Merge pull request #1547 from milaboratory/allels-help

Allels help
milaboratory · Feb 22, 2024 · 3a94f2d · 3a94f2d
2 parents 4f60620 + 20aa8f6
commit 3a94f2d
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 25 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 = "4.6.0-50-rework_analyze"
+val mixcrAlgoVersion = "4.6.0-52-alleles-help"
 // 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/regression/cli-help/findAlleles.txt b/regression/cli-help/findAlleles.txt
@@ -1,6 +1,6 @@
 Usage: mixcr findAlleles [--export-library <path.(json|fasta)>]... [--export-alleles-mutations
-                         <path>] [--dont-remove-unused-genes] [-O <key=value>]... [--report <path>]
-                         [--json-report <path>] [--use-local-temp] [--threads <n>]
+                         <path.(t|c)sv>] [--dont-remove-unused-genes] [-O <key=value>]... [--report
+                         <path>] [--json-report <path>] [--use-local-temp] [--threads <n>]
                          [--force-overwrite] [--no-warnings] [--verbose] [--help]
                          (--output-template <template.clns> | --no-clns-output) (input_file.
                          (clns|clna)|directory)...
@@ -28,33 +28,33 @@ of V and J genes and the same features to align.
                              For `.json` library will be written in reqpseqio format.
                              For `.fasta` library will be written in FASTA format with gene name
                                and feature name in description.
-      --export-alleles-mutations <path>
+      --export-alleles-mutations <path.(t|c)sv>
                              Path where to write descriptions and stats for all result alleles,
                                existed and new.
                              Available columns:
                              alleleName - result allele name. If it's de nova, then before `-M` is
                                nearest known variant, M{n} is how many mutations in it and CDR3-{n}
                                is how many mutations in CDR3, at the end uniq hash.
-                             geneName - gene name, the same for zigotes
+                             geneName - gene name, the same for zygotes
                              type - gene type (Variable or Joining)
-                             status - DE_NOVO: De novo alleles
-                             FOUND_KNOWN_VARIANT: Found allele identical to one of genes in the
-                               library, but it's not a main variant. If enoughInfo is true, then it
-                               was found by main algorithm. In some cases gene name could be
-                               changed. If enoughInfo is false, then it was found by aligning
-                               clones on known variants.
-                             ALIGNED_ON_KNOWN_VARIANT: Not enough info for search, but aligned on
-                               known allele
-                             NOT_CHANGED_AFTER_SEARCH: Not changed after search
-                             NO_CLONES_TO_SEARCH: No clones to search
+                             status - DE_NOVO: New allelic variant, that doesn't match with any
+                               from the library.
+                             FOUND_KNOWN_VARIANT: Known variant from the library, found by main
+                               algorithm.
+                             ALIGNED_ON_KNOWN_VARIANT: Not enough info for main algorithm, but the
+                               main bulk of clones aligns with a known allele from the database.
+                             NOT_CHANGED_AFTER_SEARCH: The found allele variant is the same as was
+                               used in `align` command (clones are not changed).
+                                If enoughInfo is true, then it was found by main algorithm,
+                               otherwise, by aligning clones on known variants.
+                             NO_CLONES_TO_SEARCH: All clones were filtered out before search (there
+                               are filters by productivity and reads/UMI count).
                              COULD_NOT_BE_ALIGNED_ON_KNOWN_VARIANT: Have clones to search, but not
-                               enough info for search and can't be aligned on known variant
-                             REMOVED_BECAUSE_NO_TOP_HITS_IN_RESULT_FILES: No top hits in result
-                               files
-                             REMOVED_BECAUSE_NO_TOP_HITS_IN_SOURCE_FILES: No top hits in source
-                               files
+                               enough info for search and can't be aligned on known variant.
+                             REMOVED_BECAUSE_NO_TOP_HITS_IN_RESULT_FILES: Removed because no top
+                               hits in any result clone.
                              REMOVED_BECAUSE_NOT_REPRESENTED_IN_SOURCE_FILES: Not represented in
-                               source files
+                               source clones.
                              .
                              enoughInfo - is there enough clones to search by the main algorithm.
                              mutations - list of mutations for de nova alleles.
@@ -76,7 +76,7 @@ of V and J genes and the same features to align.
                              filteredForAlleleSearchClonesCountWithNegativeScoreChange - how many
                                clones get lower score after aligned to this allele and was used in
                                searching before
-                             scoreDelta - stats for score change of clones afeter aligning to this
+                             scoreDelta - stats for score change of clones after aligning to this
                                allele.
       --dont-remove-unused-genes
                              Don't remove genes that weren't used in clones in the result library

