Refactor samplesheet validation and input handling in ChIP-seq pipeline

Author: cursoragent

assets/schema_input.json

@@ -9,14 +9,13 @@
         "properties": {
             "sample": {
                 "type": "string",
-                "pattern": "^\\S+$",
-                "errorMessage": "Sample name must be provided and cannot contain spaces",
+                "pattern": "^[a-zA-Z0-9_.-]+$",
+                "errorMessage": "Sample name must be provided and can only contain alphanumeric characters, underscores, dots and dashes",
                 "meta": ["id"]
             },
             "fastq_1": {
                 "type": "string",
                 "format": "file-path",
-                "exists": true,
                 "pattern": "^\\S+\\.f(ast)?q\\.gz$",
                 "errorMessage": "FastQ file for reads 1 must be provided, cannot contain spaces and must have extension '.fq.gz' or '.fastq.gz'"
             },
@@ -25,7 +24,8 @@
                 "anyOf": [
                     {
                         "type": "string",
-                        "pattern": "^\\S+\\.f(ast)?q\\.gz$"
+                        "pattern": "^\\S+\\.f(ast)?q\\.gz$",
+                        "format": "file-path"
                     },
                     {
                         "type": "string",
@@ -35,27 +35,59 @@
             },
             "replicate": {
                 "type": "integer",
-                "errorMessage": "Replicate id not an integer!",
+                "minimum": 1,
+                "errorMessage": "Replicate id must be provided and must be a positive integer",
                 "meta": ["replicate"]
             },
             "antibody": {
-                "type": "string",
-                "pattern": "^\\S+$",
-                "errorMessage": "Antibody entry cannot contain spaces",
+                "anyOf": [
+                    {
+                        "type": "string",
+                        "pattern": "^[a-zA-Z0-9_.-]+$",
+                        "minLength": 1
+                    },
+                    {
+                        "type": "string",
+                        "maxLength": 0
+                    }
+                ],
+                "errorMessage": "Antibody entry can only contain alphanumeric characters, underscores, dots and dashes",
                 "meta": ["antibody"]
             },
             "control": {
-                "type": "string",
-                "pattern": "^\\S+$",
-                "errorMessage": "Control entry cannot contain spaces",
+                "anyOf": [
+                    {
+                        "type": "string",
+                        "pattern": "^[a-zA-Z0-9_.-]+$",
+                        "minLength": 1
+                    },
+                    {
+                        "type": "string",
+                        "maxLength": 0
+                    }
+                ],
+                "errorMessage": "Control sample name can only contain alphanumeric characters, underscores, dots and dashes",
                 "meta": ["control"]
             },
             "control_replicate": {
-                "type": "integer",
-                "errorMessage": "Replicate id not an integer!",
+                "anyOf": [
+                    {
+                        "type": "integer",
+                        "minimum": 1
+                    },
+                    {
+                        "type": "string",
+                        "maxLength": 0
+                    }
+                ],
+                "errorMessage": "Control replicate id must be a positive integer when specified",
                 "meta": ["control_replicate"]
             }
         },
-        "required": ["sample", "fastq_1"]
+        "required": ["sample", "fastq_1", "replicate"],
+        "dependentRequired": {
+            "antibody": ["control"],
+            "control": ["antibody", "control_replicate"]
+        }
     }
 }

bin/check_samplesheet.py

@@ -71,10 +71,8 @@ workflow NFCORE_CHIPSEQ {
     //
     // WORKFLOW: Run nf-core/chipseq workflow
     //
-    ch_samplesheet = Channel.value(file(params.input, checkIfExists: true))
-
     CHIPSEQ(
-        ch_samplesheet,
+        PIPELINE_INITIALISATION.out.samplesheet,
         ch_versions,
         PREPARE_GENOME.out.fasta,
         PREPARE_GENOME.out.fai,