U H$xea @s dZddlZddlZddlZddlZddlZddlZddlZddlZddl m Z ddl m Z m Z mZmZmZmZmZmZmZddlmZmZmZddlZddlZddlmZddlmZmZm Z ddl!m"Z"m#Z$dd l%m&Z&dd l'm(Z(m)Z)m*Z*m+Z+m,Z,m-Z-m.Z.m/Z/m0Z0m1Z1m2Z2m3Z3m4Z4m5Z5m6Z6m7Z7m8Z8m9Z9m:Z:dd l;mZ>dd l?m@Z@mAZAdd lBmCZCmDZDddlEmFZFmGZGmHZHddlImJZJmKZKmLZLddlMmNZNmOZOddlPmQZQeRZSGdddeZTGdddeUZVedddZWeXe eYeYfdddZZeXe e eYdfdddZ[eHee eXee eXeGdd d!Z\eeXd"d#d$Z]eeXeeXeXeXe^d%d&d'Z_eeXeXe eXe eXe eXe^d(d)d*Z`e eXe eXeeXe^fd+d,d-Zae^dd.d/ZbeeXe^e^e eXdfd0d1d2Zce^dd3d4d5ZdGd6d7d7Zee eeeefdd8d9ZfeeYeeYe^d:d;d<Zge eXe eXeYe^d=d>d?Zhe^e^e eXeYe^e^e^d@dAdBZiee(ddCdDZjdEdFZkdGdHZlejmjnfe>ddIdJZodKdLZpdMdNZqdOdPZrdQdRZseYe eYdSdTdUZtdVdWZue>eeXeXe eXeYe^evedXdYdZZwexd[kreyeldS)\a cutadapt version {version} Copyright (C) 2010 Marcel Martin and contributors 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)Path) TupleOptionalSequenceListAnyIteratorUnionDictIterable)ArgumentParserSUPPRESS HelpFormatter) __version__)warn_duplicate_adaptersAdapterInvalidCharacter)OneLinedumps)!make_adapters_from_specifications)SingleEndModifierLengthTagModifier SuffixRemoverPrefixSuffixAdder ZeroCapperQualityTrimmerUnconditionalCutter NEndTrimmer AdapterCutterPairedAdapterCutterErrorPairedAdapterCutterNextseqQualityTrimmer ShortenerReverseComplementerPairedEndRenamerRenamerInvalidTemplate PolyATrimmerPairedReverseComplementer) full_reportminimal_report Statistics)SingleEndPipelinePairedEndPipeline)Pipeline run_pipeline) InputPaths OutputFiles FileOpener)available_cpu_countProgress DummyProgress) setup_loggingREPORT)HasNoQualitiescs8eZdZdZGdddeZfddZddZZS)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.CustomUsageHelpFormattercs&tdtj|d<tj||dS)Nhwidth)minshutilget_terminal_sizecolumnssuper__init__selfargskwargs __class__+lib/python3.8/site-packages/cutadapt/cli.pyrAusz8CutadaptArgumentParser.CustomUsageHelpFormatter.__init__NcCs&|tk r"|||df}||j|dS)N)r Z _add_itemZ _format_usage)rCusageZactionsgroupsprefixrDrHrHrI add_usageys z9CutadaptArgumentParser.CustomUsageHelpFormatter.add_usage)N)__name__ __module__ __qualname__rArN __classcell__rHrHrFrICustomUsageHelpFormatterts rScs0|j|d<|ddt|d<tj||dS)NZformatter_classrKz {version})rSreplacerr@rArBrFrHrIrA~s zCutadaptArgumentParser.__init__cCs<tdtjdtdtjd|dd|jd|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|jdvdddwd|jdxdytddzd{dl|jd|ddd}d|jd~ddd|jdd+gdd|jddddd|jddddd|jdddd|jdddddd|jddd&}|jddddddq|jddddddq|jdtddEddl|jddtddddl|jddtddddl|jdddddd|jdddddd|jddddd|d}|jddddd|jdddddW|jdddd|jddddd|jddddd|jdd^dddd|jdddd|jddddd|jdddd|jdddd|jdddd|jdddddq|jddd&}|jddd*dd+gd,ddӍ|jddd*dd+gd,ddӍ|jddd*dd+gd,ddӍ|jddd+gtdzddݍ|jddddodd|jddddd|jdddd|jddddd|jddddd|jdddd|jddddd|jddddd|jddtd|S)NF)rKZadd_helpZOptionsz-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-indexindexTZ store_false)destrhrerdzFinding