fix tmpdir being set from base command

aviary/aviary.py

@@ -35,7 +35,6 @@
 import logging
 import os
 from datetime import datetime
-import tempfile
 
 # Debug
 debug={1:logging.CRITICAL,
@@ -198,7 +197,7 @@ def main():
         help='Path to the location that will be treated used for temporary files. If none is specified, the TMPDIR \n'
              'environment variable will be used. Can be configured within the `configure` subcommand',
         dest='tmpdir',
-        default=tempfile.gettempdir(),
+        default=None,
     )
 
     base_group.add_argument(
@@ -1290,9 +1289,6 @@ def manage_env_vars(args):
     if args.conda_prefix is None:
         args.conda_prefix = Config.get_software_db_path('CONDA_ENV_PATH', '--conda-prefix')
 
-    if args.tmpdir is None:
-        args.tmpdir = Config.get_software_db_path('TMPDIR', '--tmpdir')
-
     try:
         if args.gtdb_path is None:
             args.gtdb_path = Config.get_software_db_path('GTDBTK_DATA_PATH', '--gtdb-path')

aviary/modules/annotation/annotation.smk

@@ -171,7 +171,7 @@ rule eggnog:
     params:
         mag_extension = config['mag_extension'],
         eggnog_db = config['eggnog_folder'],
-        tmpdir = config["tmpdir"]
+        tmpdir = config["tmpdir"] if config["tmpdir"] else "$TMPDIR",
     output:
         done = 'data/eggnog/done'
     threads:

aviary/modules/assembly/assembly.smk

@@ -367,31 +367,8 @@ rule spades_assembly:
         "envs/spades.yaml"
     benchmark:
         "benchmarks/spades_assembly.benchmark.txt"
-    shell:
-        """
-        rm -rf data/spades_assembly/tmp; 
-        minimumsize=500000;
-        actualsize=$(stat -c%s data/short_reads.filt.fastq.gz);
-        if [ -d "data/spades_assembly/" ]
-        then
-            spades.py --restart-from last --memory {params.max_memory} -t {threads} -o data/spades_assembly -k {params.kmer_sizes} --tmp-dir {params.tmpdir} > {log} 2>&1 && \
-            cp data/spades_assembly/scaffolds.fasta data/spades_assembly.fasta
-        elif [ $actualsize -ge $minimumsize ]
-        then
-            if [ {params.long_read_type} = "ont" ] || [ {params.long_read_type} = "ont_hq" ]
-            then
-                spades.py --checkpoints all --memory {params.max_memory} --meta --nanopore {input.long_reads} --12 {input.fastq} \
-                -o data/spades_assembly -t {threads}  -k {params.kmer_sizes} --tmp-dir {params.tmpdir} >> {log} 2>&1 && \
-                cp data/spades_assembly/scaffolds.fasta data/spades_assembly.fasta
-            else
-                spades.py --checkpoints all --memory {params.max_memory} --meta --pacbio {input.long_reads} --12 {input.fastq} \
-                -o data/spades_assembly -t {threads}  -k {params.kmer_sizes} --tmp-dir {params.tmpdir} >> {log} 2>&1 && \
-                cp data/spades_assembly/scaffolds.fasta data/spades_assembly.fasta
-            fi
-        else
-            mkdir -p {output.spades_folder} && touch {output.fasta}
-        fi 
-        """
+    script:
+        "scripts/spades_assembly.py"        
 
 
 # Perform short read assembly only with no other steps
@@ -458,14 +435,14 @@ rule spades_assembly_coverage:
     log:
         "logs/spades_assembly_coverage.log"
     params:
-         tmpdir = config["tmpdir"]
+         tmpdir = f"TMPDIR={config['tmpdir']}" if config["tmpdir"] else ""
     conda:
          "../../envs/coverm.yaml"
     benchmark:
         "benchmarks/spades_assembly_coverage.benchmark.txt"
     shell:
         """
-        TMPDIR={params.tmpdir} coverm contig -m metabat -t {threads} -r {input.fasta} --interleaved {input.fastq} --bam-file-cache-directory data/cached_bams/ > {output.assembly_cov} 2> {log};
+        {params.tmpdir} coverm contig -m metabat -t {threads} -r {input.fasta} --interleaved {input.fastq} --bam-file-cache-directory data/cached_bams/ > {output.assembly_cov} 2> {log};
         mv data/cached_bams/*.bam {output.bam} && samtools index -@ {threads} {output.bam} 2>> {log}
         """
 

aviary/modules/assembly/scripts/assemble_short_reads.py

@@ -29,6 +29,15 @@ def assemble_short_reads(
     :return:
     '''
 
