tbtools-cli 1.2.0__py3-none-any.whl

tbtools_cli/auto_commands.py

@@ -0,0 +1,1149 @@
+"""auto_commands：命令实现层（表驱动化 v2）。
+
+144 个同构命令（86 bridge + 58 direct Java 调用）由 ENGINE_REGISTRY 数据驱动生成，
+13 个特殊实现手写保留（hmmsearch 转发 / gxfAttr 原生 / kallisto 二进制 / fimo 二进制 /
+notung 插件 / newickRename 插件 / hmmerSearch / memeViz / gsea / tfbsShift /
+mcscanxd / quickAnno / smart —— 各含独立预检/环境/参考数据逻辑）。
+
+cli.py 通过 dir(_ac) 反射 _xxx_impl 名字注册命令，函数形态必须保留。
+doc 值为旧模块运行时 __doc__（已含编译器 docstring 处理后的真实字符）。
+"""
+import logging
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+
+from tbtools_cli.core import BUILD_DIR, JAR, ROOT, cp, ensure_bridge, run_java, run_plot, safe_temp
+
+
+# ── ENGINE_REGISTRY ──────────────────────────────────────────────
+# (命令名, kind, 类名, xmx, runner, doc)
+#   kind: bridge = 本地桥（ensure_bridge + cp(BUILD_DIR,JAR)）
+#         direct = 直连主 jar 引擎类（cp JAR + biocjava.* 全限定名）
+#   runner: plot = run_plot（xvfb 包装）/ java = run_java
+ENGINE_REGISTRY = [
+    ('admixture', 'bridge', 'AdmixtureCli', '3g', 'plot', 'admixture: admixture <qFiles.lst> <out> [sampleIDFile] [groupFile] [sor'),
+    ('admixtureViz', 'bridge', 'AdmixtureCli', '3g', 'plot', 'admixtureViz: admixtureViz <q1.txt> <q2.txt> [<q3.txt>...] <out.svg> [--id samples.txt] [--group group.txt] [--sort Qraito|Lexical|None]   # ADMIXTURE Q 矩阵可视化（GUI 逆向接口；Q 文件纯数值矩阵，样本 ID 单独 --id）'),
+    ('amazingmeta', 'bridge', 'AmazingMetaCli', '3g', 'plot', 'amazingmeta: amazingmeta <meme.xml> <newick.treefile> <out.svg|png|pdf> ['),
+    ('annocompare', 'bridge', 'StructAnnoCompareCli', '3g', 'plot', 'annocompare: annocompare <before.gff3> <after.gff3> <outDir> [runName] [r'),
+    ('bamMerge', 'direct', 'biocjava.bioDoer.GenomeAnnotation.BAMMergeByRegionCoverage', '3g', 'plot', 'bamMerge: bamMerge <gtf> <bamDir> <outDir>   # 按区域覆盖合并 BAM（多样本择优）'),
+    ('bamindex', 'bridge', 'BamIndexCli', '3g', 'plot', 'bamindex: bamindex <in.sorted.bam> [out.bai]'),
+    ('bamsort', 'bridge', 'BamSortCli', '3g', 'plot', 'bamsort: bamsort <in.bam> <out.bam> [sortOrder] [tmpDir]'),
+    ('bamstate', 'bridge', 'BamStateCli', '3g', 'plot', 'bamstate: bamstate <out.tsv> <gff3> <bam1> [<bam2> ...]'),
+    ('barplot', 'bridge', 'BarplotCli', '3g', 'plot', 'barplot: barplot <enrichment.tsv> <out> <termCol> <pvalCol> [classCol'),
+    ('barplotter', 'bridge', 'BarPlotterCli', '3g', 'plot', 'barplotter: barplotter -g <gff> -s <synteny> -c <ctl> -o <out.png>'),
+    ('batchReplace', 'direct', 'biocjava.bioDoer.Table.BatchStringReplace', '3g', 'plot', 'batchReplace: batchReplace <inFile> <outFile> <patternMap.tsv> [--partial]'),
+    ('calcRepeat', 'bridge', 'CalcRepeatCli', '3g', 'plot', 'calcRepeat: calcRepeat <genome.fa> <outRepeat.txt> [--kmerSize N] [--min'),
+    ('cddmotif', 'bridge', 'CddMotifCli', '3g', 'plot', 'cddmotif: cddmotif <cdd.hitdata.txt> <in.fasta> <out.svg|png|pdf> [new'),
+    ('circlegene', 'bridge', 'CircleGeneViewerCli', '3g', 'plot', 'circlegene: circlegene <gff> <geneID.txt> <out> [--rename f --link f --r'),
+    ('circos', 'bridge', 'CircosCli', '3g', 'plot', 'circos: circos <chrLen.txt> <link.txt> <genePos.txt> <outFile> [w] ['),
+    ('collinearRegion', 'direct', 'biocjava.bioDoer.ComparativeGenomics.MCScanX.CollinearityToRegion', '3g', 'plot', 'collinearRegion: collinearRegion <in.collinearity> <simGff> <out.txt>'),
+    ('colorscheme', 'bridge', 'ColorSchemeCli', '3g', 'plot', 'colorscheme: colorscheme <inTab> <outTab> <refColIndex>'),
+    ('conflictpaf', 'direct', 'biocjava.bioDoer.GenomeAssembly.CalculateConflictByRefAlignPAF', '3g', 'plot', 'conflictpaf: conflictpaf <in.paf> <out.tsv> [binSize]'),
+    ('ctgGroup', 'bridge', 'CtgGroupCli', '3g', 'plot', 'ctgGroup: ctgGroup <in.miniprot.gff> <polyPoid> <outContigGrpMap>'),
+    ('cubeheatmap', 'bridge', 'CubeHeatmapCli', '3g', 'plot', 'cubeheatmap: cubeheatmap <expr.tsv> <group.tsv> <out> [--log10 --minColor <c> --midColor <c> --maxColor <c>]   # 3D 立方体热图（N4: group.tsv 引擎对列数有严格假设，官方 cube_group.tsv 仍会 ArrayIndexOutOfBounds——引擎缺陷；喂分组格式: 行=样本/基因，列数须与引擎预期一致，建议先 tbtools check）'),
+    ('degramdom', 'bridge', 'DegramdomCli', '3g', 'plot', 'degramdom: degramdom <in.tsv> [out.nwk]'),
+    ('distance', 'bridge', 'DistanceCli', '3g', 'plot', 'distance: distance <in.tsv> <col1> <col2> <euclidean|pearson|pearsonDi'),
+    ('dotplot', 'direct', 'biocjava.bioDoer.JIGplotToolkit.DotPlot.dotdotdot', '3g', 'plot', 'dotplot: dotplot --inGff <gff> --genePair <pairs> --chrLayout <layout'),
+    ('dualsyn', 'bridge', 'DualSynCli', '3g', 'plot', 'dualsyn: dualsyn <simplifiedGff> <collinearity> <out> [--chr1 "1,2"] '),
+    ('efpHeat', 'direct', 'biocjava.bioDoer.SimpleEfpBrowser.generateSuperHeatMap', '3g', 'plot', 'efpHeat: efpHeat <inTGA> <sample2cc.txt> <expMat.tsv> <geneId> <out.s'),
+    ('eggnog', 'bridge', 'EggnogCli', '4g', 'java', 'eggnog: eggnog <in.fa> -o <prefix> --output_dir <outDir> --data_dir <eggNOGdb> [--cpu N] [--evalue 0.001]   # eggNOG 直系同源注释（GUI 逆向接口 EmapperPipeline；⚠️ 需先就位 eggNOG 数据库）'),
+    ('exprCorr', 'bridge', 'ExprCorrCli', '3g', 'plot', 'exprCorr: exprCorr <inFPKM> <outCorrMat>'),
+    ('fastaExtract', 'direct', 'biocjava.bioDoer.Fasta.ExtractFasta', '3g', 'plot', 'fastaExtract: fastaExtract <in.fa> <idList.txt> <out.fa> [--mode Match|Con'),
+    ('fastaSubseq', 'direct', 'biocjava.bioDoer.Fasta.ExtractFastaSubseq', '3g', 'plot', 'fastaSubseq: fastaSubseq <in.fa> <pos.txt> <out.fa>   # 按坐标提子序列（第92引擎，Ext'),
+    ('filesplit', 'bridge', 'FileSplitCli', '3g', 'plot', 'filesplit: filesplit <inFile> <numParts>'),
+    ('filterCScore', 'direct', 'biocjava.bioDoer.BLAST.FilterBlastResultByCScore', '3g', 'plot', 'filterCScore: filterCScore <in.blast.tab6> <out.tab6> [--cscore 0.5]'),
+    ('findblockdual', 'bridge', 'FindBlockDualCli', '3g', 'plot', 'findblockdual: findblockdual <queryGenome.fa> <query.gff> <subjectGenome.fa'),
+    ('findblockmultiple', 'bridge', 'FindBlockMultipleCli', '3g', 'plot', 'findblockmultiple: findblockmultiple <queryGenome.fa> <query.gff> <queryId> <ou'),
+    ('findpath', 'bridge', 'FindPathCli', '3g', 'plot', 'findpath: findpath --inGffArr <gff1,gff2,...> --inGenePairs <pairs> --'),
+    ('fqTrim', 'direct', 'biocjava.bioDoer.Fastq.FastqParallelTrimmer', '3g', 'plot', 'fqTrim: fqTrim <in.fq> <out.fq> [--b5 N] [--b3 N] [--threads N]'),
+    ('fqfaConv', 'direct', 'biocjava.bioDoer.LinuxPipe.FastqAndFasta', '3g', 'plot', 'fqfaConv: fqfaConv <input> <output> <fq2fa|fa2fq>   # FASTQ/FASTA 互转（第'),
+    ('gbar', 'bridge', 'GroupedBarCli', '3g', 'plot', 'gbar: gbar <data.tsv> <out.svg> [--header|--no-header] [--errorbar SEM|SD|CI95] [--plot BAR_ERROR|BOXPLOT|VIOLIN|SWARM] [--homoscedastic-t]   # 分组柱状图+显著性标注（GUI 逆向接口 buildPanel；数据=每行 group value）'),
+    ('gdensity', 'bridge', 'GeneDensityCli', '3g', 'java', 'gdensity: gdensity <in.gff3> <out.geneRecords> <binSize> [--feature <tag>] [--chrlen <file>]   # 基因密度 bin 分析（GUI 逆向接口 GeneDensityProfiler）'),
+    ('gel', 'direct', 'biocjava.bioDoer.JIGplotToolkit.GelImage.Marker', '3g', 'plot', 'gel: gel <FragmentRangeArr> <LaneLabels> <MarkerRange> <out>   # 凝胶电泳图（GelImage.Marker；⚠️ 无参调用会挂起 N32）'),
+    ('genedensity', 'bridge', 'GeneDensityCli', '3g', 'plot', 'genedensity: genedensity <in.gff3> <out.tsv> [binSize]'),
+    ('genelocation', 'direct', 'biocjava.bioDoer.JIGplotToolkit.GeneLocation.GeneLocation', '3g', 'plot', 'genelocation: genelocation --ChrLen <chrlen> --FeaturePos <pos> --OutGraph'),
+    ('genelocgff', 'bridge', 'GeneLocGffCli', '3g', 'plot', 'genelocgff: genelocgff <gff3> <idList> <out> [--chrLen len.tsv] [--renam'),
+    ('generic', 'bridge', 'GenericCli', '3g', 'plot', 'generic: generic <engineClass> <method[+method2]> <out> [--set field '),
+    ('gfa', 'bridge', 'VizGFACli', '3g', 'plot', 'gfa: gfa <in.gfa> <out> [width] [height]'),
+    ('gfa2fa', 'direct', 'biocjava.bioDoer.Fasta.Tools.GFAtoFasta', '3g', 'plot', 'gfa2fa: gfa2fa <in.gfa> <out.fa>   # GFA 组装图 → FASTA（第91引擎，GFAtoFast'),
+    ('goEnrich', 'bridge', 'GoEnrichCli', '4g', 'java', 'goEnrich: goEnrich <go.obo> <gene2go.tsv> <selectGenes.txt> <outDir>   # GO 富集分析（MF/CC/BP，P+BH 校正，G4 补齐）'),
+    ('goParse', 'direct', 'biocjava.bioDoer.GeneOntology.littleTools.GOtermParser', '3g', 'plot', 'goParse: goParse <gene2Go.txt> <oboFile> [--level N]   # GO 词典解析（第103'),
+    ('golevel', 'bridge', 'GoLevelCli', '3g', 'plot', 'golevel: golevel <go.obo> <gene2go.tsv> <outPrefix> [--level N] [--graph] [--width W] [--height H]   # GO 层级统计+柱状图（GUI 逆向接口；统计表纯逻辑，图需 xvfb）'),
+    ('groupCol', 'direct', 'biocjava.bioDoer.Table.TableColCollaspe', '3g', 'plot', 'groupCol: groupCol <inTable.tsv> <inGrpInfo.tsv> <outTable> [Sum|Mean|'),
+    ('groupedbar', 'bridge', 'GroupedBarCli', '3g', 'plot', 'groupedbar: groupedbar <data.tsv> <out> [plotType] [errorBarType] [hasHeader] [title]   # 分组柱状图（N5: 数据格式=每行 <group>\t<value>（重复行成组），非常规基因×样本矩阵；矩阵输入会在 GroupedBarRawData.load 崩溃——引擎缺陷）'),
+    ('gsadiag', 'bridge', 'GsaDiagCli', '3g', 'plot', 'gsadiag: gsadiag <in.fixed.gff3> <out.stat.xls> [genome.fasta] [relax'),
+    ('gxfAppend', 'direct', 'biocjava.bioDoer.GXFUtils.GxfIDAppender', '3g', 'plot', 'gxfAppend: gxfAppend <in.gff3> <out.gff3> <prefix>   # GFF seqid+ID 加前缀'),
