Merge pull request #1664 from milaboratory/fix-downsample

Fix downsample
milaboratory · May 16, 2024 · 78c52a9 · 78c52a9
2 parents 1a90bc5 + 9d98d2d
commit 78c52a9
Show file tree

Hide file tree

Showing 7 changed files with 21 additions and 19 deletions.
diff --git a/build.gradle.kts b/build.gradle.kts
@@ -134,7 +134,7 @@ val toObfuscate: Configuration by configurations.creating {
 val obfuscationLibs: Configuration by configurations.creating
 
 
-val mixcrAlgoVersion = "4.6.0-104-optional-correction-of-tags"
+val mixcrAlgoVersion = "4.6.0-106-develop"
 // may be blank (will be inherited from mixcr-algo)
 val milibVersion = ""
 // may be blank (will be inherited from mixcr-algo or milib)

diff --git a/changelogs/v4.6.1.md b/changelogs/v4.6.1.md
@@ -40,6 +40,8 @@
 - Fixed export of trees if data assembled by a feature with reference point having offset
 - Export of `VJJunction gemline` in `shmTrees` exports now export `mrca` as most plausible content
 - Fixed parsing and alignment of reads with length more than 30Kbase
+- `downsample` now supports `molecule` variant in `--downsampling` option
+- Fixed naming of output files of `downsample` command
 
 ## New Presets
 

diff --git a/regression/cli-help/downsample.txt b/regression/cli-help/downsample.txt
@@ -7,9 +7,9 @@ Downsample clonesets.
                             Paths to input files.
       --downsampling <type> Choose downsampling applied to normalize the clonesets. Possible
                               values:
-                            count-[reads|TAG]-[auto|min|fixed][-<number>]
-                             top-[reads|TAG]-[<number>]
-                             cumtop-[reads|TAG]-[percent]
+                            count-[reads|TAG|molecule]-[auto|min|fixed-<number>]
+                             top-[reads|TAG|molecule]-[<number>]
+                             cumtop-[reads|TAG|molecule]-[percent]
   -c, --chains <chains>[,<chains>...]
                             Specify chains
       --only-productive     Filter out-of-frame sequences and sequences with stop-codons.

diff --git a/regression/cli-help/overlapScatterPlot.txt b/regression/cli-help/overlapScatterPlot.txt
@@ -11,9 +11,9 @@ Plot overlap scatter-plot.
 
       --downsampling (<type>|none)
                            Choose downsampling applied to normalize the clonesets. Possible values:
-                           count-[reads|TAG]-[auto|min|fixed][-<number>]
-                            top-[reads|TAG]-[<number>]
-                            cumtop-[reads|TAG]-[percent]
+                           count-[reads|TAG|molecule]-[auto|min|fixed-<number>]
+                            top-[reads|TAG|molecule]-[<number>]
+                            cumtop-[reads|TAG|molecule]-[percent]
       --chains <chain>[,<chain>...]
                            Chains to export.
       --only-productive    Filter out-of-frame sequences and sequences with stop-codons.

diff --git a/regression/cli-help/postanalysis/individual.txt b/regression/cli-help/postanalysis/individual.txt
@@ -13,9 +13,9 @@ Run postanalysis for CDR3 metrics, diversity, V/J/VJ-usage, CDR3/V-Spectratype m
       --default-downsampling (<type>|none)
                            Default downsampling applied to normalize the clonesets. Possible
                              values:
-                           count-[reads|TAG]-[auto|min|fixed][-<number>]
-                            top-[reads|TAG]-[<number>]
-                            cumtop-[reads|TAG]-[percent]
+                           count-[reads|TAG|molecule]-[auto|min|fixed-<number>]
+                            top-[reads|TAG|molecule]-[<number>]
+                            cumtop-[reads|TAG|molecule]-[percent]
       --default-weight-function (<read>|<Tag>|none)
                            Default clonotype weight function
       --only-productive    Filter out-of-frame sequences and sequences with stop-codons.

diff --git a/regression/cli-help/postanalysis/overlap.txt b/regression/cli-help/postanalysis/overlap.txt
@@ -14,9 +14,9 @@ Overlap analysis
       --default-downsampling (<type>|none)
                            Default downsampling applied to normalize the clonesets. Possible
                              values:
-                           count-[reads|TAG]-[auto|min|fixed][-<number>]
-                            top-[reads|TAG]-[<number>]
-                            cumtop-[reads|TAG]-[percent]
+                           count-[reads|TAG|molecule]-[auto|min|fixed-<number>]
+                            top-[reads|TAG|molecule]-[<number>]
+                            cumtop-[reads|TAG|molecule]-[percent]
       --default-weight-function (<read>|<Tag>|none)
                            Default clonotype weight function
       --criteria <s>       Overlap criteria. Defines the rules to treat clones as equal. It allows

diff --git a/src/main/kotlin/com/milaboratory/mixcr/cli/postanalysis/CommandDownsample.kt b/src/main/kotlin/com/milaboratory/mixcr/cli/postanalysis/CommandDownsample.kt
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2014-2023, MiLaboratories Inc. All Rights Reserved
+ * Copyright (c) 2014-2024, MiLaboratories Inc. All Rights Reserved
  *
  * Before downloading or accessing the software, please read carefully the
  * License Agreement available at:
@@ -103,12 +103,12 @@ class CommandDownsample : MiXCRCommandWithOutputs() {
         val fileNameWithoutExtension = input.fileName.toString()
             .replace(".clna", "")
             .replace(".clns", "")
-        val suffix = when {
-            chains == null || chains!!.isEmpty() -> ""
-            chains!!.size == 1 -> "${chains!!.first()}."
-            else -> "$chains."
+        val secondSuffix = when {
+            chains == null || chains!!.isEmpty() -> null
+            chains!!.size == 1 -> "${chains!!.first()}"
+            else -> "$chains"
         }
-        val outName = "$fileNameWithoutExtension.$suffix$suffix.clns"
+        val outName = listOfNotNull(fileNameWithoutExtension, suffix, secondSuffix, "clns").joinToString(".")
         return (outPath?.resolve(outName) ?: Paths.get(outName)).toAbsolutePath()
     }