+    if not tmp_dir:
+        try:
+            tmp_dir = os.environ["TMPDIR"]
+        except KeyError:
+            tmp_dir_arg = ""
+
+    if tmp_dir:
+        tmp_dir_arg = f"--tmp-dir {tmp_dir}"
+
     # deal with read sets i.e. are we coassembling? Which assembler are we using?
     # Non co-assembled reads are handled the same for each assembler
     if read_set1 != 'none':
@@ -86,7 +95,7 @@ def assemble_short_reads(
     # Run chosen assembler
     if use_megahit:
         max_memory_in_bytes = max_memory * 1024*1024*1024
-        command = f"megahit {read_string} -t {threads} -m {max_memory_in_bytes} -o data/megahit_assembly --tmp-dir {tmp_dir}"
+        command = f"megahit {read_string} -t {threads} -m {max_memory_in_bytes} -o data/megahit_assembly {tmp_dir_arg}"
 
         with open(log, 'a') as logf:
             logf.write(f"Queueing command {command}\n")
@@ -97,7 +106,7 @@ def assemble_short_reads(
     else:
         kmers = " ".join(kmer_sizes)
         command = f"spades.py --memory {max_memory} --meta -t {threads} " \
-                f"-o data/short_read_assembly {read_string} -k {kmers} --tmp-dir {tmp_dir}"
+                f"-o data/short_read_assembly {read_string} -k {kmers} {tmp_dir_arg}"
         with open(log, 'a') as logf:
             logf.write(f"Queueing command {command}\n")
             subprocess.run(command.split(), stdout=logf, stderr=subprocess.STDOUT)

aviary/modules/assembly/scripts/spades_assembly.py

@@ -0,0 +1,89 @@
+import subprocess
+import os
+import sys
+from typing import List
+
+def spades_asssembly(
+    input_fastq: str,
+    input_long_reads: str,
+    output_fasta: str,
+    output_spades_folder: str,
+    max_memory: int,
+    threads: int,
+    kmer_sizes: List[int],
+    tmp_dir: str,
+    long_read_type: str,
+    log: str):
+
+    '''
+    Assemble short reads and long reads (if any) using spades
+    :param input_fastq: short reads fastq file
+    :param input_long_reads: long reads fastq file
+    :param output_fasta: output fasta file
+    :param output_spades_folder: output spades folder
+    :param max_memory: maximum memory to use
+    :param threads: number of threads
+    :param tmpdir: temporary directory
+    :param kmer_sizes: list of kmer sizes
+    :param long_read_type: type of long reads
+    :param log: log file
+    :return:
+    '''
+    if tmp_dir:
+        tmp_dir_arg = f"--tmp-dir {tmp_dir}"
+    else:
+        try:
+            tmp_dir = os.environ["TMPDIR"]
+            tmp_dir_arg = f"--tmp-dir {tmp_dir}"
+        except KeyError:
+            tmp_dir_arg = ""
+
+    if os.path.exists("data/spades_assembly/tmp"):
+        with open(log, 'a') as logf:
+            subprocess.run("rm -rf data/spades_assembly/tmp".split(), stdout=logf, stderr=subprocess.STDOUT)   
+    # remove existing temporary directory
+    minimumsize=500000
+    actualsize = int(subprocess.check_output('stat -c%s data/short_reads.filt.fastq.gz', shell=True))
+    # check if directory exists
+    if  os.path.exists("data/spades_assembly"):
+        # resume previous assembly
+        command = f"spades.py --restart-from last --memory {max_memory} -t {threads} " \
+                f"-o data/spades_assembly -k {kmer_sizes}  {tmp_dir_arg}" 
+        # run cmd
+        with open(log, 'a') as logf:
+            logf.write(f"Queueing command {command}\n")
+            subprocess.run(command.split(), stdout=logf, stderr=subprocess.STDOUT)
+            subprocess.run("cp data/spades_assembly/scaffolds.fasta data/spades_assembly.fasta".split(), stdout=logf, stderr=subprocess.STDOUT)
+    elif actualsize >= minimumsize:
+        if long_read_type in ["ont","ont_hq"]:
+            command = f"spades.py --checkpoints all --memory {max_memory} --meta --nanopore {input_long_reads} --12 {input_fastq} "\
+                f"-o data/spades_assembly -t {threads}  -k {kmer_sizes} {tmp_dir_arg} "       
+        else:
+            command = f"spades.py --checkpoints all --memory {max_memory} --meta --pacbio {input_long_reads} --12 {input_fastq} "\
+                f"-o data/spades_assembly -t {threads}  -k {kmer_sizes} {tmp_dir_arg} "
+        # run cmd
+        with open(log, 'a') as logf:
+            logf.write(f"Queueing command {command}\n")
+            subprocess.run(command.split(), stdout=logf, stderr=subprocess.STDOUT)        
+            subprocess.run("cp data/spades_assembly/scaffolds.fasta data/spades_assembly.fasta".split(), stdout=logf, stderr=subprocess.STDOUT)
+    else:
+        with open(log, 'a') as logf:
+            subprocess.run(f"mkdir -p {output.spades_folder} && touch {output.fasta}".split(), stdout=logf, stderr=subprocess.STDOUT)
+
+
+if __name__ == '__main__':
+    log = snakemake.log[0]
+    with open(log, 'w') as logf: pass
+
+    spades_asssembly(
+        snakemake.input.fastq,
+        snakemake.input.long_reads,
+        snakemake.output.fasta,
+        snakemake.output.spades_folder,
+        snakemake.params.max_memory,
+        snakemake.threads,
+        snakemake.params.kmer_sizes,
+        snakemake.params.tmpdir,
+        snakemake.params.long_read_type,
+        log
+    )

