Merge pull request #163 from sunbeam-labs/dev

Bugfixes

Snakefile

@@ -60,20 +60,27 @@ Pairs = ['1', '2'] if Cfg['all']['paired_end'] else ['1']
 
 # Collect host (contaminant) genomes
 sys.stderr.write("Collecting host/contaminant genomes... ")
-HostGenomeFiles = [f for f in Cfg['qc']['host_fp'].glob('*.fasta')]
-if not HostGenomeFiles:
-    sys.stderr.write(
-        "\n\nWARNING: No files detected in host genomes folder ({}). "
-        "If this is not intentional, make sure all files end in "
-        ".fasta and the folder is specified correctly.\n\n".format(
-            Cfg['qc']['host_fp']
-        ))
+if Cfg['qc']['host_fp'] == Cfg['all']['root']:
+    HostGenomeFiles = []
+else:
+    HostGenomeFiles = [f for f in Cfg['qc']['host_fp'].glob('*.fasta')]
+    if not HostGenomeFiles:
+        sys.stderr.write(
+            "\n\nWARNING: No files detected in host genomes folder ({}). "
+            "If this is not intentional, make sure all files end in "
+            ".fasta and the folder is specified correctly.\n\n".format(
+                Cfg['qc']['host_fp']
+            ))
 HostGenomes = {Path(g.name).stem: read_seq_ids(Cfg['qc']['host_fp'] / g) for g in HostGenomeFiles}
 sys.stderr.write("done.\n")
 
 sys.stderr.write("Collecting target genomes... ")
-GenomeFiles = [f for f in Cfg['mapping']['genomes_fp'].glob('*.fasta')]
-GenomeSegments = {PurePath(g.name).stem: read_seq_ids(Cfg['mapping']['genomes_fp'] / g) for g in GenomeFiles}
+if Cfg['mapping']['genomes_fp'] == Cfg['all']['root']:
+    GenomeFiles = []
+    GenomeSegments = {}
+else:
+    GenomeFiles = [f for f in Cfg['mapping']['genomes_fp'].glob('*.fasta')]
+    GenomeSegments = {PurePath(g.name).stem: read_seq_ids(Cfg['mapping']['genomes_fp'] / g) for g in GenomeFiles}
 sys.stderr.write("done.\n")
 
 # ---- Change your workdir to output_fp

docs/quickstart.rst

@@ -29,7 +29,7 @@ runs some tests to make sure everything was installed correctly.
    If you've never installed Conda before, you'll need to add it to your shell's
    path. If you're running Bash (the most common terminal shell), the following
    command will add it to your path: ``echo 'export
-   PATH=$PATH:$HOME/miniconda3/bin` > ~/.bashrc``
+   PATH=$PATH:$HOME/miniconda3/bin' > ~/.bashrc``
 
 If you see "Tests failed", check out our :ref:`troubleshooting` section or file an issue
 on our `GitHub <https://github.com/sunbeam-labs/sunbeam/issues>`_ page.

environment.yml

@@ -6,15 +6,15 @@ dependencies:
   - snakemake=4.8.1
   - ruamel.yaml
   - biopython
-  - trimmomatic
+  - trimmomatic=0.36
   - fastqc
   - blast
   - kraken=1.0
   - kraken-biom
   - vsearch
   - jellyfish=1.1.11=1
   - cutadapt
-  - bwa
+  - bwa>=0.7.17
   - pysam
   - pandas
   - megahit

sunbeamlib/config.py

@@ -65,6 +65,7 @@ def check_config(cfg):
     new_cfg = dict()
     for section, values in cfg.items():
         new_cfg[section] = validate_paths(values, root)
+    new_cfg['all']['root'] = root
     return new_cfg
 
 

sunbeamlib/scripts/init.py

@@ -57,16 +57,26 @@ def main(argv=sys.argv):
     args = parser.parse_args(argv)
 
     # Create project folder if it doesn't exist
-    project_fp = args.project_fp.resolve()
+    project_fp_exists = False
+    project_fp = args.project_fp
+
+    try:
+        project_fp = args.project_fp.resolve()
+        project_fp_exists = project_fp.exists()
+    except FileNotFoundError:
+        pass
+
+    if not project_fp_exists:
+        sys.stderr.write(
+            "Creating project folder at {}...\n".format(args.project_fp))
+        project_fp.mkdir(parents=True, exist_ok=True)
+
 
     # Check if files already exist
     config_file = check_existing(project_fp/args.output, args.force)
     samplelist_file = check_existing(project_fp/"samples.csv", args.force)
 
-    if not project_fp.exists():
-        sys.stderr.write(
-            "Creating project folder at {}...\n".format(args.project_fp))
-        project_fp.mkdir(parents=True, exist_ok=True)
+
 
     # Create config file        
     cfg = config.new(

tests/test_suite.bash

@@ -109,3 +109,26 @@ function test_guess_with_index_files_present {
     grep '{sample}_R{rp}.fastq.gz' out.txt
 #    rm -r $TEMPDIR/idx_files_present
 }
+
+# Fix for #153
+# Empty strings for _fp entries in the config should result in the relevant
+# files being ignored, even if they would normally match the pattern used.
+function test_blank_fp_behavior {
+    # Move indexes to top-level
+    for file in $TEMPDIR/indexes/*; do
+        mv $file $TEMPDIR/indexes_${file##*/}
+    done
+    # Run sunbeam with a blank string for the mapping step's genomes path, and
+    # with fasta files lying around in the top-level directory.
+    sunbeam config modify --str 'mapping: {genomes_fp: ""}' \
+        $TEMPDIR/tmp_config.yml > $TEMPDIR/blank_fp_config.yml
+    sunbeam run --configfile $TEMPDIR/blank_fp_config.yml
+    # Move indexes back to original location
+    for file in $TEMPDIR/indexes_*; do
+        mv $file ${file/indexes_/indexes\//}
+    done
+    # These files should *not* exist, because we don't accept the top-level
+    # directory as a genomes_fp directory.
+    [ ! -f $TEMPDIR/sunbeam_output/mapping/indexes_human/coverage.csv ]
+    [ ! -f $TEMPDIR/sunbeam_output/mapping/indexes_phix174/coverage.csv ]
+}