Files
mrvacommander/client/qldbtools/bin/mc-db-generate-selection
2024-11-19 15:24:41 -08:00

109 lines
2.8 KiB
Python
Executable File

#!/usr/bin/env python
""" Read a table of CodeQL DB information
and generate the selection files for
1. the VS Code CodeQL plugin
2. the gh-mrva command-line client
"""
import argparse
import logging
from argparse import Namespace
from typing import List
from pandas import DataFrame
import qldbtools.utils as utils
import numpy as np
#
#* Configure logger
#
logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s')
# Overwrite log level set by minio
root_logger = logging.getLogger()
root_logger.setLevel(logging.INFO)
#
#* Process command line
#
parser = argparse.ArgumentParser(
description=""" Read a table of CodeQL DB information
and generate the selection files for
1. the VS Code CodeQL plugin
2. the gh-mrva command-line client
""",
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('vscode_selection', type=str,
help='VS Code selection file to generate')
parser.add_argument('gh_mrva_selection', type=str,
help='gh-mrva cli selection file to generate')
parser.add_argument('-n', '--num-entries', type=int,
help='Only use N entries',
default=None)
parser.add_argument('-s', '--seed', type=int,
help='Random number seed',
default=4242)
parser.add_argument('-l', '--list-name', type=str,
help='Name of the repository list',
default='mirva-list')
args: Namespace = parser.parse_args()
#
#* Load the information
#
import pandas as pd
import sys
df0: DataFrame = pd.read_csv(sys.stdin)
if args.num_entries == None:
# Use all entries
df1: DataFrame = df0
else:
# Use num_entries, chosen via pseudo-random numbers
df1 = df0.sample(n=args.num_entries,
random_state=np.random.RandomState(args.seed))
#
#* Form and save structures
#
repos: list[str] = []
for index, row in df1[['owner', 'name', 'CID', 'path']].iterrows():
owner, name, CID, path = row
repos.append(utils.form_db_req_name(owner, name, CID))
repo_list_name: str = args.list_name
vsc = {
"version": 1,
"databases": {
"variantAnalysis": {
"repositoryLists": [
{
"name": repo_list_name,
"repositories": repos,
}
],
"owners": [],
"repositories": []
}
},
"selected": {
"kind": "variantAnalysisUserDefinedList",
"listName": repo_list_name
}
}
gh = {
repo_list_name: repos
}
import json
with open(args.vscode_selection, "w") as fc:
json.dump(vsc, fc, indent=4)
with open(args.gh_mrva_selection, "w") as fc:
json.dump(gh, fc, indent=4)
# Local Variables:
# python-shell-virtualenv-root: "~/work-gh/mrva/mrvacommander/client/qldbtools/venv/"
# End: