1
0
Fork 0
mirror of https://github.com/MillironX/taxprofiler.git synced 2024-11-21 21:46:03 +00:00
taxprofiler/bin/check_samplesheet.py
James Fellows Yates 7ce55a99e8 Fix linting
2023-02-02 13:08:16 +01:00

231 lines
8.1 KiB
Python
Executable file

#!/usr/bin/env python
from distutils import extension
import os
import sys
import errno
import argparse
def parse_args(args=None):
Description = "Reformat nf-core/taxprofiler samplesheet file and check its contents."
Epilog = "Example usage: python check_samplesheet.py <FILE_IN> <FILE_OUT>"
parser = argparse.ArgumentParser(description=Description, epilog=Epilog)
parser.add_argument("FILE_IN", help="Input samplesheet file.")
parser.add_argument("FILE_OUT", help="Output file.")
return parser.parse_args(args)
def make_dir(path):
if len(path) > 0:
try:
os.makedirs(path)
except OSError as exception:
if exception.errno != errno.EEXIST:
raise exception
def print_error(error, context="Line", context_str=""):
error_str = "ERROR: Please check samplesheet -> {}".format(error)
if context != "" and context_str != "":
error_str = "ERROR: Please check samplesheet -> {}\n{}: '{}'".format(
error, context.strip(), context_str.strip()
)
print(error_str)
sys.exit(1)
def check_samplesheet(file_in, file_out):
"""
This function checks that the samplesheet follows the following structure:
sample,run_accession,instrument_platform,fastq_1,fastq_2,fasta
2611,ERR5766174,ILLUMINA,,,ERX5474930_ERR5766174_1.fa.gz
2612,ERR5766176,ILLUMINA,ERX5474932_ERR5766176_1.fastq.gz,ERX5474932_ERR5766176_2.fastq.gz,
2612,ERR5766174,ILLUMINA,ERX5474936_ERR5766180_1.fastq.gz,,
2613,ERR5766181,ILLUMINA,ERX5474937_ERR5766181_1.fastq.gz,ERX5474937_ERR5766181_2.fastq.gz,
"""
FQ_EXTENSIONS = (".fq.gz", ".fastq.gz")
FA_EXTENSIONS = (
".fa.gz",
".fasta.gz",
".fna.gz",
".fas.gz",
)
INSTRUMENT_PLATFORMS = [
"ABI_SOLID",
"BGISEQ",
"CAPILLARY",
"COMPLETE_GENOMICS",
"DNBSEQ",
"HELICOS",
"ILLUMINA",
"ION_TORRENT",
"LS454",
"OXFORD_NANOPORE",
"PACBIO_SMRT",
]
sample_mapping_dict = {}
with open(file_in, "r") as fin:
## Check header
MIN_COLS = 4
HEADER = [
"sample",
"run_accession",
"instrument_platform",
"fastq_1",
"fastq_2",
"fasta",
]
header = [x.strip('"') for x in fin.readline().strip().split(",")]
## Check for missing mandatory columns
missing_columns = list(set(HEADER) - set(header))
if len(missing_columns) > 0:
print(
"ERROR: Missing required column header -> {}. Note some columns can otherwise be empty. See pipeline documentation (https://nf-co.re/taxprofiler/usage).".format(
",".join(missing_columns)
)
)
sys.exit(1)
## Find locations of mandatory columns
header_locs = {}
for i in HEADER:
header_locs[i] = header.index(i)
## Check sample entries
for line in fin:
## Pull out only relevant columns for downstream checking
line_parsed = [x.strip().strip('"') for x in line.strip().split(",")]
# Check valid number of columns per row
if len(line_parsed) < len(HEADER):
print_error(
"Invalid number of columns (minimum = {})!".format(len(HEADER)),
"Line",
line,
)
num_cols = len([x for x in line_parsed if x])
if num_cols < MIN_COLS:
print_error(
"Invalid number of populated columns (minimum = {})!".format(MIN_COLS),
"Line",
line,
)
lspl = [line_parsed[i] for i in header_locs.values()]
## Check sample name entries
(
sample,
run_accession,
instrument_platform,
fastq_1,
fastq_2,
fasta,
) = lspl[: len(HEADER)]
sample = sample.replace(" ", "_")
if not sample:
print_error("Sample entry has not been specified!", "Line", line)
## Check FastQ file extension
for fastq in [fastq_1, fastq_2]:
if fastq:
if fastq.find(" ") != -1:
print_error("FastQ file contains spaces!", "Line", line)
if not fastq.endswith(FQ_EXTENSIONS):
print_error(
f"FastQ file does not have extension {' or '.join(list(FQ_EXTENSIONS))} !",
"Line",
line,
)
if fasta:
if fasta.find(" ") != -1:
print_error("FastA file contains spaces!", "Line", line)
if not fasta.endswith(FA_EXTENSIONS):
print_error(
f"FastA file does not have extension {' or '.join(list(FA_EXTENSIONS))}!",
"Line",
line,
)
sample_info = []
# Check run_accession
if not run_accession:
print_error("Run accession has not been specified!", "Line", line)
else:
sample_info.append(run_accession)
# Check instrument_platform
if not instrument_platform:
print_error("Instrument platform has not been specified!", "Line", line)
else:
if instrument_platform not in INSTRUMENT_PLATFORMS:
print_error(
f"Instrument platform {instrument_platform} is not supported! "
f"List of supported platforms {', '.join(INSTRUMENT_PLATFORMS)}",
"Line",
line,
)
sample_info.append(instrument_platform)
## Auto-detect paired-end/single-end
if sample and fastq_1 and fastq_2: ## Paired-end short reads
sample_info.extend(["0", fastq_1, fastq_2, fasta])
elif sample and fastq_1 and not fastq_2: ## Single-end short/long fastq reads
sample_info.extend(["1", fastq_1, fastq_2, fasta])
elif sample and fasta and not fastq_1 and not fastq_2: ## Single-end long reads
sample_info.extend(["1", fastq_1, fastq_2, fasta])
elif fasta and (fastq_1 or fastq_2):
print_error(
"FastQ and FastA files cannot be specified together in the same library!",
"Line",
line,
)
else:
print_error("Invalid combination of columns provided!", "Line", line)
## Create sample mapping dictionary = { sample: [ run_accession, instrument_platform, single_end, fastq_1, fastq_2 , fasta ] }
if sample not in sample_mapping_dict:
sample_mapping_dict[sample] = [sample_info]
else:
if sample_info in sample_mapping_dict[sample]:
print_error("Samplesheet contains duplicate rows!", "Line", line)
else:
sample_mapping_dict[sample].append(sample_info)
## Write validated samplesheet with appropriate columns
HEADER_OUT = [
"sample",
"run_accession",
"instrument_platform",
"single_end",
"fastq_1",
"fastq_2",
"fasta",
]
if len(sample_mapping_dict) > 0:
out_dir = os.path.dirname(file_out)
make_dir(out_dir)
with open(file_out, "w") as fout:
fout.write(",".join(HEADER_OUT) + "\n")
for sample in sorted(sample_mapping_dict.keys()):
for idx, val in enumerate(sample_mapping_dict[sample]):
fout.write(f"{sample},{','.join(val)}\n")
else:
print_error("No entries to process!", "Samplesheet: {}".format(file_in))
def main(args=None):
args = parse_args(args)
check_samplesheet(args.FILE_IN, args.FILE_OUT)
if __name__ == "__main__":
sys.exit(main())