6bc @sddlZddlZddlZddlZddlmZddlmZejj ejj ejj ejj e Z dZejje eZdZddgZdZd Zejd krd Zn%ejd d krdZndZdZdZdZdZdZdZeZ e!Z"e!Z#e!Z$dZ%dZ&dZ'e!Z(dZ)e!Z*da+dZ,dZ-dZ.dZ/dZ0dZ1e!Z2e!Z3eZ4e!Z5eZ6e!Z7e!Z8e!Z9eZ:e!Z;e!Z<e!Z=e!Z>e!Z?e!Z@e!ZAe!ZBe!ZCeZDe!ZEe!ZFe!ZGdZHdZIdZJe!ZKe!ZLe!ZMe!ZNe!ZOe!ZPdZQdZRdZSd ZTe!ZUe!ZVdZWe!ZXd!ZYd"ZZd#ejj[j\d$Z]eZ^d%Z_d&Z`d'Zad(Zbd)Zcd*Zdd+Zed,Zfd-Zgd.Zhd/Zid0Zjd1Zkd2Zld3Zmd4Znd5Zod6Zpd7Zqd8Zre!Zse!Ztd9Zud:d;d<gZvd=Zwd>ZxdZydZzd?Z{dZ|dZ}d@Z~eZdAZdBZdCZdDZdEZdFZdGZdHZdIZdJZdKZdLZdMZdNZdOZd?ZdPZdQZdPZd;ZdRZdSZd?ZdZeZdTZdUZdZdZdVZd:ZdWZd<ZdZdXZdYZdZZd:Ze!Ze!ZeZdOZdZd[Zd\Zd;ZdZd]ZiZgZd^Zd ZdOZdZiZd<ZdPZdZe!Zd_Zd`Zd`ZdOZdaZdbdcdadddedfdgdhdig ZdZdZdZdZeZdZdZdZdZdZdZgZgZgZgZgZgZgZgZgZgZgZgZdZdZdjZdkZdlZddmZddnZe!deejdoZdS(piN(t LooseVersion(tbasenamet quast_libss5https://raw.githubusercontent.com/ablab/quast/master/s2.5+s3.3+tquastt metaquasttDarwintmacosxit64bittlinux_64tlinux_32s0,1000,5000,10000,25000,50000iZis0,300,1500,3000ttonegGz?g?iAgW@gV@idg?tautoii?itquast_test_outputt quast_resultstresults_s%Y_%m_%d_%H_%M_%Stjsontquast_corrected_inputs report.pdftreportttransposed_reporttreport_html_auxscontigs_report_%ssall_alignments_%s.tsvtcontigs_reportstminimap_outputtupper_bound_assemblytquast_downloaded_referencestruns_per_referencetsummaryt not_alignedtcombined_referencet krona_chartsscombined_reference.fastag?sunique_contigs_%s.tsvt reads_statsiii strivial_deletions.bedsstructural_variations.bedit busco_statss icarus.csssbuild_icarus.jsticarus_viewerss icarus.htmlsicarus_menu_templ.htmlsviewers_template.htmlsView in Icarus contig browserscontig_size_viewer.htmlsContig size viewersContig alignment viewertalignment_viewertalignment_viewer_part_salignment_viewer.htmlii2i'g?iiiXi iieiiitANYg?iXiiPi,tpdftemftepstpngtpstrawtrgbatsvgtsvgzcCs-d}tjjd}g}xtD]}|d}d|kr]|jd\}}n-|jdr|d |}}n |}}||}|jd||jddft|t|kot|knr)tSq)Wtj j d|d d d j |d tj d dS(NcSsq|jd}xRttt|D]8}||jr(tt||d|||d7}n|dkrW|d7}ntjj|dtjjdS(NsQUAST vRhs (MetaQUAST mode)Ris (QUAST-LG mode)s (RNRoRfR<tstdoutRCtflush(Rnt full_version((s@/home/psgendb/BIRCHDEV/install/quast-5.2.0/quast_libs/qconfig.pyt print_versionjs       c Cs|dkrt}n|jd|dkr7tnt}|dkrY|jdn)|dkru|jdn |jd|jdtd|dkrttt}}}nt t t }}}|dkrt }nt }|jd|jd tjd d |jd|jd |jd |ry|jd|jd|jd|jdn'|jd|jd|jd|jd||jd|jd|r|jdn|jd|jd|jd|jd|dkre|jd|jd|jd|jdtttfn|jd|jd t|jd!|dkr|jd"n)|dkr|jd#n |jd$|s|jd%n|jd&|jd'|jd(t|jd)|jd*|jd+|sS|jd,n8|jd-|jd.t|jd/|jd0|jd1t|jd2t|r|jd3n|jd4|jd5|jd6||jd7||jd8|jd9t|jd:|jd;|jd<t|ro|jd=|jd>tn|jd?|jd@|jdA|jdB||jdC|jdDt|jdE|jdFt|jdG|jdHt|jdI|jdJ|jdK|jdL|jdMt|jdN|jdO|jdP|jdQttf|jdRt|jdS|jdTt|jdUdVjt|jdW|jdX|jdY|jdZ|jd[|jd\|jd]|jd^|jd_|jd`|jda|jdb|jdc|jdd|jde|jdf|jdg|jdh|jdi|jdj|jd|jdk|jdl|jdm|jdn|jdo|jdp|jdq|jdr|jds|jdt|r|jd|jdu|jdv|jdw|jdx|jdy|r|jdzn|jd{|jd|n|jd|jd}|jd~|rK|jd|jd|jd|jdn'|jd|jd|jd|jd|jd|r|jdn|jd|jd|j dS(NR Rhs=MetaQUAST: Quality Assessment Tool for Metagenome Assemblies Ris>QUAST-LG: Quality Assessment Tool for Large Genome Assemblies s5QUAST: Quality Assessment Tool for Genome Assemblies s Version: s sUsage: python is [options] s Options: sr-o --output-dir Directory to store all result files [default: quast_results/results_] sp-r Comma-separated list of reference genomes or directory with reference genomes sj--references-list Text file with list of reference genome names for downloading from NCBI sr-g --features [type:] File with genomic feature coordinates in the references (GFF, BED, NCBI or