+    ('gxfFix', 'direct', 'biocjava.bioDoer.GXFUtils.GXFfixer.GXFFix', '3g', 'plot', 'gxfFix: gxfFix <in.gff3> <out.gff3>   # GFF 修复（重复ID前缀/CDS phase/dang'),
+    ('gxfGenepos', 'direct', 'biocjava.bioDoer.GXFUtils.GXFToGenePosFile', '3g', 'plot', 'gxfGenepos: gxfGenepos <in.gff3> <outGenepos> <outChrLen> [feature]  # G'),
+    ('gxfMatch', 'direct', 'biocjava.bioDoer.GXFUtils.GxfGenomeMatch', '3g', 'plot', 'gxfMatch: gxfMatch <in.gff3> <inGenome.fa>'),
+    ('gxfOverlap', 'direct', 'biocjava.bioDoer.GXFUtils.GXFOverlaper', '3g', 'plot', 'gxfOverlap: gxfOverlap <in.gff3> <region.txt> <out.gff3> [--ignoreStrand'),
+    ('gxfRecall', 'direct', 'biocjava.bioDoer.GXFUtils.RecallmRNAFeature', '3g', 'plot', 'gxfRecall: gxfRecall <in.gff3> <out.gff3>   # 从 gene 行恢复 mRNA 特征（第82引擎，'),
+    ('gxfRegion', 'direct', 'biocjava.bioDoer.GXFUtils.GXFRegionSummary', '3g', 'plot', 'gxfRegion: gxfRegion <in.gff3> <region.txt> <out.gff3> [--ignoreStrand]'),
+    ('gxfRename', 'direct', 'biocjava.bioDoer.GXFUtils.GXFRenamer', '3g', 'plot', 'gxfRename: gxfRename <in.gff3> <out.gff3> <renameMap.tsv>'),
+    ('gxfRepGXF', 'direct', 'biocjava.bioDoer.GXFUtils.GXFToRepresentativeGXF', '3g', 'plot', 'gxfRepGXF: gxfRepGXF <in.gff3> <out.gff3> [--featureID CDS] [--attachID'),
+    ('gxfRepIDs', 'direct', 'biocjava.bioDoer.GXFUtils.GXFToRepresentativeIDs', '3g', 'plot', 'gxfRepIDs: gxfRepIDs <in.gff3> <out.txt>'),
+    ('gxfStat', 'direct', 'biocjava.bioDoer.GXFUtils.GXFfixer.GXFstat', '3g', 'plot', 'gxfStat: gxfStat <in.gff3> <outStat.xls>   # GFF 统计（基因/mRNA/外显子/内含子/C'),
+    ('gxffilter', 'bridge', 'GxfFilterCli', '3g', 'plot', 'gxffilter: gxffilter <in.gff3|gtf> <idList.txt> <out.gff3|gtf>'),
+    ('gxfsort', 'bridge', 'GxfSortCli', '3g', 'plot', 'gxfsort: gxfsort <in.gff3|gtf> <out.sorted>'),
+    ('hicEnzyme', 'direct', 'biocjava.bioDoer.GenomeAssembly.HiCRestrictionEnzymePrediction', '3g', 'plot', 'hicEnzyme: hicEnzyme <inHiC.fastq>   # HiC 限制酶预测（第76引擎）'),
+    ('hmmExtract', 'direct', 'biocjava.bioDoer.LinuxPipe.hmmInfoExtracter', '3g', 'plot', 'hmmExtract: hmmExtract <in.hmm> <idList.txt> <out.hmm>   # 从 HMM 文件按 NAM'),
+    ('homoPhase', 'direct', 'biocjava.bioDoer.GenomeAssembly.HomoConflictBasedPartition', '3g', 'plot', 'homoPhase: homoPhase <inContigGrpMap> <outPhasedMap>'),
+    ('keggEnrich', 'bridge', 'KeggEnrichCli', '4g', 'java', 'keggEnrich: keggEnrich <reference.keg> <annotation.tsv> <selectIds.txt> <out.xls>   # KEGG 富集分析（G4 补齐，需真实 .keg 参考文件）'),
+    ('layoutheatmap', 'bridge', 'LayoutHeatmapCli', '3g', 'plot', 'layoutheatmap: layoutheatmap <layout.tsv> <expr.tsv> <out> [--options]'),
+    ('levelGo', 'direct', 'biocjava.bioDoer.GeneOntology.Grapher.LevelDoer', '3g', 'plot', 'levelGo: levelGo <gene2Go.txt> <outTable> <oboFile> [--level N]'),
+    ('makemotif', 'bridge', 'MakeMotifCli', '3g', 'java', 'makemotif: makemotif <in.seqs.txt> <out.meme> [--mol DNA|RNA|Protein]   # 等长序列→MEME motif 文件（GUI 逆向接口；产物可直接喂 fimo/mast）'),
+    ('marker', 'direct', 'biocjava.bioDoer.markerDesign.BigMarkerRandomDesign', '3g', 'plot', 'marker: marker <MarkerDist|MarkerFilter|SampleDist|BigMarkerRandomDe'),
+    ('markertools', 'bridge', 'MarkerToolsCli', '3g', 'plot', 'markertools: markertools <filter|dist|sampledist> <in.marker.tab> [maxPoi'),
+    ('mast', 'bridge', 'MastCli', '3g', 'java', 'mast: mast <sequence.fa> <motifs.meme|meme.xml> <workingDir> [--motif-to-use N] [--max-motif-pvalue 0.0001] [--max-seq-evalue 10]   # MAST motif 搜索（GUI 逆向接口 QuickRunMAST，需系统 mast；产物 mast.html/txt/xml）'),
+    ('mast2tab', 'bridge', 'Mast2TabCli', '3g', 'plot', 'mast2tab: mast2tab <mast|meme.xml> <out.tab>'),
+    ('mastExtract', 'direct', 'biocjava.bioDoer.MEME.ExtractSeq.ExtractSeqFromMastXML', '3g', 'plot', 'mastExtract: mastExtract <in.fa> <mast.xml> <out.txt>   # 从 MAST XML 提取命中'),
+    ('mastrun', 'bridge', 'MastRunCli', '3g', 'plot', 'mastrun: mastrun <meme.xml> <seq.fasta> <workingDir> [--motifs M] [--'),
+    ('mcscanx', 'bridge', 'MCScanXCli', '3g', 'plot', 'mcscanx: mcscanx <gff> <blast> <outPrefix> [--html]   # 共线性检测'),
+    ('meme', 'bridge', 'MemeCli', '3g', 'java', 'meme: meme <in.fa> <workingDir> <outMemeXml> [--nmotifs N] [--minw N] [--maxw N] [--evt 0.05] [--mod zoops|oops|anr]   # MEME motif 发现（GUI 逆向接口 QuickRunMEME，需系统 meme；产物可与 memeViz/fimo 串联）'),
+    ('meme2tab', 'bridge', 'Meme2TabCli', '3g', 'java', 'meme2tab: meme2tab <meme.xml|mast.xml> <out.tab>   # MEME/MAST XML→motif 域表（GUI 逆向接口 MEMESuiteXMLtoTab）'),
+    ('memerun', 'bridge', 'MemeRunCli', '3g', 'plot', 'memerun: memerun <in.fasta> <workingDir> [--motif N] [--minW N] [--ma'),
+    ('mggxf', 'bridge', 'MgGxfCli', '3g', 'plot', 'mggxf: mggxf <inGenePair|blastTab6> <in.simplified.gff> <out.Linked'),
+    ('microgenome', 'direct', 'biocjava.bioDoer.JIGplotToolkit.MicroGenomeViz.MicroGenomeAnnotationCircosPlot', '3g', 'plot', 'microgenome: microgenome <inGBK> <anno.tsv> <out> [micro|macro]'),
+    ('microsyn', 'bridge', 'MicroSynCli', '3g', 'plot', 'microsyn: microsyn <gxf1> <gxf2> <collinearity> <out> [--chr1 C --star'),
+    ('mirnaIdentify', 'bridge', 'MirIdentifyCli', '3g', 'plot', 'mirnaIdentify: mirnaIdentify <genome.fa> <targetSo.tsv> <outPredict.txt> <outChecklog.txt> [--checkARM BOTH|FIVE|THREE] [--maxAsy N] [--maxBulge N]   # miRNA 前体鉴定（GUI 逆向 #78 MirIdentifyCli；⚠️ 第 4 参 outChecklog 必需，docstring 原漏写 N29）'),
+    ('mirnaTarget2', 'direct', 'biocjava.bioDoer.miRNA.Target2TablePipe', '3g', 'plot', 'mirnaTarget2: mirnaTarget2 <mirna.fa> <target.fa> <out.txt> [--revCom true'),
+    ('mountain', 'bridge', 'MountainPlotCli', '3g', 'plot', 'mountain: mountain <fold.txt> <out.tsv>'),
+    ('mpattern', 'bridge', 'MotifPatternCli', '3g', 'plot', 'mpattern: mpattern <mast.xml> <out.svg> [--max-motif N] [--shape RoundRect|Rect|Oval] [--line Middle|Up|Down|Splice] [--gradient] [--show-num]   # MEME/MAST motif 序列标注图（GUI 逆向接口，postGraph(String,panel) 重载绕弹窗）'),
+    ('multiEfp', 'bridge', 'MultiSuperHeatCli', '3g', 'plot', 'multiEfp: multiEfp <inTGA> <sample2cc> <expMat1[,expMat2,...]> <geneId'),
+    ('multisyn', 'bridge', 'SeveralSpeciesCli', '3g', 'plot', 'multisyn: multisyn <gxf.lst> <collinear.lst> <out> [--genes idlist.txt'),
+    ('nwAlign', 'bridge', 'NeedlemanWunschCli', '3g', 'plot', 'nwAlign: nwAlign <seq1.fa> <seq2.fa> <out> [--protein|--dna] [--format EMBOSS|FASTA] [--gap-open N] [--gap-extend N] [--end-gap-open N] [--end-gap-extend N] [--end-weight]   # Needleman-Wunsch 全局比对（GUI 逆向接口 NeedleManWunschAlign；旧 SimpleBatchProcess 静默无产物已替换）'),
+    ('pafcomp', 'bridge', 'PafGC', '3g', 'plot', 'pafcomp: pafcomp --inPaf <paf> --outGraph <out> [--colorMode Target|Q'),
+    ('pafref', 'direct', 'biocjava.bioDoer.JIGplotToolkit.Paf.PafRefBaseCoverCalc', '3g', 'plot', 'pafref: pafref --inPaf <paf> --outTab <out.tsv>'),
+    ('pafviz', 'bridge', 'PafVizCli', '3g', 'plot', 'pafviz: pafviz <in.paf> <out.svg> [--graph-size N] [--color Target|Query|None] [--seed N] [--min-len N] [--switch-qnt] [--rc-color]   # PAF 比对 dot 图（GUI 逆向接口 PafViz.process，绕 quickShow）'),
+    ('partitionconflict', 'direct', 'biocjava.bioDoer.GenomeAssembly.ParititionByConflictFreq', '3g', 'plot', 'partitionconflict: partitionconflict <inConflictFreq.tsv> <polyPoid> <outCluste'),
+    ('peakanno', 'direct', 'biocjava.bioDoer.JIGplotToolkit.MACS2viz.peakAnno', '3g', 'plot', 'peakanno: peakanno <gxf> <macs2_peak.xls> <out.tsv> [--dist N]'),
+    ('peakdist', 'bridge', 'PeakDistCli', '3g', 'plot', 'peakdist: peakdist <chrLen.tsv> <macs2_peak.xls> <out> [--chrHeight H]'),
+    ('peaktss', 'direct', 'biocjava.bioDoer.JIGplotToolkit.MACS2viz.peakTssHeatMap', '3g', 'plot', 'peaktss: peaktss <gxf> <macs2_peak.xls> <out.svg/png> [--dist N] [--b'),
+    ('pep2codon', 'bridge', 'Pep2CodonCli', '3g', 'plot', 'pep2codon: pep2codon <cds.fa> <pep.aln.fa> <codon.aln.out>'),
+    ('pfammotif', 'bridge', 'PfamMotifCli', '3g', 'plot', 'pfammotif: pfammotif <pfamscan.txt> <in.fasta> <out.svg|png|pdf> [newic'),
+    ('phylotree', 'bridge', 'PhyloTreeCli', '3g', 'plot', 'phylotree: phylotree <in.nwk> <out> [vertical] [width] [height]'),
+    ('pileup', 'bridge', 'PileUpCli', '3g', 'plot', 'pileup: pileup <blast.xml> <out.svg> [--query NAME]'),
+    ('plotrna', 'direct', 'biocjava.bioDoer.JIGplotToolkit.miRCoverage.PlotRNAfold', '3g', 'plot', 'plotrna: plotrna <genomeFA> <region> <SAM> [--directPDF out.pdf]'),
+    ('preparespecies', 'direct', 'biocjava.bioDoer.ComparativeGenomics.PrepareSpecies', '3g', 'plot', 'preparespecies: preparespecies <prefix> <inGenome.fa> <inGFF> <outGenome.fa>'),
+    ('qdot', 'bridge', 'QuickGenomeDotCli', '3g', 'plot', 'qdot: qdot <blast.tab> <in.gff> <chrLayout.txt> <out.svg> [--point-size N] [--highlight genes.txt]   # 基因组 dot plot（插件 P00380 CLI 化；blast/gff/chrLayout 可由 mcscanxd 产出，绕开插件 quickShow GUI 崩溃直驱 dotdotdot）'),
+    ('qpcr', 'bridge', 'QpcrCli', '3g', 'plot', 'qpcr: qpcr <data.txt> <out> [w] [h]   (data: name       mean    sd)'),
+    ('qpcrExp', 'bridge', 'QpcrDdctCli', '3g', 'plot', 'qpcrExp: qpcrExp <in.qpcr.tab> <out.xls>'),
+    ('qpcrproc', 'bridge', 'QpcrProcCli', '3g', 'plot', 'qpcrproc: qpcrproc <in.qpcr.tab> <out.xls>'),