adaptersaParameters -a, -g, -b specify adapters to be removed from each read (or from R1 if data is paired-end. If specified multiple times, only the best matching adapter is trimmed (but see the --times option). Use notation 'file:FILE' to read adapter sequences from a FASTA file.)Z descriptionz-az --adaptercSsd|fSNZbackrHxrHrHrIz%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.)rkrerhmetavarrordz-gz--frontcSsd|fSNZfrontrHrqrHrHrIrsrtaSequence 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 --anywherecSsd|fSNZanywhererHrqrHrHrIrsrtaSequence 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%%))rkrwrhrdz --no-indelsindelszNAllow only mismatches in alignments. Default: allow both mismatches and indels)rerorhrdz-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.)rerhrordz--action)trimZretainmaskZ lowercasenoner~zWhat 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-trimreZ store_constr)roreconstrdz--mask-adapterrzAdditional read modificationsz-uz--cutZLENzRemove LEN bases from each read (or R1 if paired; use -U option for R2). If LEN is positive, remove bases from the beginning. If LEN is negative, remove bases from the end. Can be used twice if LENs have different signs. Applied *before* adapter trimming.)rerhrkrwrdz--nextseq-trimz3'CUTOFFzhNextSeq-specific quality trimming (each read). Trims also dark cycles appearing as high-quality G bases.)rkrhrwrdz-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.)rhrwrdz--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--poly-azTrim poly-A tailsz--lengthz-lZLENGTHzShorten 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'.)rwrdz--strip-suffixzKRemove this suffix from read names if present. Can be given multiple times.z-xz--prefixrJzUAdd 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--max-average-error-ratez --max-aerZ ERROR_RATEzcas --max-expected-errors (see above), but divided by length to account for reads of varying length.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:).ZOutputz--quietzPrint only error messages.)rhrerdz--report)ZfullminimalzAWhich type of report to print: 'full' or 'minimal'. Default: fullz--jsonZFILEz"Dump report in JSON format to FILEz-oz--outputzWrite 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))rerrordz --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 optionsznThe -A/-G/-B/-U/-Q options work like their lowercase counterparts, but are applied to R2 (second read in pair)z-AcSsd|fSrprHrqrHrHrIrs^rt adapters2z 3' adapter to be removed from R2)rkrorerhrwrdz-GcSsd|fSrxrHrqrHrHrIrsartz 5' adapter to be removed from R2z-BcSsd|fSryrHrqrHrHrIrsdrtz"5'/3 adapter to be removed from R2z-Ucut2zRemove LENGTH bases from R2)rorerhrkrwrdz-Qquality_cutoff2z7Quality-trimming cutoff for R2. Default: same as for R1)rorhrwrdz-pz--paired-outputzWrite