TXT) sx Optional 'type' can be specified for extracting only a specific feature type from GFF s8-r Reference genome file sq-g --features [type:] File with genomic feature coordinates in the reference (GFF, BED, NCBI or TXT) sR-m --min-contig Lower threshold for contig length [default: %d] sS-t --threads Maximum number of threads [default: 25% of CPUs] s:These are basic options. To see the full list, use --help sAdvanced options: s-s --split-scaffolds Split assemblies by continuous fragments of N's and add such "contigs" to the comparison s|-l --labels "label, label, ..." Names of assemblies to use in reports, comma-separated. If contain spaces, use quotes s\-L Take assembly names from their parent directory names s_-e --eukaryote Genome is eukaryotic (primarily affects gene prediction) s[ --fungus Genome is fungal (primarily affects gene prediction) s] --large Use optimal parameters for evaluation of large genomes sq In particular, imposes '-e -m %d -i %d -x %d' (can be overridden manually) s-k --k-mer-stats Compute k-mer-based quality metrics (recommended for large genomes) This may significantly increase memory and time consumption on large genomes sT --k-mer-size Size of k used in --k-mer-stats [default: %d] s7 --circos Draw Circos plot sG-f --gene-finding Predict genes using MetaGeneMark sF-f --gene-finding Predict genes using GeneMark-ES s-f --gene-finding Predict genes using GeneMarkS (prokaryotes, default) or GeneMark-ES (eukaryotes, use --eukaryote) sv --mgm Use MetaGeneMark for gene prediction (instead of the default finder, see above) st --glimmer Use GlimmerHMM for gene prediction (instead of the default finder, see above) s| --gene-thresholds Comma-separated list of threshold lengths of genes to search with Gene Finding module s4 [default: %s] sP --rna-finding Predict ribosomal RNA genes using Barrnap s\-b --conserved-genes-finding Count conserved orthologs using BUSCO (only on Linux) sl --operons File with operon coordinates in the reference (GFF, BED, NCBI or TXT) so --est-ref-size Estimated reference size (for computing NGx metrics without a reference) s --max-ref-number Maximum number of references (per each assembly) to download after looking in SILVA database. sZ Set 0 for not looking in SILVA at all [default: %s] s --blast-db Custom BLAST database (.nsq file). By default, MetaQUAST searches references in SILVA database s --use-input-ref-order Use provided order of references in MetaQUAST summary plots (default order: by the best average value) se --contig-thresholds Comma-separated list of contig length thresholds [default: %s] s --x-for-Nx Value of 'x' for Nx, Lx, etc metrics reported in addition to N50, L50, etc (0, 100) [default: %s] s{ --reuse-combined-alignments Reuse the alignments from the combined_reference stage on runs_per_reference stages. sg-u --use-all-alignments Compute genome fraction, # genes, # operons in QUAST v1.