+    ('quickFamily', 'direct', 'biocjava.bioDoer.BLAST.ReciprocalBlast.QuickGeneFamilyIdentification', '3g', 'plot', 'quickFamily: quickFamily <refPep.fa> <familyIds.txt> <queryPep.fa> <outPr'),
+    ('recipBlast', 'direct', 'biocjava.bioDoer.BLAST.ReciprocalBlast.ReciprocalBlast', '3g', 'plot', 'recipBlast: recipBlast <query.fa> <subject.fa> <outPrefix> [--queryIds i'),
+    ('regionAnno', 'direct', 'biocjava.bioDoer.GXFUtils.RegionGXFOverlapAnnotation', '3g', 'plot', 'regionAnno: regionAnno <in.gff3> <region.txt> <outTab> [--flankLen N] [-'),
+    ('regiondepth', 'bridge', 'RegionDepthCli', '3g', 'plot', 'regiondepth: regiondepth <in.sam> <region> <out.depth> [scaleFactor]'),
+    ('rnaplot', 'bridge', 'RNAplotCli', '3g', 'plot', 'rnaplot: rnaplot <seq.fa|rawSeq> <out> [--colorMap "seq1=R,G,B;seq2=R'),
+    ('sambamcov', 'bridge', 'SamBamCovCli', '3g', 'plot', 'sambamcov: sambamcov <in.bam> <out.tsv> [binSize] [countMode]'),
+    ('sepChr', 'direct', 'biocjava.bioDoer.GenomeAssembly.SeperateChrByAlleles', '3g', 'plot', 'sepChr: sepChr <gene2chr.tsv> <in.miniprot.gff> <outMap>'),
+    ('seqconvert', 'bridge', 'SeqConverterCli', '3g', 'plot', 'seqconvert: seqconvert -i <in> -o <out> -iF <fmt> -oF <fmt>'),
+    ('seqlentrack', 'bridge', 'SeqLenTrackCli', '3g', 'plot', 'seqlentrack: seqlentrack <seqlen.txt> <out.svg|png|pdf> [newick.treefile]'),
+    ('simplehmmscan', 'bridge', 'SimpleHmmscanCli', '3g', 'plot', 'simplehmmscan: simplehmmscan <pfamA.hmm> <target.pep> <idList.txt> <out.txt'),
+    ('sricher', 'bridge', 'SimpleEnricherCli', '3g', 'java', 'sricher: sricher <in.tsv> <out.xls> <totalAnnoIdx> <totalHitIdx> <selAnnoIdx> <selHitIdx> [--header]   # 简单富集（GUI 逆向接口 SimpleEnricher，超几何+BH；goEnrich 轻量版无需 OBO）'),
+    ('supercircos', 'bridge', 'SuperCircosCli', '3g', 'plot', 'supercircos: supercircos <config.cfg> <out> [width] [height]'),
+    ('tableAppend', 'direct', 'biocjava.bioDoer.Table.TableAppend', '3g', 'plot', 'tableAppend: tableAppend <inTab1> <inTab2> <outTab> [--c1 N] [--c2 N]   #'),
+    ('tableCast', 'direct', 'biocjava.bioDoer.Table.TableCast', '3g', 'plot', 'tableCast: tableCast <inLong.txt> <outMatrix>'),
+    ('tableColSel', 'direct', 'biocjava.bioDoer.Table.TableColSelector', '3g', 'plot', 'tableColSel: tableColSel <inTable> <outTable> <idList.txt> [--mode Match|'),
+    ('tableColSelect', 'bridge', 'TableColManipCli', '3g', 'plot', 'tableColSelect: tableColSelect <inTable> <outTable> <colName1> [colName2...]'),
+    ('tableCollapse', 'bridge', 'TableCollapseCli', '3g', 'plot', 'tableCollapse: tableCollapse <inTable> <keyColIndex> <outTable> [hasHeader '),
+    ('tableMelt', 'direct', 'biocjava.bioDoer.Table.TableMelt', '3g', 'plot', 'tableMelt: tableMelt <inTable> <outTable>   # 宽表转长表（第88引擎，TableMelt）'),
+    ('tableMerge', 'direct', 'biocjava.bioDoer.Table.TableMerger', '3g', 'plot', 'tableMerge: tableMerge --inFileArr "f1,f2,..." --inColIndexArr "0,1,..." --outTable <out> [--defaultNAvalue NA] [--appendMergedKey true|false] [--rmKeyColumns true|false]   # 按键合并多个表格(TableMerger；⚠️ ArgsParser 式，旧 docstring 位置参数写法已废弃 N3；位置参数兼容见 _tableMerge_impl)'),
+    ('tableSplit', 'direct', 'biocjava.bioDoer.Table.TableSplitByCol', '3g', 'plot', 'tableSplit: tableSplit <inTab> <outDir> [--colIndex N] [--suffix .txt]'),
+    ('tableTranspose', 'direct', 'biocjava.bioDoer.Table.TableTransposer', '3g', 'plot', 'tableTranspose: tableTranspose <inTable> <outTable>   # 表格转置（第95引擎，TableTran'),
+    ('tableUniq', 'direct', 'biocjava.bioDoer.Table.TableUniq', '3g', 'plot', 'tableUniq: tableUniq <inTab> <outFile> [--colIndex N] [--showFreq true|'),
+    ('tauIndex', 'bridge', 'TauCalcCli', '3g', 'plot', 'tauIndex: tauIndex <inExpTab> <outTAU>'),
+    ('trimmsa', 'bridge', 'TrimMSACli', '3g', 'plot', 'trimmsa: trimmsa <in.aln.fa> <out.aln.fa> [ratio]'),
+    ('twoSeqBlast', 'direct', 'biocjava.bioDoer.BLAST.CompareTwoSeqSet', '3g', 'plot', 'twoSeqBlast: twoSeqBlast <query.fa> <subject.fa> <out.txt> [--prog blastp'),
+    ('upset', 'bridge', 'UpSetCli', '3g', 'plot', 'upset: upset <set1.txt> <set2.txt> [<set3.txt>...] <out.svg> [--min-overlap N] [--rank1 Size|Count|Name] [--rank2 ...] [--rank3 ...] [--size-mode/--count-mode/--name-mode Increasing|Decreasing]   # UpSet 集合图（GUI 逆向接口 UpSetPlot.plot，绕 show 弹窗）'),
+    ('venn2', 'direct', 'biocjava.bioDoer.JJplot2Toolkit.WonderfulVenn.Venn2', '2g', 'plot', 'venn2: venn2 --List1 <setA.txt> --List2 <setB.txt> --label1 A --label2 B --graph <out> --prefix <out> [--bgNum N]'),
+    ('venn3', 'direct', 'biocjava.bioDoer.JJplot2Toolkit.WonderfulVenn.Venn3', '2g', 'plot', 'venn3: venn3 --List1 <A> --List2 <B> --List3 <C> --label1..3 <labels> --graph <out> --prefix <out>'),
+    ('venn4', 'direct', 'biocjava.bioDoer.JJplot2Toolkit.WonderfulVenn.Venn4Ellipse', '2g', 'plot', 'venn4: venn4 --List1 <A> --List2 <B> --List3 <C> --List4 <D> --label1..4 <labels> --graph <out> --prefix <out>'),
+    ('venn5', 'bridge', 'Venn5Cli', '3g', 'plot', 'venn5: venn5 <out> <setA.txt> <setB.txt> <setC.txt> <setD.txt> <setE.txt> [labels]'),
+    ('venn6', 'bridge', 'Venn6Cli', '3g', 'plot', 'venn6: venn6 <out> <setA..F.txt> [labels]'),
+    ('violin', 'bridge', 'ViolinCli', '3g', 'plot', 'violin: violin <in.tsv> <out> [width] [height]'),
+    ('virusRecomb', 'direct', 'biocjava.bioDoer.VirusDetect.RecombinationAnalysis', '3g', 'plot', 'virusRecomb: virusRecomb <inDB.fa> <inContig.fa> <outDir>   # 病毒重组分析（第77引'),
+    ('visualizeblock', 'bridge', 'VisualizeCli', '3g', 'plot', 'visualizeblock: visualizeblock <inBlockOut> <out.pdf> [--labels "Genome1,Gen'),
+    ('kaks', 'direct', 'biocjava.bioIO.BioSoftPipeServer.PairWiseKaKsCalculator', '2g', 'java', 'kaks --inCDS <cds.fa> --inGenePair <pairs.txt> --outKaks <out.xls> [--inCPU N] [--inPep pep.fa]   # 成对 Ka/Ks 计算（GUI 逆向 PairWiseKaKsCalculator；自带 ArgsParser：--key value；inGenePair 为 ID1\\tID2 每行，缺文件时自动全两两配对并翻译 CDS）'),
+    ('sixframe', 'bridge', 'SixFrameTranlaterCli', '2g', 'java', 'sixframe <in.fa> <out.fa>   # 六框翻译（GUI 逆向 #16 SixFrameTranlater：setInFile/setOutFile/process；输出每条序列 6 框 12 条；注意引擎类名拼写 SixFrameTranlater 少一个 s）'),
+    ('longestorf', 'direct', 'biocjava.bioIO.ORF.GetLongestORF', '2g', 'java', 'longestorf --inFa <seq.fa> --outORFs <out.fa>   # 批量最长完整 ORF 预测（GUI 逆向 #17 GetLongestORF：setFastaFile/setOutFile/startPredict；自带 ArgsParser）'),
+    ('protparam', 'direct', 'biocjava.bioWeb.ProtParamWrapper', '2g', 'java', 'protparam --inFa <pep.fa> --outTab <out.txt>   # 蛋白理化性质批量计算（GUI 逆向 #18 ProtParamWrapper.batchCalc；⚠️ 联网 POST Expasy。输出：分子量/pI/不稳定指数/脂肪族指数/GRAVY）'),
+    ('seqpattern', 'direct', 'biocjava.bioIO.FastX.QuickLocateSeqPattern', '2g', 'java', 'seqpattern --inFasta <seq.fa> --pattern <ATG|regex> --outTab <out.gff3> [--overlap] [--maxSeqLen <len>]   # 序列模式定位（GUI 逆向 #20 QuickLocateSeqPattern：正则找模式→GFF3；--key value 空格分隔）'),
+    ('bed2gff3', 'bridge', 'RegionBedToGFF3Cli', '2g', 'java', 'bed2gff3 <in.bed> <out.gff3> [genome.fa]   # exon BED→GFF3（GUI 逆向 #21 RegionBedToGFF3；⚠️ BED 第4列须为 ID:链向:编码 如 G01:+:C；同 ID 多行合并出 mRNA+exon）'),
+    ('careclassify', 'bridge', 'PlantCAREResultClassifyCli', '2g', 'java', 'careclassify <plantcare.tab> <out.xls>   # PlantCARE 顺式元件分类（GUI 逆向 #22；第8列 motif 名查 jar 内置 97 类表，行尾追加大类/亚类；查不到 NA）'),
+    ('subtree', 'bridge', 'GetSubNewickTreeCli', '2g', 'java', 'subtree <tree.nwk> <idList.txt> <out.nwk> [--contain]   # Newick 子树提取（GUI 逆向 #23 GetSubNewickTreeGUIPanel→PhyloTreeMan.getSubTree；--contain 模糊匹配；引擎重算诱导子树内部枝长）'),
+    ('protsim', 'bridge', 'CalculateSimilarityCli', '2g', 'java', 'protsim <pep.fa> <out.matrix>   # 蛋白两两相似度矩阵（GUI 逆向 #24 ProteinPairwiseSimilarityMatrixGUIPanel→CalculateSimilarity；百分比矩阵 TSV）'),
+    ('iqtree', 'bridge', 'QuickRunIQtreeCli', '2g', 'java', 'iqtree <aln.fa> <outPrefix> [--model MFP] [--ufboot 1000] [--boot N] [--freerate] [--asc] [--threads N] [--redo]   # IQ-TREE ML 建树（GUI 逆向 #28 QuickRunIQtree；⚠️ UFBoot 须 ≥1000 否则引擎静默失败；产物 outPrefix.treefile；依赖系统 iqtree）'),
+    ('trimal', 'bridge', 'QuickTrimALCli', '2g', 'java', 'trimal <in.aln> <out.aln> [--mode gappyout|strict|strictplus|automated1] [--format fasta|clustal|phylip|nexus|mega|nbrf] [--keepheader]   # trimAl 比对修剪（GUI 逆向 #29 QuickTrimAL；默认 automated1；依赖系统 trimal；muscle→trimal→iqtree 管线）'),
+    ('gblocks', 'bridge', 'JgblocksCli', '2g', 'java', 'gblocks <in.aln.fa> <out.aln.fa> [--is 0.5] [--fs 0.85] [--cp 8] [--bl1 15] [--bl2 10] [--nongap 0.5] [--gaptreat none|half|all]   # Gblocks 保守区修剪（GUI 逆向 #30 Jgblocks 纯 Java 实现；main 仅 in/out 全参数须 setter；⚠️ 高歧异比对可能 validSites=0 合法）'),
+    ('goAnno', 'direct', 'biocjava.bioDoer.GeneOntology.Annotation.GoAnnoPipe', '3g', 'java', 'goAnno --IdmappingDb <idmapping.DB.gz> --BlastxAnnoFile <blastx.xml> [--inPutFileType BlastxXml|Query2GiTable] [--maxEvalue 1e-5] [--minQueryCov 0.33] [--outDir dir] [--isDoDbFormat]   # GO 注释管道（GUI 逆向 #32 GoAnnotationGUIPanel→GoAnnoPipe；自带 ArgsParser；⚠️ idmappingDb 须 gzip 格式「ID; ID\\tGO:num; GO:num」；产物 outDir/<输入名>.xls）'),
+    ('fasplit', 'direct', 'biocjava.bioIO.FastX.FastaIndex.QuickSpiltFasta', '2g', 'java', 'fasplit --inFa <in.fa> --outPre <prefix> --NumPerFile <N> [--byCount true]   # FASTA 按记录数拆分（GUI 逆向 #33 QuickSpiltFasta；⚠️ 与 filesplit 不同：按记录不切行，产物 prefix.N.split.fa）'),
+    ('famerge', 'bridge', 'FastaMergerCli', '2g', 'java', 'famerge <out.fa> <in1.fa> <in2.fa> [...]   # 多 FASTA 合并（GUI 逆向 #33 FastaMergerAndSpliter.Merge）'),
+    ('clearchar', 'bridge', 'FileCleanerCli', '2g', 'java', 'clearchar <in.txt> <out.txt>   # 文件非法字符清理（GUI 逆向 #34 FileCleaner.simplifyFile；非可打印 ASCII/非 tab→_，空白行跳过，逐行报告）'),
+    ('gb2fa', 'bridge', 'GenBank2FastaCli', '2g', 'java', 'gb2fa <in.gb> <out.fa>   # GenBank→FASTA 转换（GUI 逆向 #36 genBank2Fasta；头含 locus/accession/organism/definition）'),
+    ('findhomolog', 'direct', 'biocjava.bioIO.BioSoftPipeServer.FindBestHomology', '3g', 'java', 'findhomolog --inQueryProteinSet <query.pep> --inSubjectProteinSet <subject.pep> --targetIDs <ID[,ID2]> --outDir <dir> [--threads N] [--extendClade] [--sensitive N] [--similarity 0.x] [--weightCov 0.x] [--plot] [--directGraph]   # 最优同源查找（GUI 逆向 #37 FindBestHomology，同引擎覆盖 GenomeAnnotationSlim+FindBestHomology 两面板；自带 ArgsParser；BLAST+可选建树）'),
+    ('taxparse', 'bridge', 'TaxonomyBatchCli', '2g', 'java', 'taxparse <idList.txt> <out.xls>   # 物种名批量分类解析（GUI 逆向 #38 TaxonomyParserGUIPanel→NCBITaxonomy；⚠️ 联网 NCBI eutils；输出 9 级分类+透传列；单次版=tbtools tool NCBITaxonomy）'),
+    ('srr2ena', 'bridge', 'GetENALinksCli', '2g', 'java', 'srr2ena <srrList.txt> <out.xls>   # SRR→ENA 下载链接解析（GUI 逆向 #39 GetENALinksOfSRR；⚠️ 联网 ENA filereport API+引擎自带 0~3s 限速；17 字段含 fastq_ftp/aspera）'),
+    ('sraxml2tab', 'direct', 'biocjava.bioIO.SRAtools.ParseSRAXml2Table', '2g', 'java', 'sraxml2tab --sraFullXML <sra.xml> --outTab <out.xls>   # SRA XML→信息表（GUI 逆向 #40 ParseSRAXml2Table；自带 ArgsParser；离线 JDOM 解析；XML 从 efetch db=sra 获取）'),
+    ('sranum2info', 'direct', 'biocjava.bioWeb.EntrezUtils.BatchGetSRARecordInfo', '2g', 'java', 'sranum2info --sraIdList <srrList.txt> --outTabInfo <out.xls>   # SRR 批量信息表（GUI 逆向 #41 BatchGetSRARecordInfo；⚠️ 联网 NCBI Entrez+限速；自带 ArgsParser；SRA 组 3/3 全清）'),
+    ('blat', 'bridge', 'BlatExecutorCli', '2g', 'java', 'blat <db.fa> <query.fa> <out> [--format blast9|psl|pslx|axt|maf|sim4|wublast|blast|blast8] [--minScore N] [--minIdentity 0.x] [--noHead] [--mode auto|dnadna|dnarna] [--tileSize N] [--stepSize N] [--maxGap N] [--maxIntron N] [--extra "opts"]   # BLAT 序列比对（GUI 逆向 #42 BlatExecutor；org.ucsc.blat 纯 Java 实现内嵌 jar 无需外部二进制）'),
+    ('seqrecommend', 'bridge', 'AssemblyRecommandCli', '2g', 'java', 'seqrecommend <genomeSize1n_bp> [--polyploid] [--het 0.01] [--level Minimum|Draft|Haplotyped_Resolved|Haplotyped_T2T]   # 基因组组装测序量推荐（GUI 逆向 #43 AssemblyGenomeDataSizeRecommand；纯计算离线；Hifi/HiC 深度+数据量）'),
+    ('seqfetch', 'direct', 'biocjava.bioWeb.EntrezUtils.NcbiSmartSeqFetchEntrezUtils', '2g', 'java', 'seqfetch --inFile <idList.txt> --outSeqFile <out.fa> --outReport <report.txt> [--targetDb nuccore|protein] [--preferDb db] [--format fasta] [--greedyMode] [--apiKey KEY] [--auditFile f]   # NCBI 智能序列下载（GUI 逆向 #44 NcbiSmartSeqFetchEntrezUtils；⚠️ 联网 Entrez+限速；ID 自动检测/转换/审计；支持 apiKey 提速）'),
+    ('pubmed', 'bridge', 'PubmedSearchCli', '2g', 'java', 'pubmed <query> <out.xls>   # PubMed 文献检索汇总（GUI 逆向 #45 PubmedSearch.process；⚠️ 联网 eutils；输出期刊/标题/年份/IF/DOI 表）'),
+    # N13: gwas 分组补命令（vcfAddID/mimicVqsr 原只在 CLI_TOOLS，gwas 分组是空壳）
+    ('vcfAddID', 'direct', 'biocjava.bioDoer.GWAS.VCFAddID', '2g', 'java', 'vcfAddID: vcfAddID --inFile <vcf> --outFile <out.vcf>   # VCF 加 ID 列（GWAS；ArgsParser --inFile/--outFile，支持 .gz）'),
+    ('mimicVqsr', 'direct', 'biocjava.bioDoer.GWAS.MimicVqsrCutoffFind', '2g', 'java', 'mimicVqsr: mimicVqsr --inFile <vcf> --outFile <out.txt>   # VCF 质量指标（QD/MQ/FS/SOR；GWAS）'),
+]
+
+
+
+_IMPL_REGISTRY = {}  # 第四轮评审: 显式 impl 注册表(替代隐式 globals 注入)
+
+def parse_kv_args(args, spec, raw_flags=()):
+    """通用参数解析(替代各 impl 重复的 while 循环)。
+
+    spec: {flag: (attr, converter)} —— 取值 flag,converter 可选(如 int);
+    raw_flags: 出现即收集到 raw 的 flag(如 --super5)。
+    返回 (kwargs, positional, raw);kwargs 只含出现的 flag,缺省值由调用方自行默认。
+    """
+    kwargs, pos, raw = {}, [], []
+    i = 0
+    n = len(args)
+    while i < n:
+        a = args[i]
+        if a in spec and i + 1 < n:
+            attr, conv = spec[a]
+            val = args[i + 1]
+            kwargs[attr] = conv(val) if conv else val
+            i += 2
+        elif a in raw_flags:
+            raw.append(a)
+            i += 1
+        else:
+            pos.append(a)
+            i += 1
+    return kwargs, pos, raw
+
+def _warn_gtf_input(args, cmd):
+    """N28/N22/N36: Gxf 族命令输入格式预检——GTF 引擎 NPE、BED 误报 GFF3/GTF 识别。"""
+    for a in args:
+        if not a.startswith("-") and a.lower().endswith((".gtf", ".gtf.gz")):
+            print(f"⚠️ 格式提醒: {cmd} 对 GTF 输入支持有限（GENCODE 真 GTF 会触发引擎 NPE，N28）", file=sys.stderr)
+            print("   （建议先转 GFF3 再输入；如确认可忽略）", file=sys.stderr)
+            break
+        if not a.startswith("-") and a.lower().endswith((".bed", ".bed.gz")):
+            # N22/N36: 引擎 GFF3/GTF 识别对 BED 报误导性错误（"can not decide"）+ GxfStat NPE
+            print(f"⚠️ 格式提醒: {cmd} 不支持 BED 输入（引擎会报『can not decide GFF3 or GTF』并可能 NPE）", file=sys.stderr)
+            print("   （请提供 GFF3/GTF 注释文件）", file=sys.stderr)
+            break
+
+
+# N32: 无参调用会挂起/弹窗的命令（引擎无参读 stdin/等 GUI）——直接打印用法退出
+_NOARG_HANG = {"gel"}
+
+def _make_impl(cmd, kind, cls, xmx, runner, doc):
+    """工厂：按注册表条目生成 _xxx_impl 闭包（保持 (args, verbose, quiet) 签名）"""
+    # N28: Gxf 族引擎对 GTF 输入解析 NPE（GENCODE 真 GTF 实证），输入预检警告
+    gxf_cmd = cmd.startswith("gxf") or cmd in ("gsadiag", "annocompare", "genedensity", "gblocks")
+    if kind == "bridge":
+        def impl(args, verbose=False, quiet=False):
+            if cmd in _NOARG_HANG and not args:
+                print(f"用法: {doc.split(':', 1)[1].strip() if ':' in doc else cmd}", file=sys.stderr)
+                return 1
+            if gxf_cmd:
+                _warn_gtf_input(args, cmd)
+            ensure_bridge(cls)
+            java_args = ["java", f"-Xmx{xmx}", "-cp", cp(BUILD_DIR, JAR), cls] + args
+            if runner == "plot":
+                return run_plot(java_args, verbose=verbose, quiet=quiet, command_name=cmd)
+            return run_java(java_args, verbose=verbose, quiet=quiet, command_name=cmd)
+    else:  # direct
+        def impl(args, verbose=False, quiet=False):
+            if cmd in _NOARG_HANG and not args:
+                print(f"用法: {doc.split(':', 1)[1].strip() if ':' in doc else cmd}", file=sys.stderr)
+                return 1
+            if gxf_cmd:
+                _warn_gtf_input(args, cmd)
+            # N27: direct 类也含 build/（fake jaxb DatatypeConverter 等），否则 JDK9+ 缺 javax.xml.bind
+            java_args = ["java", f"-Xmx{xmx}", "-cp", cp(BUILD_DIR, JAR), cls] + args
+            if runner == "plot":
+                return run_plot(java_args, verbose=verbose, quiet=quiet, command_name=cmd)
+            return run_java(java_args, verbose=verbose, quiet=quiet, command_name=cmd)
+    impl.__doc__ = doc
+    impl.__name__ = f"_{cmd}_impl"
+    return impl
+
+
+_HANDWRITTEN = {} if False else None  # placeholder
+for _cmd, _kind, _cls, _xmx, _runner, _doc in ENGINE_REGISTRY:
+    if hasattr(sys.modules[__name__], f"_{_cmd}_impl"):
+        # 手写实现在 ENGINE_REGISTRY 后同文件定义时会覆盖注册; 冲突条目应在注册表删除(mirnatarget/msy 已删)
+        raise SystemError(f"注册表与手写实现冲突: {_cmd}(手写 {f'_{_cmd}_impl'} 存在)——请从 ENGINE_REGISTRY 删除该条目")
+    # 设计说明(第三轮审查评估): 表驱动工厂统一产出 _<name>_impl,cli_load 按名查找;
+    # 命令清单的单一数据源是 ENGINE_REGISTRY,command_metadata.json 是其投影(gen_metadata 生成,--check 防漂移),
+    # 不再反向依赖 JSON 生成 click(避免运行时依赖生成物、且 JSON 不含 runner/xmx/doc 等运行时信息)。
+    # 第四轮评审: 显式注册表(替代隐式 globals 注入),cli_load 优先查 _IMPL_REGISTRY
+    _fn = _make_impl(_cmd, _kind, _cls, _xmx, _runner, _doc)
+    _IMPL_REGISTRY[_cmd] = _fn
+    globals()[f"_{_cmd}_impl"] = _fn
+
+
+# ── 特殊实现（手写保留）──────────────────────────────────────────
+
+def _hmmsearch_impl(args, verbose=False, quiet=False):
+    """hmmsearch: hmmsearch <pfamA.hmm> <target.pep> <idList.txt> <out.txt>   # HMM Search 域扫描（= simpleHmmscan 引擎，调系统 hmmsearch，G1 补齐别名）"""
+    return _simplehmmscan_impl(args, verbose=verbose, quiet=quiet)  # type: ignore[name-defined]  # noqa: F821  # 注册表动态生成
+
+
+def _gxfAttr_impl(args, verbose=False, quiet=False):
+    """gxfAttr: gxfAttr <in.gff3|gtf> <out.tsv> [--feature mRNA] [--attrs ID,Name,Parent]   # GXF 属性/ID 对照表提取（G7 补齐，Python 原生，jar 无此引擎）"""
+    import re as _re
+    # 解析参数
+    pos, feature, attrs = [], "mRNA", None
+    _kw, pos, _ = parse_kv_args(args, {
+        "--feature": ("feature", None),
+        "--attrs": ("attrs", lambda s: [a.strip() for a in s.split(",") if a.strip()])})
+    feature = _kw.get("feature", "mRNA"); attrs = _kw.get("attrs")
+    if len(pos) < 2:
+        print("用法: gxfAttr <in.gff3|gtf> <out.tsv> [--feature mRNA] [--attrs ID,Name,Parent]", file=sys.stderr)
+        return 1
+    in_gxf, out_tsv = pos[0], pos[1]
+    if not os.path.isfile(in_gxf):
+        print(f"❌ 输入文件不存在: {in_gxf}", file=sys.stderr)
+        return 2
+    def parse_attrs(field):
+        """兼容 GFF3 (k=v;) 与 GTF (k "v";) 属性列"""
+        d = {}
+        for m in _re.finditer(r'([^\s;=]+)\s*=\s*"?([^";]+)"?\s*;?', field):
+            d[m.group(1)] = m.group(2)
+        if not d:  # GTF 风格
+            for m in _re.finditer(r'([^\s;]+)\s+"([^"]+)"\s*;?', field):
