aboutsummaryrefslogtreecommitdiff
path: root/scripts
diff options
context:
space:
mode:
authorFrederick Muriuki Muriithi2022-04-27 19:06:45 +0300
committerFrederick Muriuki Muriithi2022-04-27 19:14:37 +0300
commitaadd9aa5dd4c552b573828ddac581a8b7064b0e2 (patch)
treefbf17e017af4f736ec5ea797cd777396768decfe /scripts
parent903af1c0b1f2cc695ea4e0c31438f9205571d15d (diff)
downloadgn-uploader-aadd9aa5dd4c552b573828ddac581a8b7064b0e2.tar.gz
Enable managing app via setup.py
While the application is developed with GNU Guix, the end user might not be using it, and therefore, this commit provides a way for the user to install the application with the usual python package management systems.
Diffstat (limited to 'scripts')
-rw-r--r--scripts/__init__.py0
-rw-r--r--scripts/qc.py108
2 files changed, 108 insertions, 0 deletions
diff --git a/scripts/__init__.py b/scripts/__init__.py
new file mode 100644
index 0000000..e69de29
--- /dev/null
+++ b/scripts/__init__.py
diff --git a/scripts/qc.py b/scripts/qc.py
new file mode 100644
index 0000000..09758cb
--- /dev/null
+++ b/scripts/qc.py
@@ -0,0 +1,108 @@
+"""Implements the command-line interface for the qc application"""
+import os
+import sys
+import argparse
+
+import magic
+
+from quality_control.errors import ParseError
+from quality_control.parsing import (
+ FileType,
+ parse_file,
+ strain_names,
+ parse_errors,
+ parse_strains)
+
+
+def is_file_mime(filepath, mimetype):
+ """Check that `filepath` has a mimetype of `mimetype` or `text/plain`"""
+ return magic.from_file(filepath, mime=True) in ("text/plain", mimetype)
+
+def cli_argument_parser():
+ """Create the parser for the CLI arguments"""
+ parser = argparse.ArgumentParser(
+ prog="qc", description = (
+ "Command-Line Interface program for quality control of data files"))
+ parser.add_argument(
+ "filetype",
+ help="The type of file to check",
+ choices=("average", "standard-error"))
+ parser.add_argument(
+ "filepath",
+ help=(
+ "The path to the file to be checked."
+ "If an absolute path is not provided, then the file will be relative to"
+ f"\t'{os.getcwd()}'"))
+ default_strains_file = os.path.join(
+ os.path.dirname(os.path.dirname(__file__)), "etc/strains.csv")
+ parser.add_argument(
+ "-s", "--strainsfile",
+ help=(
+ "Path to the file containing allowable strains/samples. "
+ f"[default '{default_strains_file}']"),
+ default=default_strains_file)
+
+ parser.add_argument(
+ "-v", "--verbose",
+ help="Controls whether to show extra output",
+ default=False, action="store_true")
+ return parser
+
+def check(filepath, filetype, strains, verbose=False):
+ """Check the file and print out results"""
+ try:
+ for line_num, line in enumerate(parse_file(
+ filepath, filetype, strains), start=1):
+ if verbose:
+ print(f"Checked line: {line_num}")
+
+ print(f"Successfully checked the file. No errors found.")
+ except ParseError as pe:
+ print("line\terrors")
+ for line_num, error in enumerate(
+ parse_errors(filepath, filetype, strains,
+ pe.args[0]["line_number"]),
+ start = pe.args[0]["line_number"] + 1):
+ print(f"{line_num}\t{' '.join(error['message'])}")
+
+
+def main():
+ """Entry point function"""
+ argparser = cli_argument_parser()
+ args = argparser.parse_args()
+ if not os.path.exists(args.filepath):
+ print("The file '{args.filepath}' does not exist.", file=sys.stderr)
+ return 1
+
+ if not os.path.exists(args.strainsfile):
+ print(f"The file '{args.strainsfile}' does not exist.", file=sys.stderr)
+ return 2
+
+ if not is_file_mime(args.filepath, "text/tab-separated-values"):
+ print(
+ f"The file '{args.filepath}' MUST be a tab-separated file.",
+ file=sys.stderr)
+ return 3
+
+ if not is_file_mime(args.strainsfile, "text/csv"):
+ print(
+ f"The file '{args.strainsfile}' MUST be a tab-separated file.",
+ file=sys.stderr)
+ return 4
+
+ if args.verbose:
+ print(f"Parsing the strain names from '{args.strainsfile}'")
+
+ strains = strain_names(parse_strains(os.path.realpath(args.strainsfile)))
+
+ filepath = os.path.realpath(args.filepath)
+ if args.verbose:
+ print(f"Checking '{filepath}' for errors")
+
+ check(
+ filepath, (
+ FileType.AVERAGE if args.filetype == "average"
+ else FileType.STANDARD_ERROR), strains)
+
+if __name__ == "__main__":
+ main()