* style. sl By default, QUAST filters Minimap's alignments to keep only best ones sQ-i --min-alignment The minimum alignment length [default: %s] sc --min-identity The minimum alignment identity (80.0, 100.0) [default: %.1f] sd-a --ambiguity-usage Use none, one, or all alignments of a contig when all of them sd are almost equally good (see --ambiguity-score) [default: %s] s --ambiguity-score Score S for defining equally good alignments of a single contig. All alignments are sorted s by decreasing LEN * IDY% value. All alignments with LEN * IDY% < S * best(LEN * IDY%) are sa discarded. S should be between 0.8 and 1.0 [default: %.2f] sd --unique-mapping Disable --ambiguity-usage=all for the combined reference run, sk i.e. use user-specified or default ('%s') value of --ambiguity-usage sh --strict-NA Break contigs in any misassembly event when compute NAx and NGAx. sx By default, QUAST breaks contigs only by extensive misassemblies (not local ones) sy-x --extensive-mis-size Lower threshold for extensive misassembly size. All relocations with inconsistency st less than extensive-mis-size are counted as local misassemblies [default: %s] sx --local-mis-size Lower threshold on local misassembly size. Local misassemblies with inconsistency sj less than local-mis-size are counted as (long) indels [default: %s] su --scaffold-gap-max-size Max allowed scaffold gap length difference. All relocations with inconsistency sz less than scaffold-gap-size are counted as scaffold gap misassemblies [default: %s] sy --unaligned-part-size Lower threshold for detecting partially unaligned contigs. Such contig should have se at least one unaligned fragment >= the threshold [default: %s] sq --skip-unaligned-mis-contigs Do not distinguish contigs with >= 50% unaligned bases as a separate group s] By default, QUAST does not count misassemblies in them sx --fragmented Reference genome may be fragmented into small pieces (e.g. scaffolded reference) sy --fragmented-max-indent Mark translocation as fake if both alignments are located no further than N bases s] from the ends of the reference fragments [default: %s] sC Requires --fragmented option sl --upper-bound-assembly Simulate upper bound assembly based on the reference genome and reads s --upper-bound-min-con Minimal number of 'connecting reads' needed for joining upper bound contigs into a scaffold sY [default: %d for mate-pairs and %d for long reads] s --est-insert-size Use provided insert size in upper bound assembly simulation [default: auto detect from reads or %d] s --report-all-metrics Keep all quality metrics in the main report even if their values are '-' for all assemblies or if they are not applicable (e.g., reference-based metrics in the no-reference mode) sT --plots-format Save plots in specified format [default: %s]. s< Supported formats: %s s, se --memory-efficient Run everything using one thread, separately per each assembly. sh This may significantly reduce memory consumption on large genomes s --space-efficient Create only reports and plots files. Aux files including .stdout, .stderr, .coords will not be created. s This may significantly reduce space consumption on large genomes. Icarus viewers also will not be built sk-1 --pe1 File with forward paired-end reads (in FASTQ format, may be gzipped) sk-2 --pe2 File with reverse paired-end reads (in FASTQ format, may be gzipped) s --pe12 File with interlaced forward and reverse paired-end reads. (in FASTQ format, may be gzipped) sj --mp1 File with forward mate-pair reads (in FASTQ format, may be gzipped) sj --mp2 File with reverse mate-pair reads (in FASTQ format, may be gzipped) s --mp12 File with interlaced forward and reverse mate-pair reads (in FASTQ format, may be gzipped) sg --single File with unpaired short reads (in FASTQ format, may be gzipped) s_ --pacbio File with PacBio reads (in FASTQ format, may be gzipped) sh --nanopore File with Oxford