+                d[m.group(1)] = m.group(2)
+        return d
+    rows, attr_keys = [], []
+    with open(in_gxf, encoding="utf-8", errors="replace") as f:
+        for line in f:
+            if line.startswith("#") or not line.strip():
+                continue
+            parts = line.rstrip("\n").split("\t")
+            if len(parts) < 9:
+                continue
+            feat = parts[2]
+            if feature.lower() != "all" and feat.lower() != feature.lower():
+                continue
+            d = parse_attrs(parts[8])
+            d["__feature__"] = feat
+            d["__chr__"], d["__start__"], d["__end__"], d["__strand__"] = parts[0], parts[3], parts[4], parts[6]
+            rows.append(d)
+            for k in d:
+                if not k.startswith("__") and k not in attr_keys:
+                    attr_keys.append(k)
+    if not rows:
+        print(f"⚠️ 未提取到 feature={feature} 的记录（--feature all 提取全部）", file=sys.stderr)
+        return 1
+    # 列序: feature/坐标 + 用户指定 attrs + 其余按出现序
+    lead = ["__feature__", "__chr__", "__start__", "__end__", "__strand__"]
+    if attrs:
+        ordered = attrs + [k for k in attr_keys if k not in attrs]
+    else:
+        preferred = ["ID", "Name", "Parent", "gene_id", "gene_name", "transcript_id", "symbol", "gene", "product", "Note"]
+        ordered = [k for k in preferred if k in attr_keys] + [k for k in attr_keys if k not in preferred]
+    header = ["feature", "chr", "start", "end", "strand"] + ordered
+    with open(out_tsv, "w", encoding="utf-8") as f:
+        f.write("\t".join(header) + "\n")
+        for d in rows:
+            f.write("\t".join([d.get(k, "") for k in lead] + [d.get(k, "") for k in ordered]) + "\n")
+    if not quiet:
+        print(f"[gxfAttr] {len(rows)} 条 {feature} 记录 × {len(ordered)} 属性列 → {out_tsv}", file=sys.stderr)
+    return 0
+
+
+def _kallisto_impl(args, verbose=False, quiet=False):
+    """kallisto: kallisto <transcriptome.fa> <reads.fq[,reads2.fq]> <outAbundance> [--kmer N] [--threads N] [--bootstrap N] [--bias] [--single] [--frag-len N] [--frag-sd N]   # RNA-seq 定量（插件 P00740 CLI 化，直调 kallisto 二进制——插件 wrapper 的 Linux 分支有拼接 bug 已绕开）"""
+    # A(kallisto Windows 二进制选择, V1 §3 P1-3): 按平台优先 .exe（Linux ELF 在 Windows 报 WinError 193）
+    bin_path = os.path.join(ROOT, "plugins", "lib", "bin",
+                            "kallisto.exe" if os.name == "nt" else "kallisto")
+    if not os.path.isfile(bin_path) and os.name == "nt":
+        bin_path = os.path.join(ROOT, "plugins", "lib", "bin", "kallisto")  # 回退旧路径
+    if not os.path.isfile(bin_path):
+        print(f"❌ kallisto 二进制缺失: {bin_path}", file=sys.stderr)
+        return 1
+    libs_dir = os.path.join(ROOT, "plugins", "lib", "kallisto-libs")
+    # 解析参数
+    pos, kmer, threads, bootstrap, bias, single, frag_len, frag_sd = [], 31, 4, 0, False, False, 200, 30
+    _kw, pos, _raw = parse_kv_args(args, {
+        "--kmer": ("kmer", int), "--threads": ("threads", int),
+        "--bootstrap": ("bootstrap", int), "--frag-len": ("frag_len", int),
+        "--frag-sd": ("frag_sd", int)}, raw_flags=("--bias", "--single"))
+    kmer = _kw.get("kmer", 21); threads = _kw.get("threads", 2)
+    bootstrap = _kw.get("bootstrap", 100); frag_len = _kw.get("frag_len", 400)
+    frag_sd = _kw.get("frag_sd", 20); bias = "--bias" in _raw; single = "--single" in _raw
+    if len(pos) < 3:
+        print("用法: kallisto <transcriptome.fa> <reads.fq[,reads2.fq]> <outAbundance> [--kmer N] [--threads N] [--bootstrap N] [--bias] [--single] [--frag-len N] [--frag-sd N]", file=sys.stderr)
+        return 1
+    tx, reads_str, out_ab = pos[0], pos[1], pos[2]
+    reads = [r.strip() for r in reads_str.split(",")]
+    import shutil
+    env = dict(os.environ)
+    env["PATH"] = os.path.dirname(bin_path) + os.pathsep + env.get("PATH", "")
+    if os.path.isdir(libs_dir):
+        env["LD_LIBRARY_PATH"] = libs_dir + os.pathsep + env.get("LD_LIBRARY_PATH", "")
+    idx = os.path.join(os.path.dirname(os.path.abspath(out_ab)), os.path.basename(out_ab) + ".kallisto.idx")
+    try:
+        # 1) index
+        os.unlink(idx)
+    except OSError:
+        pass
+    r = subprocess.run([bin_path, "index", "-k", str(kmer), "-i", idx, tx], env=env, capture_output=True, text=True)
+    if r.returncode != 0:
+        print(f"❌ kallisto index 失败:{chr(10)}{r.stderr[-800:]}", file=sys.stderr)
+        return r.returncode
+    # 2) quant
+    tmp_dir = tempfile.mkdtemp(prefix="kallisto_quant_")
+    try:
+        q = [bin_path, "quant", "-i", idx, "-o", tmp_dir, "-t", str(threads)]
+        if bias: q.append("--bias")
+        if bootstrap: q += ["-b", str(bootstrap)]
+        if single:
+            q += ["--single", "-l", str(frag_len), "-s", str(frag_sd)]
+        q += reads
+        r = subprocess.run(q, env=env, capture_output=True, text=True)
+        if r.returncode != 0:
+            print(f"❌ kallisto quant 失败:{chr(10)}{r.stderr[-800:]}", file=sys.stderr)
+            return r.returncode
+        abund = os.path.join(tmp_dir, "abundance.tsv")
+        if not os.path.isfile(abund):
+            print("❌ 未找到 abundance.tsv（quant 输出异常）", file=sys.stderr)
+            return 1
+        shutil.copy2(abund, out_ab)
+        if not quiet:
+            print(f"[kallisto] {os.path.basename(out_ab)} 已写出（kmer={kmer} threads={threads}{' bias' if bias else ''}{' single' if single else ''}）", file=sys.stderr)
+        return 0
+    finally:
+        shutil.rmtree(tmp_dir, ignore_errors=True)
+        try: os.unlink(idx)
+        except OSError as _e:
+            logging.getLogger(__name__).debug("io: %s", _e)  # 静默容错(第五轮评审:留痕)
+            pass
+
+
+def _fimo_impl(args, verbose=False, quiet=False):
+    """fimo: fimo --o <outDir> <motifs.meme> <promoter.fa>   # MEME FIMO motif 扫描（插件 P00552 等价，直调系统 fimo；meme-suite）"""
+    import shutil as _sh
+    fimo_bin = _sh.which("fimo")
+    if not fimo_bin:
+        print("❌ 未找到 fimo（meme-suite），安装: apt install meme-suite", file=sys.stderr)
+        return 1
+    r = subprocess.run([fimo_bin] + args, capture_output=True, text=True)
+    if r.returncode != 0:
+        print(f"❌ fimo 失败:\n{r.stderr[-600:]}", file=sys.stderr)
+        return r.returncode
+    if not quiet:
+        print("[fimo] 完成 (退出码 0)", file=sys.stderr)
+    return 0
+
+
+def _xml2blasttab_impl(args, verbose=False, quiet=False):
+    """xml2blasttab: xml2blasttab <in.xml> <out.txt>   # BLAST XML→标准 12 列表（GUI 逆向 #25 BlastXmlToBlastFoolTable.xml2ShowerTable；QueryID/SubjectID/Identity/E-value/BitScore...）"""
+    ensure_bridge("BlastXmlConvertCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, JAR), "BlastXmlConvertCli", "blasttab"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="xml2blasttab")
+
+
+def _xml2pairwise_impl(args, verbose=False, quiet=False):
+    """xml2pairwise: xml2pairwise <in.xml> <out.txt>   # BLAST XML→网页 pairwise 对齐文本（GUI 逆向 #25 BlastXMLToPairwise.parse；⚠️ 需 Hsp_query-frame/Hsp_hit-frame 字段）"""
+    ensure_bridge("BlastXmlConvertCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, JAR), "BlastXmlConvertCli", "pairwise"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="xml2pairwise")
+
+
+def _fa2tab_impl(args, verbose=False, quiet=False):
+    """fa2tab: fa2tab <in.fa> <out.tab>   # FASTA→表格 ID\\t序列（GUI 逆向 #26 FastaTable.fa2tab；与 tab2fa 往返一致）"""
+    ensure_bridge("FastaTableConvertCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, JAR), "FastaTableConvertCli", "fa2tab"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="fa2tab")
+
+
+def _tab2fa_impl(args, verbose=False, quiet=False):
+    """tab2fa: tab2fa <in.tab> <out.fa>   # 表格→FASTA（GUI 逆向 #26 FastaTable.tab2fa）"""
+    ensure_bridge("FastaTableConvertCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, JAR), "FastaTableConvertCli", "tab2fa"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="tab2fa")
+
+
+def _muscle_impl(args, verbose=False, quiet=False):
+    """muscle: muscle <in.fa> <out.aln> [--super5] [--threads N]   # MUSCLE 多序列比对（GUI 逆向 #27 MuscleGUIPanel→QuickRunMUSCLE；⚠️ 引擎硬编码 muscle3 -in/-out 语法在 v5 系统崩 → Python 直调自动适配；依赖系统 muscle）"""
+    import shutil as _sh
+    muscle_bin = _sh.which("muscle")
+    if not muscle_bin:
+        print("❌ 未找到 muscle，安装: apt install muscle", file=sys.stderr)
+        return 1
+    pos, extra = [], []
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a in ("--super5",):
+            extra.append(a); i += 1
+        elif a == "--threads" and i + 1 < len(args):
+            extra += ["-threads", args[i + 1]]; i += 2
+        else:
+            pos.append(a); i += 1
+    if len(pos) < 2:
+        print("用法: muscle <in.fa> <out.aln> [--super5] [--threads N]", file=sys.stderr)
+        return 1
+    in_fa, out_aln = pos[0], pos[1]
+    if not os.path.isfile(in_fa):
+        print(f"❌ 输入文件不存在: {in_fa}", file=sys.stderr)
+        return 2
+    # 版本探测：muscle -version 输出含 "muscle 5"/"MUSCLE v5" 即 v5 语法
+    vr = subprocess.run([muscle_bin, "-version"], capture_output=True, text=True)
+    vtxt = (vr.stdout + vr.stderr).lower()
+    is_v5 = "muscle 5" in vtxt or "muscle v5" in vtxt or "v5." in vtxt
+    super5 = "--super5" in extra