diff --git a/src/main/kotlin/com/milaboratory/mixcr/cli/CommandFindAlleles.kt b/src/main/kotlin/com/milaboratory/mixcr/cli/CommandFindAlleles.kt
@@ -150,7 +150,7 @@ class CommandFindAlleles : MiXCRCommandWithOutputs() {
             "Path where to write descriptions and stats for all result alleles, existed and new.",
             "Available columns:",
             "alleleName - result allele name. If it's de nova, then before `-M` is nearest known variant, M{n} is how many mutations in it and CDR3-{n} is how many mutations in CDR3, at the end uniq hash.",
-            "geneName - gene name, the same for zigotes",
+            "geneName - gene name, the same for zygotes",
             "type - gene type (Variable or Joining)",
             "status - ${AlleleSearchResult.Status.helpDescription}.",
             "enoughInfo - is there enough clones to search by the main algorithm.",
@@ -165,10 +165,10 @@ class CommandFindAlleles : MiXCRCommandWithOutputs() {
             "filteredForAlleleSearchNaivesCount - how many naive clones aligned to this allele and was used in searching.",
             "filteredForAlleleSearchClonesCount - how many clones aligned to this allele and was used in searching.",
             "filteredForAlleleSearchClonesCountWithNegativeScoreChange - how many clones get lower score after aligned to this allele and was used in searching before",
-            "scoreDelta - stats for score change of clones afeter aligning to this allele."
+            "scoreDelta - stats for score change of clones after aligning to this allele."
         ],
         names = ["--export-alleles-mutations"],
-        paramLabel = "<path>",
+        paramLabel = "<path.(t|c)sv>",
         order = OptionsOrder.main + 10_200
     )
     var allelesMutationsOutput: Path? = null

diff --git a/src/test/kotlin/com/milaboratory/mixcr/cli/CommandFindAllelesTest.kt b/src/test/kotlin/com/milaboratory/mixcr/cli/CommandFindAllelesTest.kt
@@ -12,9 +12,13 @@
 package com.milaboratory.mixcr.cli
 
 import com.milaboratory.app.ValidationException
+import com.milaboratory.mixcr.alleles.AlleleSearchResult
 import com.milaboratory.util.TempFileDest
 import com.milaboratory.util.TempFileManager
+import io.kotest.assertions.asClue
 import io.kotest.matchers.collections.shouldContainExactlyInAnyOrder
+import io.kotest.matchers.string.shouldContain
+import io.kotest.matchers.string.shouldNotContain
 import io.kotest.matchers.string.shouldStartWith
 import org.junit.Assert
 import org.junit.Before
@@ -39,6 +43,20 @@ class CommandFindAllelesTest {
         file3.toFile().createNewFile()
     }
 
+    @Test
+    fun `check that help contains all not hidden statuses`() {
+        AlleleSearchResult.Status.values().filter { !it.hideInReport }.forEach { status ->
+            status.asClue {
+                AlleleSearchResult.Status.helpDescription shouldContain status.name
+            }
+        }
+        AlleleSearchResult.Status.values().filter { it.hideInReport }.forEach { status ->
+            status.asClue {
+                AlleleSearchResult.Status.helpDescription shouldNotContain status.name
+            }
+        }
+    }
+
     @Test
     fun outputsWrittenToCommonDirectory() {
         val p = Main.parseArgs(