Nanopore reads (in FASTQ format, may be gzipped) sm --ref-sam SAM alignment file obtained by aligning reads to reference genome file sm --ref-bam BAM alignment file obtained by aligning reads to reference genome file s --sam Comma-separated list of SAM alignment files obtained by aligning reads to assemblies (use the same order as for files with contigs) s --bam Comma-separated list of BAM alignment files obtained by aligning reads to assemblies (use the same order as for files with contigs) sn Reads (or SAM/BAM file) are used for structural variation detection and sL coverage histogram building in Icarus sX --sv-bedpe File with structural variations (in BEDPE format) sSpeedup options: st --no-check Do not check and correct input fasta files. Use at your own risk (see manual) s8 --no-plots Do not draw plots sS --no-html Do not build html reports and Icarus viewers sB --no-icarus Do not build Icarus viewers sx --no-snps Do not report SNPs (may significantly reduce memory consumption on large genomes) sM --no-gc Do not compute GC% and GC-distribution sy --no-sv Do not run structural variation detection (make sense only if reads are specified) s --no-read-stats Do not align reads to assemblies Reads will be aligned to reference and used for coverage analysis, upper bound assembly simulation, and structural variation detection. Use this option if you do not need read statistics for assemblies. s] --fast A combination of all speedup options except --no-check sHidden options: s0-d --debug Run in a debug mode sI--no-portable-html Do not embed CSS and JS files in HTML report sD-j --save-json Save the output also in the JSON format sF-J --save-json-to Save the JSON output to a particular path s--read-support Use read coverage specified in contig names (SPAdes/Velvet style) for calculating Avg contig read support sR--cov File with read coverage (for Icarus alignment viewer) sV--phys-cov File with physical coverage (for Icarus alignment viewer) sOther: s} --silent Do not print detailed information about each step to stdout (log file is not affected) sw --test Run MetaQUAST on the data from the test_data folder, output to quast_test_output s --test-no-ref Run MetaQUAST without references on the data from the test_data folder, output to quast_test_output. s MetaQUAST will download SILVA 16S rRNA database (~170 Mb) for searching reference genomes sF Internet connection is required ss --test Run QUAST on the data from the test_data folder, output to quast_test_output sz --test-sv Run QUAST with structural variants detection on the data from the test_data folder, sB output to quast_test_output s?-h --help Print full usage message s4-v --version Print version sW --help-hidden Print this usage message with all hidden options s?Online QUAST manual is available at http://quast.sf.net/manual (!RNRoRCRAtFalseRftLARGE_MIN_CONTIGtLARGE_MIN_ALIGNMENTtLARGE_EXTENSIVE_MIS_THRESHOLDtDEFAULT_MIN_CONTIGtDEFAULT_MIN_ALIGNMENTtDEFAULT_EXT_MIS_SIZEt META_MIN_IDYtDEFAULT_MIN_IDYR<Rjtunique_kmer_lent genes_lengthstmax_referencestcontig_thresholdstx_for_additional_Nxtambiguity_usagetambiguity_scoretlocal_misassembly_min_lengthtscaffolds_gap_thresholdtunaligned_part_sizetMIN_CONNECT_MPtMIN_CONNECT_LRtoptimal_assembly_default_IStplot_extensionR5tsupported_plot_extensionsRq( t show_hiddenRntshorttstreamRht m_defaultt i_defaultt