@platforma-open/milaboratories.mixcr-clonotyping-2.workflow 3.0.3 → 3.1.0

package/package.json

@@ -1,16 +1,16 @@
 {
   "name": "@platforma-open/milaboratories.mixcr-clonotyping-2.workflow",
-  "version": "3.0.3",
+  "version": "3.1.0",
   "description": "Tengo-based template",
   "dependencies": {
-    "@platforma-sdk/workflow-tengo": "^5.0.2",
+    "@platforma-sdk/workflow-tengo": "^5.0.3",
     "@platforma-open/milaboratories.software-mixcr": "4.7.0-223-develop"
   },
   "devDependencies": {
     "@platforma-sdk/tengo-builder": "^2.1.15",
     "@platforma-sdk/test": "^1.42.20",
     "vitest": "~2.1.9",
-    "typescript": "~5.5.4"
+    "typescript": "~5.6.3"
   },
   "scripts": {
     "build": "rm -rf dist && pl-tengo check && pl-tengo build",

package/src/export-report.tpl.tengo

@@ -0,0 +1,115 @@
+self := import("@platforma-sdk/workflow-tengo:tpl")
+smart := import("@platforma-sdk/workflow-tengo:smart")
+ll := import("@platforma-sdk/workflow-tengo:ll")
+exec := import("@platforma-sdk/workflow-tengo:exec")
+assets := import("@platforma-sdk/workflow-tengo:assets")
+pcolumn := import("@platforma-sdk/workflow-tengo:pframes.pcolumn")
+times := import("times")
+text := import("text")
+maps := import("@platforma-sdk/workflow-tengo:maps")
+pframes := import("@platforma-sdk/workflow-tengo:pframes")
+xsv := import("@platforma-sdk/workflow-tengo:pframes.xsv")
+slices := import("@platforma-sdk/workflow-tengo:slices")
+pConstants := import("@platforma-sdk/workflow-tengo:pframes.constants")
+pt := import("@platforma-sdk/workflow-tengo:pt")
+qcReportColumns := import(":qc-report-columns")
+
+json := import("json")
+
+self.defineOutputs("qcReportTable")
+
+mixcrSw := assets.importSoftware("@platforma-open/milaboratories.software-mixcr:main")
+ptablerSw := assets.importSoftware("@platforma-open/milaboratories.software-ptabler:main")
+
+
+self.body(func(inputs) {
+    clnsData := inputs.clnsData
+    isSingleCell := inputs.isSingleCell
+    hasUmi := inputs.hasUmi
+    sampleIdAxisSpec := inputs.sampleIdAxisSpec
+    chains := inputs.chains
+
+    chainInfos := {
+	"IGHeavy": { mixcrFilter: "IGH", name: "IG Heavy", shortName: "Heavy" },
+	"IGLight": { mixcrFilter: "IGK,IGL", name: "IG Light", shortName: "Light" },
+	"TCRAlpha": { mixcrFilter: "TRA", name: "TCR Alpha", shortName: "Alpha" },
+	"TCRBeta": { mixcrFilter: "TRB", name: "TCR Beta", shortName: "Beta" },
+	"TCRGamma": { mixcrFilter: "TRG", name: "TCR Gamma", shortName: "Gamma" },
+	"TCRDelta": { mixcrFilter: "TRD", name: "TCR Delta", shortName: "Delta" }
+    }
+
+    chainsForMixcr := []
+
+    for chain in chains {
+        chainsForMixcr += text.split(chainInfos[chain].mixcrFilter, ",")
+    }
+    
+    // Get all clns files from the input data
+    clnsFiles := clnsData.inputs()
+    
+    // Build the exportReportsTable command
+    exportReportCmd := exec.builder().
+        software(mixcrSw).
+        secret("MI_LICENSE", "MI_LICENSE").
+        arg("exportReportsTable")
+    
+    // Add all clns files as input
+    for i, clnsFile in clnsFiles {
+        fileName := json.decode(i)[0] + ".clns"
+        exportReportCmd.addFile(fileName, clnsFile)
+        exportReportCmd.arg(fileName)
+    }
+
+    exportReportCmd.arg("qc-report.tsv").saveFile("qc-report.tsv")
+    
+    // Run the command
+    result := exportReportCmd.run()
+
+    rawTsvFile := result.getFile("qc-report.tsv")
+    
+    // Use pTabler to remove ".clns" from fileName column
+    wf := pt.workflow().
+        inMediumQueue().
+        mem("8GiB").
+        cpu(2)
+    
+    // Load the raw TSV file as a DataFrame
+    df := wf.frame(rawTsvFile, {
+        xsvType: "tsv",
+        inferSchema: true
+    })
+    
+        // Add the sampleId column and remove ".clns" from fileName using pTabler string methods
+    processedDf := df.withColumns(
+        pt.col("fileName").strSlice(0, pt.col("fileName").strLenChars().minus(5)).alias("sampleId")
+    )
+
+    // For now, let's just use the processed DataFrame without zero column filtering
+    // since the columns() method is not available in this pTabler version
+    finalDf := processedDf
+    
+    // Save the final DataFrame back to TSV
+    finalDf.save("qc-report-processed.tsv", {
+        xsvType: "tsv"
+    })
+    
+    // Run the pTabler workflow
+    wfResult := wf.run()
+    
+    tsvFile := wfResult.getFile("qc-report-processed.tsv")
+
+    qcReportColumns := qcReportColumns(hasUmi, isSingleCell, sampleIdAxisSpec, chainsForMixcr)
+    reportColumnsSpec := qcReportColumns.reportColumnsSpec
+
+    qcReportTable := xsv.importFile(
+		tsvFile,
+		"tsv",
+		reportColumnsSpec,
+		{ cpu: 1, mem: "16GiB" }
+	)
+
+    
+    return {
+        qcReportTable: qcReportTable
+    }
+})

package/src/main.tpl.tengo

@@ -11,6 +11,7 @@ pframes := import("@platforma-sdk/workflow-tengo:pframes")
 
 calculatePresetInfoTpl := assets.importTemplate(":calculate-preset-info")
 processTpl := assets.importTemplate(":process")
+exportReportTpl := assets.importTemplate(":export-report")
 
 wf.setPreRun(assets.importTemplate(":prerun"))
 
@@ -176,6 +177,10 @@ wf.body(func(args) {
 	if !is_undefined(libraryImportHandle) {
 		outputs.libraryImportHandle = libraryImportHandle
 	}
+	
+	qcReportTable := runMixcr.output("qcReportTable")
+
+	outputs.qcReportTable = pframes.exportFrame(qcReportTable)
 
 	return {
 		outputs: outputs,

package/src/process.tpl.tengo

@@ -1,7 +1,7 @@
 // process
 
 self := import("@platforma-sdk/workflow-tengo:tpl")
-
+render := import("@platforma-sdk/workflow-tengo:render")
 ll := import("@platforma-sdk/workflow-tengo:ll")
 assets := import("@platforma-sdk/workflow-tengo:assets")
 pframes := import("@platforma-sdk/workflow-tengo:pframes")
@@ -19,7 +19,7 @@ times := import("times")
 mixcrAnalyzeTpl := assets.importTemplate(":mixcr-analyze")
 mixcrExportTpl := assets.importTemplate(":mixcr-export")
 aggregateByClonotypeKeyTpl := assets.importTemplate(":aggregate-by-clonotype-key")
-
+exportReportTpl := assets.importTemplate(":export-report")
 processSingleCellTpl := assets.importTemplate(":process-single-cell")
 
 self.awaitState("InputsLocked")
@@ -99,6 +99,7 @@ self.body(func(inputs) {
 		}
 	}
 	isSingleCell := len(presetSpecForBack.cellTags) > 0
+	hasUmi := !is_undefined(presetSpecForBack.umiTags) && len(presetSpecForBack.umiTags) > 0
 
 	// calculating clns annotations
 
@@ -678,6 +679,14 @@ self.body(func(inputs) {
 		}
 	}
 
+	qcReportTable := render.create(exportReportTpl, {
+		clnsData: mixcrResults.outputData("clns"),
+		isSingleCell: isSingleCell,
+		hasUmi: hasUmi,
+		sampleIdAxisSpec: sampleIdAxisSpec,
+		chains: chains
+	})
+
 	return {
 		"qc.spec": mixcrResults.outputSpec("qc"),
 		"qc.data": mixcrResults.outputData("qc"),
@@ -693,6 +702,7 @@ self.body(func(inputs) {
 
 		"resultsToCache": resultsToCache,
 
-		clonotypes: clonotypes.build()
+		clonotypes: clonotypes.build(),
+		qcReportTable: qcReportTable.output("qcReportTable")
 	}
 })