Reorder Snakefile to make sure the tests work

There were some failed tests due to changes in the order of checks in
the Snakefile, which caused the wrong error to be thrown when there were
multiple missing files. This commit fixes that issue.

The KNOWN_SITES argument has been made mandatory for now, and this
requirement has also been added to the test cases.

Snakefile

@@ -41,6 +41,22 @@ if DBSNP is None:
 if not Path(DBSNP).exists():
     raise FileNotFoundError("{0} does not exist".format(DBSNP))
 
+SCONFIG = config.get("SAMPLE_CONFIG")
+if SCONFIG is None:
+    raise ValueError("You must set --config SAMPLE_CONFIG=<path>")
+if not Path(SCONFIG).exists():
+    raise FileNotFoundError("{0} does not exist".format(SCONFIG))
+
+KNOWN_SITES = config.get("KNOWN_SITES")
+if KNOWN_SITES is None:
+    raise ValueError("You must set --config KNOWN_SITES=<path>")
+
+# The user can specify multiple known sites
+KNOWN_SITES = KNOWN_SITES.split(',')
+for filename in KNOWN_SITES:
+    if not Path(filename).exists():
+        raise FileNotFoundError("{0} does not exist".format(filename))
+
 
 # these are all optional
 BED = config.get("BED", "")  # comma-separated list of BED files
@@ -49,19 +65,10 @@ FEMALE_THRESHOLD = config.get("FEMALE_THRESHOLD", 0.6)
 FASTQ_COUNT = config.get("FASTQ_COUNT")
 MAX_BASES = config.get("MAX_BASES", "")
 
-# Make sure all files with known sites exist
-known_sites = config.get("KNOWN_SITES","")
-
-# Did we get multiple KNOWN SITES?
-known_sites = known_sites.split(',')
-for filename in known_sites:
-    if not Path(filename).exists():
-        raise FileNotFoundError("{0} does not exist".format(filename))
-
 # Generate the input string for basrecalibration
-known_sites_argument = ''
-for argument, filename in zip(itertools.repeat('-knownSites'), known_sites):
-    known_sites_argument +=' {} {}'.format(argument, filename)
+BSQR_known_sites = ''
+for argument, filename in zip(itertools.repeat('-knownSites'), KNOWN_SITES):
+    BSQR_known_sites +=' {} {}'.format(argument, filename)
 
 def fsrc_dir(*args):
     """Wrapper around snakemake's srcdir to work like os.path.join"""
@@ -78,12 +85,6 @@ tsvpy = fsrc_dir("src", "stats_to_tsv.py")
 fqsc = fsrc_dir("src", "safe_fastqc.sh")
 
 # sample config parsing
-SCONFIG = config.get("SAMPLE_CONFIG")
-if SCONFIG is None:
-    raise ValueError("You must set --config SAMPLE_CONFIG=<path>")
-if not Path(SCONFIG).exists():
-    raise FileNotFoundError("{0} does not exist".format(SCONFIG))
-
 with open(config.get("SAMPLE_CONFIG")) as handle:
     SAMPLE_CONFIG = json.load(handle)
 SAMPLES = SAMPLE_CONFIG['samples'].keys()
@@ -323,12 +324,12 @@ rule baserecal:
     output:
         grp = "{sample}/bams/{sample}.baserecal.grp"
     params:
-        knownsites = known_sites_argument
+        known_sites = BSQR_known_sites
     singularity: "docker://broadinstitute/gatk3:3.7-0"
     shell: "java -XX:ParallelGCThreads=1 -jar /usr/GenomeAnalysisTK.jar -T "
            "BaseRecalibrator -I {input.bam} -o {output.grp} -nct 8 "
            "-R {input.ref} -cov ReadGroupCovariate -cov QualityScoreCovariate "
-           "-cov CycleCovariate -cov ContextCovariate {params.knownsites}"
+           "-cov CycleCovariate -cov ContextCovariate {params.known_sites}"
 
 rule gvcf_scatter:
     """Run HaplotypeCaller in GVCF mode by chunk"""

tests/test_dry_run.yml

@@ -4,6 +4,7 @@
     JAVA=README.md REFERENCE=tests/data/ref.fa GATK=README.md
     DBSNP=tests/data/database.vcf.gz ONETHOUSAND=tests/data/database.vcf.gz
     HAPMAP=tests/data/database.vcf.gz
+    KNOWN_SITES=tests/data/database.vcf.gz
     SAMPLE_CONFIG=tests/data/sample_config.json
   exit_code: 0
   stdout: