Fetch sample/case names from database

Fetch the sample/case names from the database rather than from a
static file in the repository.

Issue:
https://issues.genenetwork.org/issues/quality-control/read-samples-from-database-by-species

1 files changed, 17 insertions, 28 deletions

diff --git a/scripts/qc.py b/scripts/qc.py
index 7d41d6c..e8573a9 100644
--- a/scripts/qc.py
+++ b/scripts/qc.py
@@ -1,9 +1,9 @@
 """Implements the command-line interface for the qc application"""
 import os
 import sys
-import argparse
 import mimetypes
 from typing import Union, Callable
+from argparse import ArgumentParser
 
 from functional_tools import take
 
@@ -11,17 +11,18 @@ from quality_control.utils import make_progress_calculator
 from quality_control.errors import InvalidValue, DuplicateHeading
 from quality_control.parsing import FileType, strain_names, collect_errors
 
+from qc_app.db_utils import database_connection
+
+from .cli_parser import init_cli_parser
+
 
 def is_file_mime(filepath:str, mimetype:str) -> bool:
     """Check that `filepath` has a mimetype of `mimetype` or `text/plain`"""
     the_type = mimetypes.guess_type(filepath)[0]
     return the_type 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"))
+def add_file_validation_arguments(parser: ArgumentParser) -> ArgumentParser:
+    """File validation specific CLI arguments."""
     parser.add_argument(
         "filetype",
         help="The type of file to check",
@@ -32,14 +33,6 @@ def cli_argument_parser():
             "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(
         "-c", "--count", type=int,
@@ -54,6 +47,14 @@ def cli_argument_parser():
         default=False, action="store_true")
     return parser
 
+def cli_argument_parser():
+    """Create the parser for the CLI arguments"""
+    theparser = init_cli_parser(
+        "qc",
+        "Command-Line Interface program for quality control of data files")
+    theparser.add_argument("speciesid", type=int, help="ID of the species.")
+    return add_file_validation_arguments(theparser)
+
 def make_progress_indicator(
         verbose: bool, progress_calc_fn: Callable) -> Union[Callable, None]:
     """Utility to display the progress"""
@@ -106,26 +107,14 @@ def main():
         print(f"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(os.path.realpath(args.strainsfile))
+    with database_connection(args.databaseuri) as dbconn:
+        strains = strain_names(dbconn, args.speciesid)
 
     filepath = os.path.realpath(args.filepath)
     if args.verbose: