Fix gtdb dirrectory input

CHANGELOG.md

@@ -11,6 +11,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### `Fixed`
 
+- [#582](https://github.com/nf-core/mag/pull/575) - Fix GTDB database input when directory supplied (fix by @jfy133)
+
 ### `Dependencies`
 
 ### `Deprecated`

conf/test.config

@@ -29,5 +29,7 @@ params {
     busco_db                      = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     busco_clean                   = true
     skip_gtdbtk                   = true
+    gtdbtk_min_completeness       = 0
     skip_concoct                  = true
+
 }

conf/test_adapterremoval.config

@@ -29,6 +29,7 @@ params {
     max_unbinned_contigs        = 2
     busco_db                    = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness     = 0
     clip_tool                   = 'adapterremoval'
     skip_concoct                = true
     bin_domain_classification   = true

conf/test_ancient_dna.config

@@ -28,6 +28,7 @@ params {
     max_unbinned_contigs                 = 2
     busco_db                             = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     skip_gtdbtk                          = true
+    gtdbtk_min_completeness              = 0
     ancient_dna                          = true
     binning_map_mode                     = 'own'
     skip_spades                          = false

conf/test_bbnorm.config

@@ -35,6 +35,7 @@ params {
     busco_db                      = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     busco_clean                   = true
     skip_gtdbtk                   = true
+    gtdbtk_min_completeness       = 0
     bbnorm                        = true
     coassemble_group              = true
 }

conf/test_binrefinement.config

@@ -29,6 +29,7 @@ params {
     max_unbinned_contigs          = 2
     busco_db                      = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     skip_gtdbtk                   = true
+    gtdbtk_min_completeness       = 0
     refine_bins_dastool           = true
     refine_bins_dastool_threshold = 0
     // TODO not using 'both' until #489 merged

conf/test_busco_auto.config

@@ -25,6 +25,7 @@ params {
     min_length_unbinned_contigs = 1
     max_unbinned_contigs        = 2
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness     = 0
     skip_prokka                 = true
     skip_prodigal               = true
     skip_quast                  = true

conf/test_host_rm.config

@@ -26,5 +26,6 @@ params {
     max_unbinned_contigs        = 2
     busco_db                    = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness     = 0
     skip_concoct                = true
 }

conf/test_hybrid.config

@@ -25,5 +25,6 @@ params {
     max_unbinned_contigs        = 2
     busco_db                    = "https://busco-data.ezlab.org/v5/data/lineages/bacteria_odb10.2024-01-08.tar.gz"
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness     = 0
     skip_concoct                = true
 }

conf/test_hybrid_host_rm.config

@@ -27,4 +27,5 @@ params {
     skip_binqc                  = true
     skip_concoct                = true
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness     = 0
 }

conf/test_nothing.config

@@ -39,5 +39,6 @@ params {
     skip_prokka                   = true
     skip_binqc                    = true
     skip_gtdbtk                   = true
+    gtdbtk_min_completeness       = 0
     skip_concoct                  = true
 }

conf/test_virus_identification.config

@@ -28,6 +28,7 @@ params {
     reads_minlength             = 150
     coassemble_group            = true
     skip_gtdbtk                 = true
+    gtdbtk_min_completeness       = 0
     skip_binning                = true
     skip_prokka                 = true
     skip_spades                 = true

subworkflows/local/gtdbtk.nf

@@ -66,13 +66,16 @@ workflow GTDBTK {
         // Expects to be tar.gz!
         ch_db_for_gtdbtk = GTDBTK_DB_PREPARATION ( gtdb ).db
     } else if ( gtdb.isDirectory() ) {
-        // Make up meta id to match expected channel cardinality for GTDBTK
+        // The classifywf module expects a list of the _contents_ of the GTDB
+        // database, not just the directory itself (I'm not sure why). But
+        // for now we generate this list before putting into a channel,
+        // then grouping again to pass to the module.
+        // Then make up meta id to match expected channel cardinality for GTDBTK
+        gtdb_dir = gtdb.listFiles()
         ch_db_for_gtdbtk = Channel
-                            .of(gtdb)
-                            .map{
-                                [ it.toString().split('/').last(), it ]
-                            }
-                            .collect()
+                            .of(gtdb_dir)
+                            .map{['gtdb', it]}
+                            .groupTuple()
     } else {
         error("Unsupported object given to --gtdb, database must be supplied as either a directory or a .tar.gz file!")
     }