+    if is_v5:
+        cmd = [muscle_bin, "-super5" if super5 else "-align", in_fa, "-output", out_aln]
+        cmd += [e for e in extra if e != "--super5"]
+    else:
+        cmd = [muscle_bin, "-in", in_fa, "-out", out_aln]
+    r = subprocess.run(cmd, capture_output=True, text=True)
+    if r.returncode != 0 or not os.path.isfile(out_aln):
+        print(f"❌ muscle 失败:\n{r.stderr[-500:]}", file=sys.stderr)
+        return r.returncode or 1
+    if not quiet:
+        n = sum(1 for l in open(out_aln) if l.startswith(">"))
+        print(f"[muscle] 比对完成({'v5' if is_v5 else 'v3'} 语法): {n} 条 → {out_aln}", file=sys.stderr)
+    return 0
+
+
+def _bestid_impl(args, verbose=False, quiet=False):
+    """bestid: bestid --inQuery <query.pep> --Subject <subject.pep> --OutPrefix <outPrefix> [--useDiamond] [--threads N]   # 双向 BLAST 最优 ID 转换（GUI 逆向 #31 BestIDConverter；RBH 互撞 Excellent/Poor；⚠️ 引擎强制 --threads，CLI 公共 -t 会吃掉它——缺省自动注入 4，精确指定用 --config parameter.txt；依赖 blastp/diamond）"""
+    if "--threads" not in args:
+        args = args + ["--threads", "4"]
+    java_args = ["java", "-Xmx3g", "-cp", JAR,
+                 "biocjava.bioDoer.BLAST.ReciprocalBlast.BestIDConverter"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="bestid")
+
+
+def _getseqdb_impl(args, verbose=False, quiet=False):
+    """getseqdb: getseqdb <dbPrefix> <idList.txt> <out.fa> [--entry ID]   # 从 BLAST 库批量提取序列（GUI 逆向 #35 GetSeqFromBlastDBGUIPanel $5：blastdbcmd -db X -entry_batch ids -out Y；⚠️ 库须 makeblastdb -parse_seqids 建，否则 Skipped；依赖系统 blastdbcmd）"""
+    import shutil as _sh
+    blastdbcmd = _sh.which("blastdbcmd")
+    if not blastdbcmd:
+        print("❌ 未找到 blastdbcmd（ncbi-blast+），安装: apt install ncbi-blast+", file=sys.stderr)
+        return 1
+    pos, entry = [], None
+    i = 0
+    while i < len(args):
+        if args[i] == "--entry" and i + 1 < len(args):
+            entry = args[i + 1]; i += 2
+        else:
+            pos.append(args[i]); i += 1
+    if entry:
+        if len(pos) < 2:
+            print("用法: getseqdb <dbPrefix> <out.fa> --entry ID", file=sys.stderr)
+            return 1
+        cmd = [blastdbcmd, "-db", pos[0], "-entry", entry, "-out", pos[1]]
+        out_fa = pos[1]
+    else:
+        if len(pos) < 3:
+            print("用法: getseqdb <dbPrefix> <idList.txt> <out.fa> [--entry ID]", file=sys.stderr)
+            return 1
+        if not os.path.isfile(pos[1]):
+            print(f"❌ ID 列表不存在: {pos[1]}", file=sys.stderr)
+            return 2
+        cmd = [blastdbcmd, "-db", pos[0], "-entry_batch", pos[1], "-out", pos[2]]
+        out_fa = pos[2]
+    r = subprocess.run(cmd, capture_output=True, text=True)
+    if r.returncode != 0 or not os.path.isfile(out_fa):
+        print(f"❌ blastdbcmd 失败:\n{r.stderr[-400:]}", file=sys.stderr)
+        return r.returncode or 1
+    if not quiet:
+        n = sum(1 for l in open(out_fa) if l.startswith(">"))
+        print(f"[getseqdb] 提取 {n} 条 → {out_fa}", file=sys.stderr)
+    return 0
+
+
+def _genomefilter_impl(args, verbose=False, quiet=False):
+    """genomefilter: genomefilter <in.fa> <out.fa> --min-len <N> [--gxf <in.gff3>]   # 按序列长度过滤（GUI 逆向 #19 GenomeLengthFilterGUIPanel：QuickStatFasta 统计 → 按 minLen 过滤 ID → ExtractFasta 提取；可选 GXF 同过滤）"""
+    pos, min_len, gxf = [], None, None
+    _kw, pos, _ = parse_kv_args(args, {"--min-len": ("min_len", int), "--gxf": ("gxf", None)})
+    min_len = _kw.get("min_len"); gxf = _kw.get("gxf")
+    if len(pos) < 2 or min_len is None:
+        print("用法: genomefilter <in.fa> <out.fa> --min-len <N> [--gxf <in.gff3>]", file=sys.stderr)
+        return 1
+    in_fa, out_fa = pos[0], pos[1]
+    if not os.path.isfile(in_fa):
+        print(f"❌ 输入文件不存在: {in_fa}", file=sys.stderr)
+        return 2
+    # 1) QuickStatFasta 统计长度
+    stat_out = os.path.join(tempfile.gettempdir(), f"genomefilter_stat_{os.getpid()}.txt")
+    r = subprocess.run(["java", "-Xmx2g", "-cp", JAR,
+                        "biocjava.bioIO.FastX.FastaIndex.QuickStatFasta",
+                        "--inFasta", in_fa, "--outPutFile", stat_out],
+                       capture_output=True, text=True)
+    if r.returncode != 0 or not os.path.isfile(stat_out):
+        print(f"❌ statFasta 失败:\n{r.stderr[-400:]}", file=sys.stderr)
+        return r.returncode or 1
+    # 2) 过滤 ID（表头: Original_ID Simplified_ID ... Length）
+    keep_ids = []
+    with open(stat_out, encoding="utf-8", errors="replace") as f:
+        header = f.readline().rstrip("\n").split("\t")
+        try:
+            len_idx = header.index("Length")
+            id_idx = 0
+        except ValueError:
+            len_idx, id_idx = 3, 0
+        for line in f:
+            if not line.strip():
+                continue
+            cols = line.rstrip("\n").split("\t")
+            if len(cols) <= len_idx:
+                continue
+            try:
+                ln = int(cols[len_idx])
+            except ValueError:
+                continue
+            if ln >= min_len:
+                keep_ids.append(cols[id_idx])
+    if not keep_ids:
+        print(f"⚠️ 无序列 ≥ {min_len} bp，输出为空", file=sys.stderr)
+        return 1
+    id_list = os.path.join(tempfile.gettempdir(), f"genomefilter_ids_{os.getpid()}.txt")
+    with open(id_list, "w", encoding="utf-8") as f:
+        f.write("\n".join(keep_ids) + "\n")
+    # 3) ExtractFasta 按 ID 提取
+    r = subprocess.run(["java", "-Xmx2g", "-cp", JAR,
+                        "biocjava.bioDoer.Fasta.ExtractFasta",
+                        "--inFa", in_fa, "--inIDList", id_list, "--outFa", out_fa,
+                        "--processMode", "Extract", "--matchMode", "Match", "--caseInSensitive", "false"],
+                       capture_output=True, text=True)
+    os.unlink(id_list)
+    try: os.unlink(stat_out)
+    except OSError as _e:
+        logging.getLogger(__name__).debug("io: %s", _e)
+        pass
+    if r.returncode != 0 or not os.path.isfile(out_fa):
+        print(f"❌ ExtractFasta 失败:\n{r.stderr[-400:]}", file=sys.stderr)
+        return r.returncode or 1
+    # 4) 可选 GXF 同过滤
+    if gxf:
+        if os.path.isfile(gxf):
+            keep_set = set(keep_ids)
+            out_gxf = out_fa + ".gxf"
+            with open(gxf, encoding="utf-8", errors="replace") as fi, open(out_gxf, "w", encoding="utf-8") as fo:
+                for line in fi:
+                    if not line.strip() or line.startswith("#"):
+                        continue
+                    chr_name = line.split("\t", 1)[0].strip()
+                    if chr_name in keep_set:
+                        fo.write(line)
+            if not quiet:
+                print(f"[genomefilter] GXF 同过滤: {out_gxf}", file=sys.stderr)
+        else:
+            print(f"⚠️ GXF 文件不存在，跳过: {gxf}", file=sys.stderr)
+    if not quiet:
+        print(f"[genomefilter] {len(keep_ids)}/{len(open(in_fa).readlines())} 序列保留 → {out_fa}", file=sys.stderr)
+    return 0
+
+
+def _notung_impl(args, verbose=False, quiet=False):
+    """notung: notung <gene.nwk> -s <species.nwk> --reconcile [Notung 原生参数]   # 基因树-物种树 reconcile（duplication/loss 推断，插件 P00651 CLI 化）"""
+    notung = os.path.join(ROOT, "plugins", "lib", "Notung-2.9.1.5.jar")
+    if not os.path.isfile(notung):
+        print(f"❌ Notung 引擎缺失: {notung}", file=sys.stderr)
+        return 1
+    java_args = ["java", "-Xmx2g", "-jar", notung] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="notung")
+
+
+def _newickRename_impl(args, verbose=False, quiet=False):
+    """newickRename: newickRename --inNwk <tree.nwk> --renameMap <map.tsv> --outNwk <out.nwk>   # 树叶批量重命名（插件 P00690 CLI 化，map 为 OldName\\tNewName）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_NewickRenamer.jar")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    java_args = ["java", "-Xmx1g", "-cp", f"{pjar}:{JAR}", "newickRenamer.NewickRenamer"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="newickRename")
+
+
+def _hmmerSearch_impl(args, verbose=False, quiet=False):
+    """hmmerSearch: hmmerSearch <target.fa> <hmmDb> <out.tsv>   # Advanced HMMer 全库扫描+domtblout 解析（插件 P00680 CLI 化，无需 idList）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_HmmerSuite.jar")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    ensure_bridge("HmmerSuiteCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, pjar, JAR), "HmmerSuiteCli"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="hmmerSearch")
+
+
+def _memeViz_impl(args, verbose=False, quiet=False):
+    """memeViz: memeViz <meme.xml> <out.svg> [width] [height]   # MEME motif 批量可视化（插件 P00700 CLI 化，每 motif 一面板）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Batch_MEME_Motif_Viz.jar")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    ensure_bridge("BatchVizMotifsCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, pjar, JAR), "BatchVizMotifsCli"] + args
+    return run_plot(java_args, verbose=verbose, quiet=quiet, command_name="memeViz")
+
+
+def _gsea_impl(args, verbose=False, quiet=False):
+    """gsea: gsea <go.obo> <query2go.tsv> <rank.rnk> <outDir>   # GO 预排序 GSEA（插件 P00342 CLI 化，GSEAPreranked 全套报告；⚠️ set_min=15 小基因集会被过滤）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_GSEAWrapper.jar")
