add tests for header issue

gbouras13 · Mar 12, 2024 · db9e7e9 · db9e7e9
1 parent c31939f
commit db9e7e9
Show file tree

Hide file tree

Showing 6 changed files with 2,033 additions and 9 deletions.
diff --git a/bin/input_commands.py b/bin/input_commands.py
@@ -61,8 +61,8 @@ def get_input():
         "-g",
         "--gene_predictor",
         action="store",
-        help='User specified gene predictor. Use "-g phanotate" or "-g prodigal" or "-g prodigal-gv" or "-g genbank". \nDefaults to phanotate (not required unless prodigal is desired).',
-        default="phanotate",
+        help='User specified gene predictor. Use "-g phanotate" or "-g prodigal" or "-g prodigal-gv" or "-g genbank". \nDefaults to phanotate usually and prodigal-gv in meta mode.',
+        default="default",
     )
     parser.add_argument(
         "-m",

diff --git a/bin/pharokka.py b/bin/pharokka.py
@@ -65,6 +65,12 @@ def main():
 
     # set the gene_predictor
     gene_predictor = args.gene_predictor
+    # set to phanotate by default and prodigal-gv in meta mode
+    if gene_predictor == "default":
+        if args.meta is True:
+            gene_predictor = "prodigal-gv"
+        else:
+            gene_predictor = "phanotate"
 
     # instantiate outdir
     out_dir = instantiate_dirs(args.outdir, args.meta, args.force)

diff --git a/bin/pharokka_proteins.py b/bin/pharokka_proteins.py
@@ -82,7 +82,7 @@ def main():
     logger.info("Checking dependencies.")
     check_dependencies(False)  # to check pharokka_proteins.py, don't need mash
 
-    # instantiation/checking fasta and gene_predictor
+    # instantiation/checking fasta
     validate_fasta(args.infile)
     validate_threads(args.threads)
 

diff --git a/bin/processes.py b/bin/processes.py
@@ -379,7 +379,7 @@ def tidy_phanotate_output(out_dir):
         names=col_list,
         skiprows=2,
         dtype=dtype_dict,
-        comment='#'  # to skip the headers
+        comment="#",  # to skip the headers
     )
     # get rid of the headers and reset the index
     phan_df = phan_df[phan_df["start"] != "#id:"]
@@ -440,7 +440,7 @@ def tidy_prodigal_output(out_dir, gv_flag):
         index_col=False,
         names=col_list,
         dtype=dtype_dict,
-        comment='#'  # to skip the headers
+        comment="#",  # to skip the headers
     )
 
     # meta mode brings in some Nas so remove them