From f210c5e7f65f57a09a455d536a287a00edd53fbb Mon Sep 17 00:00:00 2001
From: renewiegandt <rene.wiegandt@mpi-bn.mpg.de>
Date: Fri, 12 Apr 2019 14:44:32 +0200
Subject: [PATCH] Improved error handling if parameter is missing

---
 pipeline.nf | 92 ++++++++++++++++++++++++++++++++---------------------
 1 file changed, 55 insertions(+), 37 deletions(-)

diff --git a/pipeline.nf b/pipeline.nf
index 44104e7..c50ba2d 100644
--- a/pipeline.nf
+++ b/pipeline.nf
@@ -68,7 +68,60 @@ disable_mo_clu = 1
 //evaluation
 	params.max_uropa_runs = 10
 
-if (params.bigwig == "" || params.bed == "" || params.organism == "" || params.genome_fasta == "" || params.motif_db == "" || params.config == "" || (params.gtf_annotation == "" && params.gtf_merged == "" ) || "${params.help}" != "0" ) {
+/*
+Checking for parameter input!
+*/
+int_params = ["window_length", "step", "min_size_fp", "max_size_fp", "kmer",
+             "aprox_motif_len", "motif_occurrence", "min_seq_length", "global",
+             "sequence_coverage", "memory", "throw_away_seq", "strand",
+             "min_seq", "motif_min_key", "motif_max_key", "iteration",
+             "edge_weight", "best_motif", "min_gap", "gap_penalty", "edge_weight",
+						 "threads", "max_uropa_runs"]
+file_params = ["bigwig", "bed", "genome_fasta", "motif_db", "config", "gtf_annotation",]
+all_params = int_params + file_params + ["organism" , "identity", "tfbsscan_method",
+							"percentage", "tomtom_treshold", "motif_similarity_thresh", "out",
+							"tissues", "gtf_merged", "cluster_motif", "tfbs_path", "help", "seed"]
+req_params = file_params + ["organism"]
+
+valid_organism = ["hg38", "hg19", "mm9", "mm10"]
+valid_tfbsscan_methods = ["moods","fimo"]
+val_missing = false
+send_help = false
+missing_params = []
+
+params.each { key, value ->
+	if (req_params.contains(key)){
+		if (key == "gtf_annotation" && value == "" && params.gtf_merged == ""){
+			val_missing = true
+		} else {
+			if (value == ""){
+				val_missing = true
+				missing_params + key
+			}
+		}
+	}
+	if (!(all_params.contains(key))){
+		println("Warning: Parameter $key is unknown. Please check for typos or the parameter list!")
+	}
+  if(int_params.contains(key)) {
+    if (!("${value}" ==~ /\d+/ )){
+      println("ERROR: $key needs to be an Integer")
+      System.exit(2)
+    }
+  }
+  if(file_params.contains(key) || (key == "gtf_merged" && value != "") ) {
+    if(!file(value).exists()) {
+      println("ERROR: $key not found. Please check the given path.")
+      System.exit(2)
+    }
+  }
+}
+
+if (val_missing){
+	send_help = true
+	println("Error: Following required parameters are missing: $missing_params")
+}
+if (send_help || "${params.help}" != "0") {
 	log.info """
 	Usage: nextflow run pipeline.nf --bigwig [BigWig-file] --bed [BED-file] --genome_fasta [FASTA-file] --motif_db [MEME-file] --config [UROPA-config-file]
 
@@ -79,7 +132,7 @@ if (params.bigwig == "" || params.bed == "" || params.organism == "" || params.g
 		--motif_db		 Path to motif-database in MEME-format
 		--config		 Path to UROPA configuration file
 		--gtf_annotation	Path to gtf annotation file
-	 	--organism 		 Input organism [hg38 | hg19 | mm9 | mm10]
+		--organism 		 Input organism [hg38 | hg19 | mm9 | mm10]
 		--out			 Output Directory (Default: './out/')
 
 	Optional arguments:
@@ -133,7 +186,6 @@ if (params.bigwig == "" || params.bed == "" || params.organism == "" || params.g
 		Evaluation:
 		--max_uropa_runs INT	 Maximum number UROPA runs running parallelized (Default: 10)
 	All arguments can be set in the configuration files
-	 ```
 	"""
 	System.exit(2)
 } else {
@@ -152,40 +204,6 @@ if (params.bigwig == "" || params.bed == "" || params.organism == "" || params.g
 
 
 
-/*
-Checking for parameter input!
-*/
-int_params = ["window_length", "step", "min_size_fp", "max_size_fp", "kmer",
-             "aprox_motif_len", "motif_occurrence", "min_seq_length", "global",
-             "sequence_coverage", "memory", "throw_away_seq", "strand",
-             "min_seq", "motif_min_key", "motif_max_key", "iteration",
-             "edge_weight", "best_motif", "min_gap", "gap_penalty", "edge_weight",
-						 "threads", "max_uropa_runs"]
-req_params = ["bigwig", "bed", "genome_fasta", "motif_db", "config"]
-all_params = int_params + req_params + ["organism" , "identity", "tfbsscan_method",
-							"percentage", "tomtom_treshold", "motif_similarity_thresh", "out",
-							"tissues", "gtf_merged", "cluster_motif", "tfbs_path", "help", "gtf_annotation", "seed"]
-
-valid_organism = ["hg38", "hg19", "mm9", "mm10"]
-valid_tfbsscan_methods = ["moods","fimo"]
-
-params.each { key, value ->
-	if (!(all_params.contains(key))){
-		println("Warning: Parameter $key is unknown. Please check for typos or the parameter list!")
-	}
-  if(int_params.contains(key)) {
-    if (!("${value}" ==~ /\d+/ )){
-      println("ERROR: $key needs to be an Integer")
-      System.exit(2)
-    }
-  }
-  if(req_params.contains(key) || (key == "gtf_merged" && value != "") ) {
-    if(!file(value).exists()) {
-      println("ERROR: $key not found. Please check the given path.")
-      System.exit(2)
-    }
-  }
-}
 if (!("${params.identity}" ==~ /^0\.[8-9][[0-9]*]?|^1(\.0)?/ )){
   println("ERROR: --identity needs to be float in range 0.8 to 1.0")
   System.exit(2)