+    dep = os.path.join(ROOT, "plugins", "lib", "Dependency")
+    if not (os.path.isfile(pjar) and os.path.isdir(dep)):
+        print(f"❌ GSEA 插件或 Dependency 缺失: {pjar}", file=sys.stderr)
+        return 1
+    ensure_bridge("GSEAWrapperCli")
+    java_args = ["java", "-Xmx4g", "-cp", cp(BUILD_DIR, pjar, JAR), "GSEAWrapperCli"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="gsea")
+
+
+def _tfbsShift_impl(args, verbose=False, quiet=False):
+    """tfbsShift: tfbsShift <query.pep> <outPrefix> [threads]   # 植物 TF 结合 motif 偏移分析（插件 P00551 CLI 化，参考数据内置 ath.pep+binding.motifs）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_PlantTFbindingMotifShift.jar")
+    ath = os.path.join(ROOT, "plugins", "lib", "plantTF", "ath.pep")
+    motifs = os.path.join(ROOT, "plugins", "lib", "plantTF", "binding.motifs")
+    if not (os.path.isfile(pjar) and os.path.isfile(ath) and os.path.isfile(motifs)):
+        print(f"❌ TFBS 插件或参考数据缺失: {pjar}", file=sys.stderr)
+        return 1
+    ensure_bridge("MotifShiftCli")
+    java_args = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, pjar, JAR), "MotifShiftCli", ath, motifs] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="tfbsShift")
+
+
+def _mcscanxd_impl(args, verbose=False, quiet=False):
+    """mcscanxd: mcscanxd <wkDir> <genome1.fa> <genome2.fa> <gxf1> <gxf2> [threads] [blastHits] [evalue]   # OneStep MCScanX-SuperFast（插件 P00370 CLI 化，diamond 加速，二进制随包）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_OneStepMCScanX_Diamond.jar")
+    bin_dir = os.path.join(ROOT, "plugins", "lib", "bin")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    if not os.path.isfile(os.path.join(bin_dir, "diamond")):
+        print(f"❌ diamond 二进制缺失: {bin_dir}/diamond", file=sys.stderr)
+        return 1
+    os.environ["PATH"] = bin_dir + os.pathsep + os.environ.get("PATH", "")
+    ensure_bridge("MCScanXFastCli")
+    java_args = ["java", "-Xmx4g", "-cp", cp(BUILD_DIR, pjar, JAR), "MCScanXFastCli"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="mcscanxd")
+
+
+def _quickAnno_impl(args, verbose=False, quiet=False):
+    """quickAnno: quickAnno <query.pep> <swissprotDb.fa> <out.txt> [threads] [maxHits]   # diamond 蛋白快速注释（插件 P00480 CLI 化；⚠️ db 需带描述行，否则 Top 词频为空报错）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_QuickProteinAnno.jar")
+    bin_dir = os.path.join(ROOT, "plugins", "lib", "bin")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    if not os.path.isfile(os.path.join(bin_dir, "diamond")):
+        print(f"❌ diamond 二进制缺失: {bin_dir}/diamond", file=sys.stderr)
+        return 1
+    os.environ["PATH"] = bin_dir + os.pathsep + os.environ.get("PATH", "")
+    ensure_bridge("QuickProteinAnnoCli")
+    java_args = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, pjar, JAR), "QuickProteinAnnoCli"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="quickAnno")
+
+
+def _smart_impl(args, verbose=False, quiet=False):
+    """smart: smart <in.fa> <out.txt>   # SMART 域注释（插件 P00060 CLI 化；⚠️ 联网 POST EMBL ismart.embl.de，约 10-60s，输出域位置+类型）"""
+    pjar = os.path.join(ROOT, "plugins", "lib", "Plugin_BatchSMART.jar")
+    if not os.path.isfile(pjar):
+        print(f"❌ 插件缺失: {pjar}", file=sys.stderr)
+        return 1
+    ensure_bridge("SubmitSMARTCli")
+    java_args = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, pjar, JAR), "SubmitSMARTCli"] + args
+    return run_java(java_args, verbose=verbose, quiet=quiet, command_name="smart")
+
+
+# ── P0/P1 修复实现（N19/N23/N25：覆盖错误的注册表/表驱动路径，覆盖顺序=后定义者胜）──
+
+def _findBestHomologyBatch_impl(args, verbose=False, quiet=False):
+    """findBestHomologyBatch: findBestHomologyBatch <query.pep> <subject.pep> <outDir> [--targetIds ID[,ID2...]] [--threads N] [--plot false] [--sensitive 5,10]
+       # 最优同源批量查找（FindBestHomologyBatch；N19 修复）
+       # ⚠️ 引擎真实参数: --inQueryProteinSet/--inSubjectProteinSet/--targetIdList(GeneName\tID1[,ID2])/--outDir；
+       #    旧写法 --queryFasta/--subjectFasta/--outTable 是错误参数名，引擎拒参仍返 ec=0（静默成功+输出不创建）。
+       # ⚠️ 引擎不自动创建 --outDir；targetIdList 必需，缺省时自动取 query 全部 ID（targetName=All）。
+       #    产物: <outDir>/<targetName>.s<subjectId>.ids（最佳同源 ID 对，逗号分隔）。"""
+    pos, kw = [], {}
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a.startswith("--") and i + 1 < len(args):
+            kw[a[2:]] = args[i + 1]
+            i += 2
+        else:
+            pos.append(a)
+            i += 1
+    query = kw.get("inQueryProteinSet") or kw.get("queryFasta") or (pos[0] if pos else None)
+    subject = kw.get("inSubjectProteinSet") or kw.get("subjectFasta") or (pos[1] if len(pos) > 1 else None)
+    out_dir = kw.get("outDir") or kw.get("outTable") or (pos[2] if len(pos) > 2 else None)
+    if not query or not subject or not out_dir:
+        print("用法: findBestHomologyBatch <query.pep> <subject.pep> <outDir> [--targetIds ID,...] [--threads N] [--plot false]", file=sys.stderr)
+        return 1
+    for f in (query, subject):
+        if not os.path.isfile(f):
+            print(f"❌ 输入文件不存在: {f}", file=sys.stderr)
+            return 2
+    try:
+        os.makedirs(out_dir, exist_ok=True)  # 引擎不自动创建
+    except OSError as e:
+        print(f"❌ 无法创建 outDir: {out_dir}: {e}", file=sys.stderr)
+        return 2
+    target_list = kw.get("targetIdList") or kw.get("targetIds")
+    tmp_targets = None
+    if not (target_list and os.path.isfile(target_list)):
+        ids = [l[1:].split()[0] for l in open(query, encoding="utf-8", errors="replace") if l.startswith(">")]
+        if not ids:
+            print("❌ query FASTA 无序列头", file=sys.stderr)
+            return 3
+        tmp_targets = safe_temp(prefix="tb_fbh.", suffix=".tsv")
+        with open(tmp_targets, "w") as fh:
+            fh.write(f"{kw.get('targetName') or 'All'}\t{','.join(ids)}\n")
+        target_list = tmp_targets
+    try:
+        jargs = ["java", "-Xmx4g", "-cp", JAR,
+                 "biocjava.bioIO.BioSoftPipeServer.FindBestHomologyBatch",
+                 "--inQueryProteinSet", query, "--inSubjectProteinSet", subject,
+                 "--targetIdList", target_list, "--outDir", out_dir,
+                 "--useDiamond", str(kw.get("useDiamond", "false")).lower(),
+                 "--threads", str(kw.get("threads", 2)),
+                 "--plot", str(kw.get("plot", "false")).lower()]
+        for opt in ("sensitive", "similarity", "weightCov", "extendClade"):
+            if opt in kw:
+                jargs += [f"--{opt}", kw[opt]]
+        ec = run_java(jargs, verbose=verbose, quiet=quiet, command_name="findBestHomologyBatch")
+        # 防引擎拒参仍 ec=0（N19 静默成功）: 校验产物
+        if ec == 0:
+            outs = [f for f in os.listdir(out_dir) if f.endswith(".ids")] if os.path.isdir(out_dir) else []
+            if not outs:
+                print(f"❌ 引擎未产出结果（检查 --targetIds 与输入格式），outDir={out_dir}", file=sys.stderr)
+                ec = 1
+        return ec
+    finally:
+        if tmp_targets:
+            try:
+                os.unlink(tmp_targets)
+            except Exception:
+                pass
+
+
+def _gffCdsPhaseCorrector_impl(args, verbose=False, quiet=False):
+    """gffCdsPhaseCorrector: gffCdsPhaseCorrector --inGff <in.gff3> --outGff <out.gff3> [--problemGff <p.gff3>] [--report <r.txt>]
+       # CDS phase 校正（GffCdsPhaseCorrector；N25 修复：引擎为位置参数式 <in> <correct> <problematic> <report>，
+       #    原注册表直通导致 --inGff/--outGff 被当字面量文件、警告覆盖输入——现显式转位置参数并避免覆盖已存在输出）。"""
+    pos, kw = [], {}
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a.startswith("--") and i + 1 < len(args):
+            kw[a[2:]] = args[i + 1]
+            i += 2
+        else:
+            pos.append(a)
+            i += 1
+    in_gff = kw.get("inGff") or (pos[0] if pos else None)
+    out_gff = kw.get("outGff") or (pos[1] if len(pos) > 1 else None)
+    if not in_gff or not out_gff:
+        print("用法: gffCdsPhaseCorrector --inGff <in.gff3> --outGff <out.gff3>", file=sys.stderr)
+        return 1
+    if not os.path.isfile(in_gff):
+        print(f"❌ 输入文件不存在: {in_gff}", file=sys.stderr)
+        return 2
+    if os.path.abspath(in_gff) == os.path.abspath(out_gff):
+        print("❌ inGff 与 outGff 不能相同（引擎会覆盖输入，N25）", file=sys.stderr)
+        return 2
+    prob = kw.get("problemGff") or (out_gff + ".problem.gff3")
+    rep = kw.get("report") or (out_gff + ".report.txt")
+    od = os.path.dirname(os.path.abspath(out_gff))
+    if od:
+        os.makedirs(od, exist_ok=True)
+    jargs = ["java", "-Xmx2g", "-cp", JAR,
+             "biocjava.bioDoer.GXFUtils.GffCdsPhase.GffCdsPhaseCorrector",
+             in_gff, out_gff, prob, rep]
+    return run_java(jargs, verbose=verbose, quiet=quiet, command_name="gffCdsPhaseCorrector")
+
+
+def _mirnatarget_impl(args, verbose=False, quiet=False):
+    """mirnatarget: mirnatarget <mirna.fa> <target.fa> <out.tsv> [--evalue X]
+       # miRNA 靶标预测完整管线（N23 修复）：ssearch36 -w 100 -W 25 -E X -m 10 -T 1 -i -U <mirna> <target> → TargetScoreCli
+       #   原表驱动误把本命令直通 TargetScoreCli（其输入是 ssearch36 的 m10，不是 FASTA），
+       #   导致第 2 参数被当输出清零、第 3 参被忽略、输出永不落盘——现恢复完整管线并校验输出。
+       # 依赖: ssearch36（fasta36 套件）。产物: miRNA\ttarget\tstrand\tbeg\tend\tscore\t..."""
