feat: BWA mem2 index - remove prefix parameter and determine prefix f…

…rom output (#515)

* remove prefix parameter and determine prefix from output

* snakefmt

* case that .btw.2bit.64 is the first output

* typo in comment

* assertions for output files

* raise errors instead of assertions

* reformat

* Apply suggestions from code review

Co-authored-by: Johannes Köster <johannes.koester@uni-due.de>
Co-authored-by: Filipe G. Vieira <fgarrettvieira@gmail.com>

* fmt

* simplify code

* fix access to only element in prefixes set

Co-authored-by: Johannes Köster <johannes.koester@uni-due.de>
Co-authored-by: Filipe G. Vieira <fgarrettvieira@gmail.com>

bio/bwa-mem2/index/test/Snakefile

@@ -9,7 +9,5 @@ rule bwa_mem2_index:
         "{genome}.pac",
     log:
         "logs/bwa-mem2_index/{genome}.log",
-    params:
-        prefix=lambda w: w.genome,
     wrapper:
         "master/bio/bwa-mem2/index"

bio/bwa-mem2/index/wrapper.py

@@ -15,14 +15,24 @@
 elif len(snakemake.input) > 1:
     raise ValueError("Please provide exactly one reference genome as input.")
 
-# Prefix that should be used for the database
-prefix = None
-if "prefix" in snakemake.params.keys():
-    prefix = snakemake.params["prefix"]
-else:
-    prefix = splitext(snakemake.output[0])[0]
+valid_suffixes = {".0123", ".amb", ".ann", ".bwt.2bit.64", ".pac"}
 
-if len(prefix) > 0:
-    prefix = "-p " + prefix
 
-shell("bwa-mem2 index" " {prefix}" " {snakemake.input[0]}" " {log}")
+def get_valid_suffix(path):
+    for suffix in valid_suffixes:
+        if path.endswith(suffix):
+            return suffix
+
+
+prefixes = set()
+for s in snakemake.output:
+    suffix = get_valid_suffix(s)
+    if suffix is None:
+        raise ValueError(f"{s} cannot be generated by bwa-mem2 index (invalid suffix).")
+    prefixes.add(s[: -len(suffix)])
+
+if len(prefixes) != 1:
+    raise ValueError("Output files must share common prefix up to their file endings.")
+(prefix,) = prefixes
+
+shell("bwa-mem2 index -p {prefix} {snakemake.input[0]} {log}")