From 2886d21cfd3fc6509cc169191bc6d406b18f708f Mon Sep 17 00:00:00 2001
From: Christoph Stritt <christoph.stritt@unibas.ch>
Date: Tue, 16 Jan 2024 10:51:49 +0000
Subject: [PATCH] Revert "Merge branch 'scicore' of
https://git.scicore.unibas.ch/TBRU/PacbioSnake into scicore"
This reverts commit 0d6ddf818d1c47682c9aba56b0d25c55f5ff1dc2
---
.gitignore | 2 -
assembly/README.md | 8 +--
.../cluster_config.yaml} | 10 +--
assembly/config/config.yaml | 14 ++--
assembly/config/environment.yml | 1 -
assembly/run_assembly_pipeline.py | 65 +++++++------------
6 files changed, 36 insertions(+), 64 deletions(-)
rename assembly/{cluster/config.yaml => config/cluster_config.yaml} (86%)
diff --git a/.gitignore b/.gitignore
index 155279f..479190f 100755
--- a/.gitignore
+++ b/.gitignore
@@ -7,5 +7,3 @@ assembly/resources/bakta_db
facienda.md
variantcalling/container/pggb_latest.sif
variantcalling/.snakemake
-assembly/logs
-assembly/.fontconfig
diff --git a/assembly/README.md b/assembly/README.md
index 177b319..8c339e5 100755
--- a/assembly/README.md
+++ b/assembly/README.md
@@ -13,18 +13,14 @@ The user needs to provide two things to run the workflow on her samples:
- a config file with some global options for the analysis
- a tab separate table, without header, that contains the sample names and the corresponding paths to the HiFi consensus reads.
-
-## Clone the directory
-
-
## Create conda environment containing snakemake and singularity
```
-conda env create -f config/environment.yml
+conda env create -f environment.yml
```
-## Run the pipeline
+## config.yml
In the file config/config.yaml some global parameters can be set:
```yaml
diff --git a/assembly/cluster/config.yaml b/assembly/config/cluster_config.yaml
similarity index 86%
rename from assembly/cluster/config.yaml
rename to assembly/config/cluster_config.yaml
index 9deb95e..a8eb216 100644
--- a/assembly/cluster/config.yaml
+++ b/assembly/config/cluster_config.yaml
@@ -9,16 +9,16 @@ cluster:
--output=logs/{rule}/{rule}-{wildcards}-%j.stdout
--error=logs/{rule}/{rule}-{wildcards}-%j.stderr
default-resources:
- - "partition=scicore"
- - "qos='1day'"
- - "time='12:00:00'"
- - "mem_mb=20000"
+ - partition=scicore
+ - qos=1day
+ - time=12:00:00
+ - mem_mb=20000
restart-times: 3
max-jobs-per-second: 10
max-status-checks-per-second: 1
local-cores: 1
latency-wait: 60
-jobs: 10
+jobs: 500
keep-going: True
rerun-incomplete: True
printshellcmds: True
diff --git a/assembly/config/config.yaml b/assembly/config/config.yaml
index ee2e9b6..eb11d3b 100755
--- a/assembly/config/config.yaml
+++ b/assembly/config/config.yaml
@@ -2,17 +2,15 @@
#
##############################
-samples: "config/samples.tsv"
-outdir: "./results"
+samples: config/samples.tsv
+outdir: ./results
ref:
- genome_size: "4.4m"
- gbf: "resources/H37Rv.gbf"
+ genome_size: 4.4m
+ gbf: resources/H37Rv.gbf
-bakta_db: "/scicore/home/gagneux/GROUP/PacbioSnake_resources/databases/bakta_db"
-container: "/scicore/home/gagneux/GROUP/PacbioSnake_resources/containers/assemblySC.sif"
-
-annotate: "No"
+bakta_db: /scicore/home/gagneux/GROUP/PacbioSnake_resources/databases/bakta_db
+container: /scicore/home/gagneux/GROUP/PacbioSnake_resources/containers/assemblySC.sif
threads_per_job: 4
diff --git a/assembly/config/environment.yml b/assembly/config/environment.yml
index 1946e19..f5e4dfb 100644
--- a/assembly/config/environment.yml
+++ b/assembly/config/environment.yml
@@ -7,4 +7,3 @@ channels:
dependencies:
- snakemake=7.32.4
- singularity=3.8.6
- - biopython
diff --git a/assembly/run_assembly_pipeline.py b/assembly/run_assembly_pipeline.py
index 490b900..388c616 100755
--- a/assembly/run_assembly_pipeline.py
+++ b/assembly/run_assembly_pipeline.py
@@ -2,6 +2,7 @@
import argparse
import os
+import yaml
import sys
def get_args():
@@ -11,17 +12,15 @@ def get_args():
# Parameter groups
parser_io = parser.add_argument_group('INPUT/OUTPUT')
- parser_cluster = parser.add_argument_group('CLUSTER CONFIGURATION (not implemented yet)')
+ parser_cluster = parser.add_argument_group('CLUSTER CONFIGURATION')
# INPUT/OUTPUT
- parser_io.add_argument('-s', '--samples', required=True, help='Absolute path to tab-separated table, no header, with sample name and path to fastq with HiFi reads.')