+    pos, evalue, extra = [], 1.0, []
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a == "--evalue" and i + 1 < len(args):
+            try:
+                evalue = float(args[i + 1])
+            except ValueError:
+                pass
+            i += 2
+        elif a.startswith("--") and i + 1 < len(args):
+            extra += [a, args[i + 1]]
+            i += 2
+        else:
+            pos.append(a)
+            i += 1
+    if len(pos) < 3:
+        print("用法: mirnatarget <mirna.fa> <target.fa> <out.tsv> [--evalue X]", file=sys.stderr)
+        return 1
+    mirna, target, out = pos[0], pos[1], pos[2]
+    for f in (mirna, target):
+        if not os.path.isfile(f):
+            print(f"❌ 输入文件不存在: {f}", file=sys.stderr)
+            return 2
+    if os.path.abspath(out) in (os.path.abspath(mirna), os.path.abspath(target)):
+        print("❌ 输出文件与输入同名会覆盖输入（N23），请换输出名", file=sys.stderr)
+        return 2
+    ssearch = shutil.which("ssearch36")
+    if not ssearch:
+        print("❌ 缺少依赖 ssearch36（fasta36 套件），请先安装", file=sys.stderr)
+        return 4
+    od = os.path.dirname(os.path.abspath(out))
+    if od:
+        os.makedirs(od, exist_ok=True)
+    tmp_m10 = safe_temp(prefix="tb_mirna.", suffix=".m10")
+    try:
+        with open(tmp_m10, "w") as fh:
+            r = subprocess.run([ssearch, "-w", "100", "-W", "25", "-E", str(evalue),
+                                "-m", "10", "-T", "1", "-i", "-U", mirna, target],
+                               stdout=fh, stderr=subprocess.PIPE)
+        if r.returncode != 0 or not os.path.isfile(tmp_m10) or os.path.getsize(tmp_m10) == 0:
+            print(f"❌ ssearch36 未产出比对（0 命中或序列格式不符；evalue={evalue}）", file=sys.stderr)
+            return 3
+        ensure_bridge("TargetScoreCli")
+        jargs = ["java", "-Xmx2g", "-cp", cp(BUILD_DIR, JAR), "TargetScoreCli", tmp_m10, out] + extra
+        ec = run_java(jargs, verbose=verbose, quiet=quiet, command_name="mirnatarget")
+        if ec == 0 and not os.path.isfile(out):
+            print(f"❌ 输出文件未生成: {out}", file=sys.stderr)
+            ec = 1
+        return ec
+    finally:
+        try:
+            os.unlink(tmp_m10)
+        except Exception:
+            pass
+
+
+
+def _msy_impl(args, verbose=False, quiet=False):
+    """msy: msy <simplifiedGff.pos> <links.txt> <chrLayout.txt> <out> [w] [h]
+       # 多物种微共线性图 / microsynteny / collinearity plot（N26 修复：原表驱动把 msy 注册为裸 GenericCli 透传，用户参数被当
+       #   engineClass 导致 ClassNotFoundException——现按 tbplot.sh 已验证调用方式显式拼参数）
+       # 格式: pos=Chr\\tGene\\tStart\\tEnd；links=GeneA\\tGeneB\\t[r,g,b]；layout=Genome: chr1 chr2"""
+    if len(args) < 4:
+        print("用法: msy <simplifiedGff.pos> <links.txt> <chrLayout.txt> <out> [w] [h]", file=sys.stderr)
+        return 1
+    pos, links, layout, out = args[0], args[1], args[2], args[3]
+    w, h = "1000", "800"
+    if len(args) >= 5:
+        w = args[4]
+    if len(args) >= 6:
+        h = args[5]
+    for f in (pos, links, layout):
+        if not os.path.isfile(f):
+            print(f"❌ 输入文件不存在: {f}", file=sys.stderr)
+            return 2
+    ensure_bridge("GenericCli")
+    jargs = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, JAR), "GenericCli",
+             "biocjava.bioDoer.JIGplotToolkit.Synteny.MultipleSpeciesSyteny", "plot", out,
+             "--set", "inSimplifiedGff", pos, "--set", "genePairInfoFile", links,
+             "--set", "chrLayoutFile", layout, "--width", w, "--height", h]
+    return run_plot(jargs, verbose=verbose, quiet=quiet, command_name="msy")
+
+
+def _getLongestCompleteORF_impl(args, verbose=False, quiet=False):
+    """getLongestCompleteORF: getLongestCompleteORF --inFa <seq.fa> --outORFs <out.fa>
+       # 批量最长完整 ORF 预测（N24 修复：原注册到 JavaFX 无 main 的 biocjava.bioIO.ORF.ORF，
+       #   改映射到 GetLongestORF（longestorf 同引擎，ArgsParser --inFa/--outORFs 实测可用））"""
+    return _longestorf_impl(args, verbose=verbose, quiet=quiet)  # type: ignore[name-defined]  # noqa: F821
+
+
+def _efpHeat_impl(args, verbose=False, quiet=False):
+    """efpHeat: efpHeat --inTGA <plant.tga> --inSample2CC <sample2cc.txt> --expMat <expMat.tsv> --geneId <ID> --outImg <out>
+       # eFP 组织表达热图（单矩阵，generateSuperHeatMap）——N27 附带修复：引擎是 --key value 式，
+       #   旧 docstring 写成位置参数导致拒参；支持位置参数 <tga> <sample2cc> <expmat> <geneId> <out> 自动转换
+       # ⚠️ 需 fake DatatypeConverter（build/，JDK9+ 无 javax.xml.bind，ensure_bridge 自动重建）"""
+    has_flag = any(a.startswith("--") for a in args)
+    if has_flag:
+        # 命名参数直通（TGA 参数校验交给引擎）
+        jargs = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, JAR),
+                 "biocjava.bioDoer.SimpleEfpBrowser.generateSuperHeatMap"] + args
+        return run_plot(jargs, verbose=verbose, quiet=quiet, command_name="efpHeat")
+    pos = [a for a in args if not a.startswith("--")]
+    if len(pos) < 5:
+        print("用法: efpHeat --inTGA <plant.tga> --inSample2CC <s2cc.txt> --expMat <exp.tsv> --geneId <ID> --outImg <out>", file=sys.stderr)
+        return 1
+    tga, s2cc, exp, gid, out = pos[0], pos[1], pos[2], pos[3], pos[4]
+    for f in (tga, s2cc, exp):
+        if not os.path.isfile(f):
+            print(f"❌ 输入文件不存在: {f}", file=sys.stderr)
+            return 2
+    jargs = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, JAR),
+             "biocjava.bioDoer.SimpleEfpBrowser.generateSuperHeatMap",
+             "--inTGA", tga, "--inSample2CC", s2cc, "--expMat", exp,
+             "--geneId", gid, "--outImg", out]
+    return run_plot(jargs, verbose=verbose, quiet=quiet, command_name="efpHeat")
+
+
+def _tableMerge_impl(args, verbose=False, quiet=False):
+    """tableMerge: tableMerge --inFileArr \"f1,f2,...\" --inColIndexArr \"0,1,...\" --outTable <out> [--defaultNAvalue NA] [--appendMergedKey true|false] [--rmKeyColumns true|false]
+       # 按键合并多个表格（N3 修复：引擎为 ArgsParser 式 --inFileArr/--inColIndexArr/--outTable，
+       #   旧 docstring 位置参数写法 <outTable> <inFile1> [...] 与引擎完全对不上；本 impl 兼容位置参数自动转换）"""
+    pos, kw = [], {}
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a.startswith("--") and i + 1 < len(args):
+            kw[a[2:]] = args[i + 1]
+            i += 2
+        else:
+            pos.append(a)
+            i += 1
+    # 位置参数兼容: <outTable> <inFile1> [<inFile2>...] → --outTable + --inFileArr
+    if "inFileArr" not in kw and len(pos) >= 2:
+        kw["outTable"] = pos[0]
+        kw["inFileArr"] = ",".join(pos[1:])
+        kw.setdefault("inColIndexArr", ",".join("0" for _ in pos[1:]))
+    if "inFileArr" not in kw or "outTable" not in kw:
+        print("用法: tableMerge --inFileArr \"f1,f2,...\" --inColIndexArr \"0,1,...\" --outTable <out>", file=sys.stderr)
+        print("   （旧位置参数写法 <outTable> <inFile1> [...] 已兼容自动转换）", file=sys.stderr)
+        return 1
+    jargs = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, JAR), "biocjava.bioDoer.Table.TableMerger",
+             "--inFileArr", kw["inFileArr"],
+             "--inColIndexArr", kw.get("inColIndexArr", "0"),
+             "--outTable", kw["outTable"]]
+    for opt in ("defaultNAvalue", "appendMergedKey", "rmKeyColumns", "appendOnly"):
+        if opt in kw:
+            jargs += [f"--{opt}", kw[opt]]
+    return run_java(jargs, verbose=verbose, quiet=quiet, command_name="tableMerge")
+
+
+def _parallelMD5Check_impl(args, verbose=False, quiet=False):
+    """parallelMD5Check: parallelMD5Check <md5_list.txt> [threads]
+       # 并行校验 MD5 列表（N6 修复：位置参数式 <md5_list_file> [threads]；
+       #   每行 `md5sum  文件名`，与 md5sum -c 同格式；原 registry 直通在参数形态错误时 ec=3）"""
+    if not args:
+        print("用法: parallelMD5Check <md5_list.txt> [threads]", file=sys.stderr)
+        print("   每行: <md5sum>  <文件名>（与 md5sum -c 同格式）", file=sys.stderr)
+        return 1
+    lst = args[0]
+    if not os.path.isfile(lst):
+        print(f"❌ MD5 列表文件不存在: {lst}", file=sys.stderr)
+        return 2
+    threads = "24"
+    if len(args) >= 2:
+        threads = args[1]
+        if not threads.isdigit():
+            print(f"❌ threads 须为数字: {threads}", file=sys.stderr)
+            return 2
+    jargs = ["java", "-Xmx2g", "-cp", JAR, "biocjava.bioDoer.FileUtils.ParallelMD5Check", lst, threads]
+    return run_java(jargs, verbose=verbose, quiet=quiet, command_name="parallelMD5Check")
+
+
+def _extractFeatureFromGTF_impl(args, verbose=False, quiet=False):
+    """extractFeatureFromGTF: extractFeatureFromGTF --inGtf <in.gtf> --inGenome <genome.fa> --outFile <out> [--targetFeature CDS|exon|...] [--targetIdTag transcript_id] [--retainAttr true|false]
+       # GTF 特征提取（N6 修复：引擎 main 硬编码 Windows 默认路径，但 toolsKit.ArgsParser 可覆盖 → 走 ArgsParser 路线；
+       #   旧注册表直通裸 main 必 FileNotFoundException；位置参数 <in.gtf> <genome.fa> <out> 兼容转换）
+       # ⚠️ 输入 GTF 坐标须与 genome 匹配（越界会 NPE）；输出 FASTA 头含 +/− 链/Location 注释"""
+    pos, kw = [], {}
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a.startswith("--") and i + 1 < len(args):
+            kw[a[2:]] = args[i + 1]
+            i += 2
+        else:
+            pos.append(a)
+            i += 1
+    in_gtf = kw.get("inGtf") or (pos[0] if pos else None)
+    genome = kw.get("inGenome") or (pos[1] if len(pos) > 1 else None)
+    out = kw.get("outFile") or (pos[2] if len(pos) > 2 else None)
+    if not in_gtf or not genome or not out:
+        print("用法: extractFeatureFromGTF --inGtf <in.gtf> --inGenome <genome.fa> --outFile <out> [--targetFeature CDS] [--targetIdTag transcript_id] [--retainAttr true]", file=sys.stderr)
+        return 1
+    for f in (in_gtf, genome):
+        if not os.path.isfile(f):
+            print(f"❌ 输入文件不存在: {f}", file=sys.stderr)
+            return 2
+    jargs = ["java", "-Xmx3g", "-cp", cp(BUILD_DIR, JAR), "biocjava.bioIO.GTF.ExtractFeaturefromGTFandGenome",
+             "--inGtf", in_gtf, "--inGenome", genome, "--outFile", out,
+             "--targetFeature", kw.get("targetFeature", "exon"),
+             "--targetIdTag", kw.get("targetIdTag", "transcript_id"),
+             "--retainAttr", kw.get("retainAttr", "false")]
+    for opt in ("onlyCheck", "maxFeatureCounts", "minFeatureCounts"):
+        if opt in kw:
+            jargs += [f"--{opt}", kw[opt]]
+    return run_java(jargs, verbose=verbose, quiet=quiet, command_name="extractFeatureFromGTF")
+
+
+def _gxfSplit_impl(args, verbose=False, quiet=False):
+    """gxfSplit: gxfSplit <in.gff3|gtf> <outPrefix> [--numOfFile N]
+       # GXF 按记录数拆分（§8.B N13 家族命令面缺口修复：RPC GxfSplit.process，走 RPC 自动拉起）"""
+    import json as _json
+    import urllib.request as _ur
+    pos, num = [], 8
+    i = 0
+    while i < len(args):
+        a = args[i]
+        if a == "--numOfFile" and i + 1 < len(args):
+            num = int(args[i + 1]); i += 2
+        else:
+            pos.append(a); i += 1
+    if len(pos) < 2:
+        print("用法: gxfSplit <in.gff3|gtf> <outPrefix> [--numOfFile N]", file=sys.stderr)
+        return 1
+    inp, prefix = pos[0], pos[1]
+    if not os.path.isfile(inp):
+        print(f"❌ 输入文件不存在: {inp}", file=sys.stderr)
+        return 2
+    from tbtools_cli.core import run_java  # noqa
+    # 走 RPC（服务器按需拉起，见 rpc call --no-autostart 之外的自动拉起）
+    from tbtools_cli.cli_rpc import _ensure_rpc  # 批次B: rpc 段已拆出
+    ok = _ensure_rpc(8765)
+    if not ok:
+        print("❌ RPC 服务器不可用，无法调用 GxfSplit", file=sys.stderr)
+        return 1
+    body = _json.dumps({"jsonrpc":"2.0","method":"GxfSplit.process",
+        "params":{"inputPath": inp, "outputPrefix": prefix, "numOfFile": num}, "id":1}).encode()
+    req = _ur.Request("http://127.0.0.1:8765/rpc", data=body, headers={"Content-Type":"application/json"})
+    resp = _json.loads(_ur.urlopen(req, timeout=120).read())
+    if resp.get("error"):
+        print(f"❌ RPC 错误: {resp['error']}", file=sys.stderr)
+        return 1
+    r = resp.get("result", {})
+    print(f"✅ 拆分完成 {num} 份 × 前缀 {prefix}（{r.get('outputPrefix', '')}）", file=sys.stderr)
+    return 0
+
+
+def _gxfIdAppender_impl(args, verbose=False, quiet=False):
+    """gxfIdAppender: gxfIdAppender <in.gff3|gtf> <out.gff3> <prefix>
+       # GXF ID/染色体名前缀追加（§8.B N13 家族命令面缺口修复：RPC GxfIdAppender.process）"""
+    import json as _json
+    import urllib.request as _ur
+    pos = [a for a in args if not a.startswith("--")]
+    if len(pos) < 3:
+        print("用法: gxfIdAppender <in.gff3|gtf> <out.gff3> <prefix>", file=sys.stderr)
+        return 1
+    inp, out, prefix = pos[0], pos[1], pos[2]
+    if not os.path.isfile(inp):
+        print(f"❌ 输入文件不存在: {inp}", file=sys.stderr)
+        return 2
+    from tbtools_cli.cli_rpc import _ensure_rpc  # 批次B: rpc 段已拆出
+    ok = _ensure_rpc(8765)
+    if not ok:
+        print("❌ RPC 服务器不可用，无法调用 GxfIdAppender", file=sys.stderr)
+        return 1
+    body = _json.dumps({"jsonrpc":"2.0","method":"GxfIdAppender.process",
+        "params":{"inputPath": inp, "outputPath": out, "prefix": prefix}, "id":1}).encode()
+    req = _ur.Request("http://127.0.0.1:8765/rpc", data=body, headers={"Content-Type":"application/json"})
+    resp = _json.loads(_ur.urlopen(req, timeout=120).read())
+    if resp.get("error"):
+        print(f"❌ RPC 错误: {resp['error']}", file=sys.stderr)
+        return 1
+    print(f"✅ ID 前缀追加完成: {out}", file=sys.stderr)
+    return 0