf|j |j dffD]\}}}t |t |krLtdj |d q7|j d krVtd d |jkrdd ksitdtd|jru|jd krwtddSdS)NzQOption --untrimmed-paired-output can only be used when trimming paired-end reads.zFOption --pair-adapters can only be used when trimming paired-end readszWhen a paired-end trimming option such as -A/-G/-B/-U, is used, a second output file needs to be specified via -p (--paired-output).zDWhen you use -p or --paired-output, you must also use the -o option.rz too-shortztoo-longzvWhen trimming paired-end data, you must use either none or both of the --{name}-output/--{name}-paired-output options.rrjzThe overlap must be at least 1.rdz8GC content must be given as percentage between 0 and 100z+--pair-adapters cannot be used with --times)rr` pair_adaptersrrrrrrrrrr]overlap gc_contenttimes)r>rrZ paired_outargnamerBrBrCcheck_argumentsas<     rc Cs|jdkrd|_|r|jdurdn|j}t||}nt|}t|tr3|r'|s3|js0|js0|jr3d|_t ||j |j ||rA|j n|j }|jdurR|t|j|j|jdurit|j}|t|d|d|jt|||||j|j|j|j|j |j t|D]} || q|jr|js|jrtdz|r|t|jn|t |jWnt!y} zt| d} ~ wwdD]1} t"|| } | durt#| } |st$| d krtd |rt$| dkr| d| df} t%|| | q|j&|_&|j'|_'|j(|_(|j)|_)|j|_|S) z Setup a processing pipeline from parsed command-line arguments. If there are any problems parsing the arguments, a CommandLineError is raised. Return an instance of Pipeline (SingleEndPipeline or PairedEndPipeline) rNrTrrjzFOption --rename cannot be combined with --prefix (-x) or --suffix (-y))rrrVz8Two minimum or maximum lengths given for single-end data)*rerr'r& isinstancerrrZoverride_untrimmed_pair_filteradd_unconditional_cutterscutrZadd_bothaddZ nextseq_trimr quality_baseZquality_cutoffrradd_adapter_cutterrrrrenamern)modifiers_applying_to_both_ends_if_pairedrJsuffixr`add_paired_modifierrr r!getattrrrsetattrZmax_nZmax_expected_errorsZdiscard_casavar)r>rrrzrZpair_filter_modepipelineZ pipeline_addrmodifierrattrparamlengthsrBrBrCpipeline_from_parsed_argss        rc Cst|j|j|j|j|jd}z||j}||j}Wnt t t fy/}zt |d}~wwt |t ||jdkrH||D]}|qA||fS)N)Z max_errorsZ min_overlapZread_wildcardsZadapter_wildcardsrrj)rZ error_raterZmatch_read_wildcardsrrZ parse_multirzrFileNotFoundErrorrrr`r debugZ enable_debug)r>Zadapter_parserrzrradapterrBrBrCadapters_from_argss(    rr cut1rcCst||gD]c\}}|s qt|dkrtdt|dkr+|d|ddkr+td|D];}|dkr4q-|dkrY|rJt|tsAJ|t|dq-t|tsQJ|t|q-t|ts`J|dt|q-qdS)NrVz0You cannot remove bases from more than two ends.rrjz0You cannot remove bases from the same end twice.) enumeraterr`rr'rrr&)r rrriZcut_argcrBrBrCrs(  rrrerr add_rc_suffix allow_indexc  Cs |r,|rtdzt|||} Wnty$} ztdt| d} ~ ww|| dSd\} } z|r:t|||| } |rCt|||| } WntyT} zt| d} ~ ww|rk|r]td| sa| ri|| | dSdS| r|rzt| |rudndd}n| }||dSdS)Nz)Cannot use --revcomp with --pair-adaptersz--pair-adapters: rz.