R2 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-filter)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)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.)rwrhrdz--too-long-paired-outputz=Write second read in a pair to this file if pair is too long.inputs*)nargsrd)r9r`Zadd_argument_group add_argumentrr intfloat)parsergrouprHrHrIget_argument_parsers                                              r)srcc Cszdd|dD}Wn0tk rH}ztd|W5d}~XYnXt|dkrdd|dg}nt|dkrxtd |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|qSrHr).0valuerHrHrI sz!parse_cutoffs..,z%Quality cutoff value not recognized: NrjrrWzJExpected one value or two values separated by comma for the quality cutoff)split ValueErrorralen)rZcutoffserHrHrI parse_cutoffss   r.c Cs|d}t|dkrtdztdd|D}Wn0tk rd}ztd|W5d}~XYnXt|dkr|d dkr|d dkrtd |d 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) :)rjrWz!Only at most one colon is allowedcss"|]}|dkrt|ndVqdS)rJNr)rfrHrHrI sz parse_lengths..zValue not recognized: NrWrrjzCannot parse 'z(': At least one length needs to be given)rrratupler)rZfieldsvaluesrrHrHrI parse_lengthss   $ r) file_opener adapter_namesadapter_names2rccCs\t|j|j|j|j|j|j|j|j|j |j |j g | |jd}| |jd}| |jd}d}} |j dk r||j|jd\}} d} } |jdk r||j|jd\} } t|jt|jt|jdk dkrtdt|j |j } | r|jrtd| dkr|jrtd| dkrXd} }d}}t||j |j |j|j |j|\}}}}n| dkrd |j krzd |j ks~td |j krd |j kst|js|j rtd d} }d}}d}}t|||j |j |j|\}}nJd}}d}}||j|j d\}}||j |j d\} }| dkr2|} t||||| | | ||| ||||||jd S) 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. combinatorialzWith --pair-adapters, you can only use {name} in your output file name template, not {name1} and {name2} (no combinatorial demultiplexing).normal{name1}{name2}zCombinatorial demultiplexing (with {name1} and {name2}) cannot be combined with --untrimmed-output or --untrimmed-paired-output)restinfowildcard too_short too_short2too_long too_long2 untrimmed untrimmed2outout2demultiplex_outdemultiplex_out2combinatorial_outcombinatorial_out2Z force_fasta)complain_about_duplicate_paths rest_file info_file wildcard_filetoo_short_outputtoo_short_paired_outputtoo_long_outputtoo_long_paired_outputuntrimmed_outputuntrimmed_paired_outputoutput paired_outputZ xopen_or_noneminimum_lengthZ xopen_pairmaximum_lengthrdiscard_trimmeddiscard_untrimmedradetermine_demultiplex_mode pair_adaptersopen_demultiplex_outAssertionErroropen_combinatorial_outr1Zfasta)rDdefault_outfilerrrrrrrrrrZdemultiplex_moderrrrrrrrrHrHrIopen_output_filess         r)pathscCsrtjdkrtjdkrdSt}|D]J}|dkr0q"t|}|rJ|sJq"||krbtd|d||q"dS)NZwin32)r|zPath zQ specified more than once as an output file. This is not supported at the moment.) rZplatform version_infosetrexistsis_fileraadd)rseenpathprHrHrIr;s r)rroutput_templatepaired_output_templatercCst}t}|rg}n*dg}|dd|D7}|dd|D7}tt|||D]x\} } | dk