aviary/modules/binning/scripts/get_coverage.py

@@ -12,7 +12,8 @@ def get_coverage(
     threads: int,
     log: str,
 ):
-    os.environ["TMPDIR"] = tmpdir
+    if tmpdir: os.environ["TMPDIR"] = tmpdir
+
     if long_reads != "none" and not os.path.exists("data/long_cov.tsv"):
         if long_read_type in ["ont", "ont_hq"]:
             coverm_cmd = f"coverm contig -t {threads} -r {input_fasta} --single {' '.join(long_reads)} -p minimap2-ont -m length trimmed_mean variance --bam-file-cache-directory data/binning_bams/ --discard-unmapped --min-read-percent-identity 0.85 --output-file data/long_cov.tsv".split()

aviary/modules/processor.py

@@ -92,7 +92,7 @@ def __init__(self,
 
 
         self.conda_prefix = args.conda_prefix
-        self.tmpdir = os.path.abspath(args.tmpdir)
+        self.tmpdir = os.path.abspath(args.tmpdir) if args.tmpdir else None
         self.resources = args.resources
         self.output = os.path.abspath(args.output)
         self.threads = args.max_threads
@@ -444,7 +444,8 @@ def run_workflow(self, cores=16, profile=None, cluster_retries=None,
         self._validate_config()
 
         cores = max(int(self.threads), cores)
-        os.environ["TMPDIR"] = self.tmpdir
+        if self.tmpdir is not None:
+            os.environ["TMPDIR"] = self.tmpdir
         for workflow in self.workflows:
             cmd = (
                 "snakemake --snakefile {snakefile} --directory {working_dir} "

aviary/modules/strain_analysis/strain_analysis.smk

@@ -65,7 +65,6 @@ rule lorikeet:
     params:
         mag_extension = config['mag_extension'],
         parallel_genomes = 8,
-        tmpdir = config['tmpdir']
     resources:
         mem_mb=int(config["max_memory"])*512
     threads:

config.yaml

@@ -50,4 +50,4 @@ ani: none
 precluster_ani: none
 precluster_method: none
 pggb_params: none
-tmpdir: /tmp
+tmpdir: none