--revcomp not implemented for paired-end readsz rc)Z rc_suffix) r`rrstrr rrrr)r rzrrrrerrrrZcutterrZadapter_cutterZadapter_cutter2r rBrBrCrsH  rccs|jdur t|jV|jrtV|jrt|jV|jD]}t|Vq|js-|j r5t |j|j V|j rAt |j dVdSdS)N)r)lengthrZtrim_nrZ length_tagrZ strip_suffixrrJrrZzero_caprr)r>rrBrBrCr5s      rcCsJt}|dkrd|dnd}tdtt|tdd|dS) z'Print the "This is cutadapt ..." headerCPythonz ()rDz$This is cutadapt %s with Python %s%szCommand line parameters: %s N)platformpython_implementationloggerrr python_versionr) cmdlineargsimplementationoptrBrBrC log_headerDs r(cCsttjdddS)z#Entry point for command-line scriptrjNr)mainrZargvrBrBrBrCmain_cliNsr+c CsHt}t}|j|d\}}tjjs#ttt||j |j dk|j dt |t |j}|j r7|j r7|d|rGt||dd||jdkrQ|d|jdkrYtn|j}t|jt|d }tjrt|j st|j stt} nt} t|} | d vsJzP|jot|jd k} t |j| | \} } t!|| t"|\}}t#|| |||}d d |D}dd |D}t$|||||}t%| | | d}t&|||| ||j'|}Wn!t(y}ztj ddd|t)|WYd}~dSd}~wwt*d||d krdndddd |j+z| }|,}Wdn 1swYWnJt-y6t.dtjdt/dYn5t0yDt/d Yn't1j2t1j3t4fyj}ztj dddt/d5|WYd}~nd}~wwt|}|j dkrzt6}nt7}t8t9d||||j:d|durddl;}|<|=|>d?d |S)!z Set up a processing pipeline from the command-line arguments, run it and return a Statistics object. default_outfile is the file to which trimmed reads are sent if the ``-o`` parameter is not used. rr)Z log_to_stderrquietrrz.cSr.rBrr/rBrBrCrr1)rrzCommand line error. Traceback:T)exc_infoNz,Processing reads on %d core%s in %s mode ...rrDz single-endz paired-endZ InterruptedrTzcutadapt: error: {}z%srtime)@r4rparse_known_argsloggingroothandlersr1r#is_any_output_stdoutr,reportrr(setup_profiler_if_requestedZprofiler^warn_if_en_dashesrcoresr-r0restimate_compression_threadsrZr[isattyr.r/rrrrrrrrrr( setup_runner buffer_sizer`rrrrunKeyboardInterruptrYr\BrokenPipeErrordnaioFileFormatErrorUnknownFileFormatEOFErrorr]r#r"logr2rpstatsdisableZStatsZ sort_statsZ print_stats)r%r start_timerr>Z leftover_argsprofilerr>rprogressrZis_interleaved_inputrrrzrr rroutfilesinpathsrunnerrrstatselapsedr;rKrBrBrCr)Ts              r)rQrPrOr>rBc Csfz|dkrt|||||||dWS||}t||||WStjtjtfy2}zt|d}~ww)Nrj)Z n_workersrB)r,openr+rFrHrGOSErrorr`) r rQrPrOr>rBrZinfilesrrBrBrCrAs  rAcCs(|rddl}|}||Sd}|S)Nr)cProfileZProfileenable) requestedrXrNrBrBrCr<sr<cCs$|D] }|drtd|qdS)Nu–uThe first character in argument '%s' is '–' (an en-dash, Unicode U+2013) and will therefore be interpreted as a file name. If you wanted to provide an option, use a regular hyphen '-'.) startswithr#warning)r>argrBrBrCr=s r=cCstdt|dS)Nr)maxr7)r>rBrBrCr?sr?cCsht|jdu|jdk|jdk|jdk|jdk|jdk|jdk|jdk|jdk|j dk|j dk|j dkg S)N-) rrrrrrrrrrrrrrBrBrCr:sr:__main__)er_rZr4r7r!rtypingrrrrrrrargparser r r rFZcutadaptr Zcutadapt.adaptersr rrZcutadapt.parserrZcutadapt.modifiersrrrrrrrrrrrrrrrr r!Zcutadapt.reportr"r#r$Zcutadapt.pipeliner%r&r'r(r)r*r+r,Zcutadapt.utilsr-r.r/r0Z cutadapt.logr1r2 getLoggerr#r3 Exceptionr`rrrrrrrrrrrrrrrrrrr(r+stdoutbufferr)rAr<r=r?r:rMr\rBrBrBrCs!$  L(#l H %%T"  - T