rh| nd} | dk rx| nd} |d| d| } |d| d| }|| d|| | f<||d|| | f<qT||fS) NNNcSsg|] }d|fqSNrH)rname2rHrHrIr^sz*open_combinatorial_out..cSsg|] }|dfqSrrH)rname1rHrHrIr_sunknownrrr)dictlist itertoolsproductrTxopen)rrrrrrrrZextrarrZfname1Zfname2path1path2rHrHrIrOs$ r)r template1 template2rrrcCst}|dk rtnd}|D]P} |d| } || d|| <|dk r|dk sPt|d| } || d|| <q|dd} |r|} |rd} n || d} |dk r|dd}|r|}|rd}q||d}nd}||| |fS)N{name}rr)rrTrr)rrrrrrrrrnamerrZuntrimmed_pathrZuntrimmed2_pathrrHrHrIrls2       r)rrrccCs|dk od|k}|dk r,|d|kkr,td|dk oZ|dk oZd|koZd|koZd|koZd|k}|rl|rltd|rtdS|r|dSd SdS) 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)ra)rrZ demultiplexZdemultiplex_combinatorialrHrHrIrs.rcCs:t|jp6|jp6|jp6|jp6|jp6|jp6|jp6|jp6|j S)z> Determine whether we should work in paired-end mode. ) boolr interleavedrr pair_filterrrrrrDrHrHrIdetermine_paireds$r)rpairedrrccCst|dkrtdn:t|dkrPtdt|dddd|Dd |d}|r|st|d krvtd q|d }nt|dkrtd d}|r||fS|fSdS) NrzIYou did not provide any input file names. Please give me something to do!rWzFYou provided {} input file names, but either one or two are expected. zThe file names were: - z - css|]}d|dVqdS)'NrH)rrrHrHrIrsz$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.)rraformatjoin)rrrZinput_filenameZinput_paired_filenamerHrHrIsetup_input_filess>     r)rrccCs|s |jrtd|jr td|r|js|js8td|jsFtd|j|jdf|j|jdf|j |j dffD]*\}}}t |t |krntdj |d qn|j d krtd d |jkrd ksntd|jr|jd krtddS)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)rrarrrrrrrrrrroverlap gc_contenttimes)rDrrZ paired_outargnamerHrHrIcheck_argumentssH     rc@s"eZdZddZedddZdS) PipelineMakercCs2||_|jdkrdn|j|_||_||_||_dS)Nr)rDrerrvr)rCrDrrvrrHrHrIrA"s zPipelineMaker.__init__rbc Csg}|t|jj|jj|j|jjdk rdt|jj|jj}|jrZ| |t |fn | ||t |jj |jj |jj|j|t|j|j|j|jj|j|jj|jj|jj |jj |jjr|jr| ttddfn | tt|jD],}|jr | |t |fq| |q|jjrR|jjsJ|jjrRtd|jjr|jjdkrz.|jrt|jjn t|jj}| |Wn,tk r}z t|W5d}~XYnX|jr|jj dkrdn|jj }t!||}nt"|}t#|t!r<|jr|js<|jj$s6|jj%s6|jj&r.rrj)rrr)rrrrZ qtrimmersrHrrIr s   r )rrrerr add_rc_suffix allow_indexc cs|r\|rtdzt|||} Wn2tk rR} ztdt| W5d} ~ XYnX| Vnd\} } z(|rxt||||} |rt||||} Wn*tk r} z t| W5d} ~ XYnX|r| s| r|rt| | |rdnddVn | | fVn,| r|rt| |rdnddVn| VdS)Nz)Cannot use --revcomp with --pair-adaptersz--pair-adapters: rz rc)Z rc_suffix)rar rstrrrr(r#) rvrrrrerrr r!ZcutterrZadapter_cutterZadapter_cutter2rHrHrIr s@ "   r ccs|jdk rt|jV|jr$tV|jr6t|jV|jD]}t|Vq<|jsX|j rht |j|j V|j r|t |j dVdS)Nr)Zlengthr"Ztrim_nrZ length_tagrZ strip_suffixrrMr rZzero_caprr)rDr rHrHrIr s      r cCsJt}|dkrd|dnd}tdtt|tdd|dS) z'Print the "This is cutadapt ..." headerZCPythonz ()rJz$This is cutadapt %s with Python %s%szCommand line parameters: %s N)rZpython_implementationloggerrrpython_versionr) cmdlineargsimplementationZoptrHrHrI log_header"sr)cCstttjdddS)z#Entry point for command-line scriptrjNr)multiprocessingZfreeze_supportmainrZargvrHrHrHrImain_cli0sr-c Cst}t}|j|d\}}tjjsFttt||j |j dk|j dt |t |j}t|j rt|j rt|d|rt||dd||jdkr|d|jdkrtn|j}t|jt|d }tjr|j s|j st} nt} t|} z|jot|j d k} t!