x_defaulttmin_idy_default((s@/home/psgendb/BIRCHDEV/install/quast-5.2.0/quast_libs/qconfig.pytusagevsJ                                                                                                                  (tdatetimeR\tplatformR<tdistutils.versionRtos.pathRR]tabspathtdirnametrealpatht__file__R^t PACKAGE_NAMER5t LIBS_LOCATIONt GIT_ROOT_URLR=tLOGGER_DEFAULT_NAMEtLOGGER_META_NAMEtsystemt platform_namet architectureRRRNt min_contigRxR~t ref_fpathRAt prokaryoteRtt is_fungust gene_findingtrna_gene_findingRRtmeta_ambiguity_scoretreuse_combined_alignmentstalignments_for_reuse_dirpathtuse_all_alignmentsRMt min_alignmentRytmin_IDYR|R{testimated_reference_sizet strict_NAtsplit_scaffoldst draw_plotst draw_circost html_reportt save_jsont metagenemarktdebugt portable_htmlttestttest_svt test_no_reftno_checkt no_check_metatunique_mappingtno_gctno_svt no_read_statst show_snpstglimmertis_combined_reftcheck_for_fragmented_reftMAX_REFERENCE_FRAGMENTSRtunaligned_mis_thresholdtall_labels_from_dirst run_buscoRltuse_kmctreport_all_metricstoptimal_assemblytoptimal_assembly_insert_sizeRtoptimal_assembly_min_IStoptimal_assembly_max_IStsilentt is_agb_modeterror_log_fpatht save_errorttest_output_dirnametdefault_results_root_dirnametnowtstrftimetoutput_dirnametmake_latest_symlinktdefault_json_dirnametcorrected_dirnamet plots_fnamet report_prefixttransposed_report_prefixt html_aux_dirtcontig_report_fname_patternticarus_report_fname_patternt detailed_contigs_reports_dirnametaligner_output_dirnametoptimal_assembly_basenametdownloaded_dirnametper_ref_dirnametmeta_summary_dirtnot_aligned_nametcombined_output_namet krona_dirnametcombined_ref_nametdownloaded_ref_min_aligned_ratetunique_contigs_fname_patterntcalculate_read_supporttuse_input_ref_ordertreads_stats_dirnametcoverage_thresholdsttrivial_deletions_fnamet sv_bed_fnamet used_colorstused_lst MAX_PE_ISt cov_fpathtphys_cov_fpatht busco_dirnametcreate_icarus_htmlticarus_css_nameticarus_script_nameticarus_dirnameticarus_html_fnameticarus_menu_template_fnameticarus_viewers_template_fnamet icarus_linktcontig_size_viewer_fnametcontig_size_viewer_nametcontig_alignment_viewer_nametone_alignment_viewer_nametalignment_viewer_part_nametalignment_viewer_fpathtMAX_SIZE_FOR_COMB_PLOTtICARUS_MAX_CHROMOSOMEStmax_contigs_num_for_size_viewertmin_contig_for_size_viewertcontig_len_deltatmin_similar_contig_sizetSHORT_INDEL_THRESHOLDtSPLIT_ALIGN_THRESHOLDRRztextensive_misassembly_thresholdtfragmented_max_indentRwRuRvtupperbound_min_connectionsRRR}tBSS_MAX_SETS_NUMBERtBSS_critical_number_of_alignstBSS_EXTENSIVE_PENALTYtBSS_LOCAL_PENALTYRUtassemblies_numtmemory_efficienttspace_efficientt analyze_gapst min_gap_sizetmin_gene_overlaptALL_FEATURES_TYPEt GC_bin_sizetGC_contig_bin_sizetGC_window_sizetGC_window_size_largetassembly_labels_by_fpathtassemblies_fpathst max_pointstmin_differencetmax_coverage_binstpcnt_covered_for_histogramtdict_of_broken_scaffoldstNs_break_thresholdRtcustom_blast_db_fpathtdownloaded_refstidentity_thresholdt min_lengtht min_bitscoreRRRtoutput_dirpatht referencetgenestoperonstdicttfeaturestlabelst sam_fpathst bam_fpathst reference_samt reference_bamtbedt reads_fpathst forward_readst reverse_readstmp_forward_readstmp_reverse_readstinterlaced_readstmp_interlaced_readst paired_readst mate_pairstunpaired_readst pacbio_readstnanopore_readstreferences_txttjson_output_dirpathRLRXRfRoRsRpR(((s@/home/psgendb/BIRCHDEV/install/quast-5.2.0/quast_libs/qconfig.pyts    6   !   $