Merge pull request #343 from sunbeam-labs/332-kraken-biom-rule-fails-…

…when-run-on-a-cluster

Kraken-biom rule fails on cluster

envs/classify.yml

@@ -1,5 +1,12 @@
 name: classify
 channels:
   - bioconda
+  - anaconda
+  - conda-forge
 dependencies:
-  - biom-format
+  - biom-format
+  - kraken-biom
+  - libgfortran-ng=7.2.0 # kraken-biom https://github.com/gammapy/gammapy/issues/1417
+#  - conda-forge::libgfortran # kraken-biom
+#  - anaconda::pathlib # kraken-biom
+  - python>=3 # biom-format

rules/classify/kraken.smk

@@ -30,13 +30,12 @@ rule kraken2_biom:
     input:
         expand(CLASSIFY_FP/'kraken'/'{sample}-taxa.tsv',
                sample=Samples.keys())
+    conda:
+        "../../envs/classify.yml"
     output:
         CLASSIFY_FP/'kraken'/'all_samples.biom'
     shell:
-        # Using pip to install because conda version is way outdated
         """
-        pip install kraken-biom && \
-        pip install --upgrade numpy && \
         kraken-biom --max D -o {output} {input}
         """
 

rules/reports/reports.smk

@@ -40,17 +40,3 @@ rule fastqc_report:
     script:
         "../../scripts/reports/fastqc_report.py"
 
-
-rule multiqc_report:
-    """Build multiqc report on qc step."""
-    input:
-        TARGET_FASTQC
-    output:
-        MULTIQC_REPORT
-    params:
-        title = Cfg['qc'].get('report_title', 'QC report'),
-        outdir = QC_FP/'reports'
-    conda:
-        "../../envs/reports.yml"
-    script:
-        "../../scripts/reports/multiqc_report.py"

rules/targets/targets.smk

@@ -56,14 +56,10 @@ TARGET_ANNOTATE = expand(
 
 
 # ---- Reports
-# MultiQC report
-MULTIQC_REPORT = QC_FP/'reports'/'multiqc_report.html'
-
 TARGET_REPORT = [
-    QC_FP/'reports'/'preprocess_summary.tsv',
-    QC_FP/'reports'/'fastqc_quality.tsv',
-    ASSEMBLY_FP/'contigs_coverage.txt',
-    MULTIQC_REPORT
+    str(QC_FP/'reports'/'preprocess_summary.tsv'),
+    str(QC_FP/'reports'/'fastqc_quality.tsv'),
+    str(ASSEMBLY_FP/'contigs_coverage.txt'),
 ]
 
 # ---- All targets

scripts/reports/multiqc_report.py

@@ -3,9 +3,11 @@
 
 report_name = snakemake.output[0].split('/')[-1]  # Get unique name from targets.rules file
 
+os.system("conda list")
 os.system("multiqc -f -i \"{a}\" -n {b} -o {c} {c}".format(
             a=snakemake.params.title, b=report_name, c=snakemake.params.outdir))
 # -f overwrites any previous reports instead of iterating the next one's name
+print(os.listdir(snakemake.params.outdir))
 
 # Different versions of multiqc seem to be inconsistent with naming conventions
 if report_name in os.listdir(snakemake.params.outdir):

tests/targets.txt

@@ -31,7 +31,5 @@ qc/cleaned/dummyecoli_2.fastq.gz
 qc/cleaned/random_1.fastq.gz
 qc/cleaned/random_2.fastq.gz
 
-qc/reports/multiqc_report.html
-
 mapping/human/coverage.csv
 mapping/phix174/coverage.csv