|j | | } t"|| t#|\} }t$| | rF|nd t%|| | |&}d d | D}dd |D}t'|||||}t(| d| i}t)dddd|j*||d krdndt+||||| |j,}Wnt-k r|j rnt.dtjdt/dYnt0k r0t/d Ynlt1t2t3t4j5t4j6t7fk r}zd k rRt?|j>dZ}t@|||jAdt|jAd kr|jAd nd || |j=dd }|BtC||Bd!W5QRX|d k rdd lD}|E|F|Gd"Hd#|S)$z Set up a processing pipeline from the command-line arguments, run it and return a Statistics object. default_outfile is the file to which trimmed reads are sent if the ``-o`` parameter is not used. rr)Z log_to_stderrquietrrz.cSsg|] }|jqSrHrr/rHrHrIrnsrz$Processing %s reads on %d core%s ...z single-endz paired-end)FTrrJZ InterruptedrUzCommand line error. Traceback:T)exc_infoz%srWgY@w)statsr'rrcoresrrrXtime)Ir6rZparse_known_argsloggingrootZhandlersr6r%is_any_output_stdoutr.reportrr)setup_profiler_if_requestedZprofilelog_system_infor_warn_if_en_dashesrr5r3r2restimate_compression_threadsrZr[isattyr4r5rrrrrrr log_adaptersrrrr0rrr/Z buffer_sizeKeyboardInterruptrYr\BrokenPipeErrorOSErrorEOFErrorr8dnaioZUnknownFileFormatZFileFormatErrorrarr*r)logr7rZjsonopen json_reportrwrite json_dumpspstatsdisableZStatsZ sort_statsZ print_stats)r'rZ start_timerrDZ leftover_argsprofilerr5rZprogressrZis_interleaved_inputZ input_pathsrvrrrrZoutfilesZinpathsr4rZ exit_codeelapsedr;rZ json_dictrLrHrHrIr+7s             r+cCs.tdtjtdtjtdtjdS)NzPython executable: %szdnaio version: %szxopen version: %s)r%rrZ executablerFrrrHrHrHrIr=sr=cCs|dk }t|rdndt|t|dD]}td|q,t|dkr^tdt|d|rtdt|t|dD]}td|q~t|dkrtdt|ddS)NzR1 adapters (%d):zAdapters (%d):r7z- %sz - (%d more)zR2 adapters (%d):)r%rrrislice)rvrrr0rHrHrIrAs  rAcCs&|rddl}|}|nd}|S)Nr)cProfileZProfileenable)Z requestedrRrNrHrHrIr<s  r<cCs$|D]}|drtd|qdS)Nu–uThe first character in argument '%s' is '–' (an en-dash, Unicode U+2013) and will therefore be interpreted as a file name. If you wanted to provide an option, use a regular hyphen '-'.) startswithr%Zwarning)rDargrHrHrIr>s  r>)r5rccCstdt|ddS)Nrrj)maxr<)r5rHrHrIr?sr?cCsht|jdk|jdk|jdk|jdk|jdk|jdk|jdk|jdk|jdk|j dk|j dk|j dkg S)N-) rrrrrrrrrrrrrrHrHrIr:sr:)r4r'rrr5rrrcc Cs@dtddgtt|||||dd}||j|dd|S)NzCutadapt reportrr|)rrr)tagZschema_versionZcutadapt_versionr&Zcommand_line_argumentsr5inputT)Zone_line)rrrr&updateZas_json)r4r'rrr5rrdrHrHrIrIs   rI__main__)zr`rrZr6r=r8rrr*pathlibrtypingrrrrrrr r r argparser r rrFrZcutadaptrZcutadapt.adaptersrrrZ cutadapt.jsonrrrKZcutadapt.parserrZcutadapt.modifiersrrrrrrrrrrr r!r"r#r$r%r&r'r(Zcutadapt.reportr)r*r+Zcutadapt.pipeliner,r-Zcutadapt.runnersr.r/Zcutadapt.filesr0r1r2Zcutadapt.utilsr3r4r5Z cutadapt.logr6r7Zcutadapt.qualtrimr8Z getLoggerr%r9 Exceptionrarr"rrrrrrrrrrrrrrrr r r r)r-stdoutbufferr+r=rAr<r>r?r:rrIrOr\rHrHrHrIs  ,  T #v     )  *1|  0v