filter for outliers

.changeset/smooth-rockets-confess.md

@@ -0,0 +1,6 @@
+---
+'@platforma-open/milaboratories.repertoire-diversity.workflow': minor
+'@platforma-open/milaboratories.repertoire-diversity.model': minor
+---
+
+Outliers removed from table and plots if option "Drop outliers" is on

model/src/index.ts

@@ -124,7 +124,8 @@ export const model = BlockModel.create()
       .getData()
       .entries.map((v) => v.obj)
       .filter(isPColumn)
-      .filter((column) => valueTypes.find((valueType) => valueType === column.spec.valueType));
+      .filter((column) => valueTypes.find((valueType) => valueType === column.spec.valueType))
+      .filter((column) => column.spec.name !== "Sample");
 
     return ctx.createPFrame([...pCols, ...upstream]);
   })

pnpm-workspace.yaml

@@ -15,6 +15,7 @@ catalog:
   '@milaboratories/software-pframes-conv': ^2.1.8
   '@platforma-open/milaboratories.software-small-binaries': ^1.15.7
   '@platforma-open/milaboratories.software-mixcr': 4.7.0-137-develop
+  '@platforma-open/milaboratories.software-ptransform': ^1.2.0
 
   '@platforma-sdk/test': ^1.22.22
 

workflow/package.json

@@ -14,6 +14,7 @@
     "@milaboratories/software-pframes-conv": "catalog:",
     "@platforma-open/milaboratories.software-small-binaries": "catalog:",
     "@platforma-open/milaboratories.software-mixcr": "catalog:",
+    "@platforma-open/milaboratories.software-ptransform": "catalog:",
     "vitest": "catalog:"
   }
 }

workflow/src/filterTables.lib.tengo

@@ -0,0 +1,36 @@
+ll := import("@platforma-sdk/workflow-tengo:ll")
+exec := import("@platforma-sdk/workflow-tengo:exec")
+assets := import("@platforma-sdk/workflow-tengo:assets")
+json := import("json")
+paggregateSw := assets.importSoftware("@platforma-open/milaboratories.software-ptransform:main")
+
+//// aggregating by-nodes output to make it uniquely addressable by it's native key
+filterOutliers := func(inputTsv) {
+    filterWorkflow := { steps: [
+        {
+            type: "filter",
+            predicate: {
+                type: "gt",
+                value: 0,
+                column: "Observed diversity"
+                }
+            }
+        ]
+    }
+
+
+    filterCmd := exec.builder().
+        printErrStreamToStdout().
+        software(paggregateSw).
+        arg("--workflow").arg("wf.json").
+        writeFile("wf.json", json.encode(filterWorkflow)).
+        arg("input.tsv").addFile("input.tsv", inputTsv).
+        arg("output.tsv").saveFile("output.tsv").
+        run()
+
+    return filterCmd.getFile("output.tsv")
+}
+
+export ll.toStrict({
+	filterOutliers: filterOutliers
+})

workflow/src/importTsv.tpl.tengo

@@ -10,6 +10,8 @@ json := import("json")
 
 self.defineOutputs("pf")
 
+filterTables := import(":filterTables")
+
 tsvImportParams := {
 	"axes": [
 		{
@@ -58,8 +60,9 @@ self.body(func(args) {
 
 	tsvMap := pColumn.resourceMapBuilder(/* key length*/ 1 )
 	for fileName, file in args.tsv {
+		filteredFile := filterTables.filterOutliers(file)
 		chain := fileName[FILE_NAME_PREFIX_LEN:(len(fileName) - FILE_NAME_SUFFIX_LEN)]
-		tsvMap.add([chain], file)
+		tsvMap.add([chain], filteredFile)
 	}
 	tsvMap = tsvMap.build()