+ parser_io.add_argument('-s', '--samples', required=True, help='Path to tab-separeted table, no header, with sample name and path to fastq with HiFi reads.')
- parser_io.add_argument('-o', '--outdir', required=True, help='Absolute path to output directory.')
+ parser_io.add_argument('-o', '--outdir', required=True, help='Output directory for the results.')
- parser_io.add_argument('-n', '--dry_run', action='store_true', help='Do snakemake dry run.')
-
- # CLUSTER CONFIG (not implemented, would have to temper with the cluster config file)
+ # CLUSTER CONFIG
parser_cluster.add_argument('-j', '--njobs', default='4', help='Number of jobs to run in parallel. [4]')
parser_cluster.add_argument('-t', '--threads', default='10', help='Threads per job. [10]' )
@@ -37,7 +36,8 @@ def main():
# Infer pipeline location from path of run_assembly_pipeline.py
pl_path = os.path.dirname(os.path.abspath(sys.argv[0]))
-
+ print(pl_path)
+
# Directories for which singularity needs to be given access
bind_dirs = [
"/scicore/home/gagneux/GROUP/tbresearch/genomes/IN_PROGRESS/PacBio_genomes/Gagneux",
@@ -47,43 +47,24 @@ def main():
pl_path
]
- # Infer folders with samples, to add them to bind_dirs
- sample_dirs = set()
- with open(args.samples) as f:
- for line in f:
- fields = line.strip().split()
- fastq_path = fields[1]
- fastq_dir = os.path.dirname(os.path.realpath(fastq_path))
- sample_dirs.add(fastq_dir)
-
- bind_dirs = bind_dirs + list(sample_dirs)
-
singularity_args = "--bind " + " --bind ".join(bind_dirs)
- if args.dry_run:
-
- cmd = [
- "snakemake -n",
- "--snakefile", pl_path + "/workflow/Snakefile",
- "--directory", pl_path,
- "--configfile", pl_path + "/config/config.yaml",
- "--config", "samples=\"" + args.samples + "\"" + " outdir=\"" + args.outdir + "\""
- ]
-
- else:
- cmd = [
- "snakemake",
- "--snakefile", pl_path + "/workflow/Snakefile",
- "--directory", pl_path,
- "--configfile", pl_path + "/config/config.yaml",
- "--profile", pl_path + "/cluster",
- "--use-singularity",
- "--singularity-args" + " \"" + singularity_args + "\"",
- # Overwrite samples and outdir parameters in configfile
- "--config", "samples=\"" + args.samples + "\"" + " outdir=\"" + args.outdir + "\""
- ]
-
- print("\n" + " ".join(cmd) + "\n")
+ cmd = [
+ "snakemake",
+ "--snakefile", pl_path + "/workflow/Snakefile",
+ "--directory", pl_path,
+ "--configfile", pl_path + "/config/config.yaml",
+ "--profile", pl_path + "/config/cluster_config.yaml",
+ # Overwrite samples and outdir parameters
+ "--config", "samples=" + args.samples,
+ "--config", "outdir=" + args.outdir,
+ "--jobs", args.njobs,
+ "--cleanup-shadow",
+ "--use-singularity",
+ "--singularity-args" + " \"" + singularity_args + "\""
+ ]
+
+ #print(" ".join(cmd))
os.system(" ".join(cmd))
if __name__ == '__main__':
--
GitLab