Skip to content

Commit

Permalink
MAINT: code refactoring
Browse files Browse the repository at this point in the history
  • Loading branch information
@Vini2
Vini2 committed Jul 31, 2023
1 parent a7d7b8d commit 8b49a34
Show file tree
Hide file tree
Showing 7 changed files with 255 additions and 173 deletions.
345 changes: 175 additions & 170 deletions graphbin2
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,10 +2,12 @@

"""graphbin2: Refined and overlapped binning of metagenomic contigs using assembly graphs."""

import argparse
import click
import os
import sys
import subprocess
import logging

from src import (graphbin2_SPAdes, graphbin2_SGA, graphbin2_Flye)

__author__ = "Vijini Mallawaarachchi, Anuradha Wickramarachchi, and Yu Lin"
__copyright__ = "Copyright 2020, GraphBin2 Project"
Expand All @@ -15,202 +17,205 @@ __maintainer__ = "Vijini Mallawaarachchi"
__email__ = "viji.mallawaarachchi@gmail.com"
__status__ = "Stable Release"

def main():

parser = argparse.ArgumentParser(description="""GraphBin2 Help. GraphBin2 is a tool which refines the binning results obtained from existing tools and,
more importantly, is able to assign contigs to multiple bins. GraphBin2 uses the connectivity and coverage information from assembly graphs to
adjust existing binning results on contigs and to infer contigs shared by multiple species.""")

parser.add_argument("--assembler",
required=True,
type=str,
help="name of the assembler used (SPAdes, SGA or Flye)")

parser.add_argument("--graph",
required=True,
type=str,
help="path to the assembly graph file")

parser.add_argument("--contigs",
required=True,
type=str,
help="path to the contigs file")

parser.add_argument("--paths",
required=False,
type=str,
help="path to the contigs.paths file")

parser.add_argument("--abundance",
required=False,
type=str,
help="path to the abundance file")

parser.add_argument("--binned",
required=True,
type=str,
help="path to the .csv file with the initial binning output from an existing tool")

parser.add_argument("--output",
required=True,
type=str,
help="path to the output folder")

parser.add_argument("--prefix",
required=False,
type=str,
default='',
help="prefix for the output file")

parser.add_argument("--depth",
required=False,
type=int,
default=5,
help="maximum depth for the breadth-first-search. [default: 5]")

parser.add_argument("--threshold",
required=False,
type=float,
default=1.5,
help="threshold for determining inconsistent vertices. [default: 1.5]")

parser.add_argument("--delimiter",
required=False,
type=str,
default=",",
help="delimiter for input/output results. Supports a comma (,), a semicolon (;), a tab ($'\\t'), a space (\" \") and a pipe (|) [default: , (comma)]")

parser.add_argument("--nthreads",
required=False,
type=int,
default=8,
help="number of threads to use. [default: 8]")

args = vars(parser.parse_args())


assembler = args["assembler"]
assembly_graph_file = args["graph"]
contigs = args["contigs"]
contig_paths = args["paths"]
abundance = args["abundance"]
contig_bins_file = args["binned"]
output_path = args["output"]
prefix = args["prefix"]
depth = args["depth"]
threshold = args["threshold"]
delimiter = args["delimiter"]
nthreads = args["nthreads"]

@click.command()
@click.option(
"--assembler",
help="name of the assembler used. (Supports SPAdes, SGA and Flye)",
type=click.Choice(["spades", "sga", "flye"], case_sensitive=False),
required=True,
)
@click.option(
"--graph",
help="path to the assembly graph file",
type=click.Path(exists=True),
required=True,
)
@click.option(
"--contigs",
help="path to the contigs file",
type=click.Path(exists=True),
required=True,
)
@click.option(
"--paths",
help="path to the contigs.paths (metaSPAdes) or assembly.info (metaFlye) file",
type=click.Path(exists=True),
required=False,
)
@click.option(
"--abundance",
help="path to the abundance file",
type=click.Path(exists=True),
required=True,
)
@click.option(
"--binned",
help="path to the .csv file with the initial binning output from an existing toole",
type=click.Path(exists=True),
required=True,
)
@click.option(
"--output",
help="path to the output folder",
type=click.Path(dir_okay=True, writable=True, readable=True),
required=True,
)
@click.option(
"--prefix",
help="prefix for the output file",
type=str,
required=False,
)
@click.option(
"--depth",
help="maximum depth for the breadth-first-search.",
type=int,
default=5,
show_default=True,
required=False,
)
@click.option(
"--threshold",
help="threshold for determining inconsistent vertices.",
type=float,
default=1.5,
show_default=True,
required=False,
)
@click.option(
"--delimiter",
help="delimiter for output results. Supports a comma (,), a semicolon (;), a tab ($'\\t'), a space (\" \") and a pipe (|) .",
type=click.Choice([",", ";", "$'\\t'", "\" \""], case_sensitive=False),
default=",",
show_default=True,
required=False,
)
@click.option(
"--nthreads",
help="number of threads to use.",
type=int,
default=8,
show_default=True,
required=False,
)
def main(
assembler,
graph,
contigs,
paths,
abundance,
binned,
output,
prefix,
depth,
threshold,
delimiter,
nthreads
):

"""
GraphBin2: Refined and Overlapped Binning of Metagenomic Contigs Using Assembly Graphs
"""

assembly_graph_file = graph
contigs_file = contigs
contig_paths = paths
abundance_file = abundance
contig_bins_file = binned
output_path = output

# Validation of inputs
#---------------------------------------------------

# Check assembler type
if not (assembler.lower() == "spades" or assembler.lower() == "sga" or assembler.lower() == "flye"):
print("\nPlease make sure to provide the correct assembler type (SPAdes, SGA or Flye).")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Check assembly graph file
if not os.path.isfile(assembly_graph_file):
print("\nFailed to open the assembly graph file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)
# Validate prefix
if prefix != None:
if not prefix.endswith("_"):
prefix = prefix + "_"
else:
prefix = ''

# Check contigs file
if not os.path.isfile(contigs):
print("\nFailed to open the contigs file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Check if paths file is provided when the assembler type is SPAdes
if assembler.lower() == "spades" and contig_paths is None:
print("\nPlease make sure to provide the path to the contigs.paths file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)
# Setup logger
#-----------------------
logger = logging.getLogger('GraphBin2 1.2.0')
logger.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
consoleHeader = logging.StreamHandler()
consoleHeader.setFormatter(formatter)
consoleHeader.setLevel(logging.INFO)
logger.addHandler(consoleHeader)

# Check contigs.paths file for SPAdes
if assembler.lower() == "spades" and not os.path.isfile(contig_paths):
print("\nFailed to open the contigs.paths file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)
# Setup output path for log file
#---------------------------------------------------

# Check if abundance file is provided when the assembler type is SGA
if assembler.lower() == "sga" and abundance is None:
print("\nPlease make sure to provide the path to the abundance file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)
fileHandler = logging.FileHandler(f"{output_path}/{prefix}graphbin2.log")
fileHandler.setLevel(logging.DEBUG)
fileHandler.setFormatter(formatter)
logger.addHandler(fileHandler)

# Check abundance file for SGA
if assembler.lower() == "sga" and not os.path.isfile(abundance):
print("\nFailed to open the abundance file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Check if abundance file is provided when the assembler type is Flye
if assembler.lower() == "flye" and abundance is None:
print("\nPlease make sure to provide the path to the abundance file.")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)
# Validation of inputs
#---------------------------------------------------

# Check abundance file for Flye
if assembler.lower() == "flye" and not os.path.isfile(abundance):
print("\nFailed to open the abundance file.")
print("Exiting GraphBin2...\nBye...!\n")
# Check if paths file is provided when the assembler type is SPAdes
if assembler.lower() == "spades" and contig_paths is None:
logger.error("Please make sure to provide the path to the contigs.paths file.")
logger.info("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Check the file with the initial binning output
if not os.path.isfile(contig_bins_file):
print("\nFailed to open the file with the initial binning output.")
print("Exiting GraphBin2...\nBye...!\n")
# Check if paths file is provided when the assembler type is SPAdes
if assembler.lower() == "flye" and contig_paths is None:
logger.error("Please make sure to provide the path to the assembly_info.txt file.")
logger.info("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Handle for missing trailing forwardslash in output folder path
if output_path[-1:] != "/":
output_path = output_path + "/"

# Create output folder if it does not exist
if not os.path.isdir(output_path):
subprocess.run("mkdir -p "+output_path, shell=True)

# Validate prefix
if args["prefix"] != '':
if args["prefix"].endswith("_"):
prefix = args["prefix"]
else:
prefix = args["prefix"]+"_"
else:
prefix = ''
# Validate paths
if assembler.lower() == "sga":
contig_paths = "None"

# Validate depth
if depth < 1:
print("\nPlease enter a valid number for depth")
print("Exiting GraphBin2...\nBye...!\n")
logger.error("Please enter a valid number for depth")
logger.info("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Validate threshold
if threshold < 1.0:
print("\nPlease enter a valid number for threshold")
print("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)

# Validate delimiter
delimiters = [",", ";", " ", "\t", "|"]

if delimiter not in delimiters:
print("\nPlease enter a valid delimiter")
print("Exiting GraphBin2...\nBye...!\n")
logger.error("Please enter a valid number for threshold")
logger.info("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)


# Validate number of threads
if nthreads <= 0:
print("\nPlease enter a valid number for the number of threads")
print("Exiting GraphBin2...\nBye...!\n")
logger.error("Please enter a valid number for the number of threads")
logger.info("Exiting GraphBin2...\nBye...!\n")
sys.exit(1)


# Start GraphBin2
#---------------------------------------------------

logger.info("Welcome to GraphBin2: Refined and Overlapped Binning of Metagenomic Contigs using Assembly Graphs.")

if assembler.lower() == "spades":
logger.info("This version of GraphBin2 makes use of the assembly graph produced by SPAdes which is based on the de Bruijn graph approach.")
elif assembler.lower() == "sga":
logger.info("This version of GraphBin2 makes use of the assembly graph produced by SGA which is based on the string graph approach.")
elif assembler.lower() == "flye":
logger.info("This version of GraphBin2 makes use of the assembly graph produced by metaFlye which is a long reads assembler based on the de Bruijn graph approach.")

logger.info("Input arguments:")
logger.info(f"Contigs file: {contigs_file}")
logger.info(f"Assembly graph file: {assembly_graph_file}")
logger.info(f"Contig paths file: {contig_paths}")
logger.info(f"Existing binning output file: {contig_bins_file}")
logger.info(f"Final binning output file: {output_path}")
logger.info(f"Depth: {depth}")
logger.info(f"Threshold: {threshold}")
logger.info(f"Number of threads: {nthreads}")

logger.info("GraphBin2 started")


# Run GraphBin2
#---------------------------------------------------
if assembler.lower() == "spades":
Expand Down
Empty file added src/__init__.py
View file
Empty file.
Empty file added src/bidirectionalmap/__init__.py
View file
Empty file.
Loading

0 comments on commit 8b49a34

Please sign in to comment.