xref: /openbmc/openbmc/poky/scripts/pull-sdpx-licenses.py (revision 8460358c3d24c71d9d38fd126c745854a6301564) 
1#! /usr/bin/env python3
2#
3# Copyright OpenEmbedded Contributors
4#
5# SPDX-License-Identifier: GPL-2.0-only
6
7import argparse
8import json
9import sys
10import urllib.request
11from pathlib import Path
12
13TOP_DIR = Path(__file__).parent.parent
14
15
16def main():
17    parser = argparse.ArgumentParser(
18        description="Update SPDX License files from upstream"
19    )
20    parser.add_argument(
21        "-v",
22        "--version",
23        metavar="MAJOR.MINOR[.MICRO]",
24        help="Pull specific version of License list instead of latest",
25    )
26    parser.add_argument(
27        "--overwrite",
28        action="store_true",
29        help="Update existing license file text with upstream text",
30    )
31    parser.add_argument(
32        "--deprecated",
33        action="store_true",
34        help="Update deprecated licenses",
35    )
36    parser.add_argument(
37        "--dest",
38        type=Path,
39        default=TOP_DIR / "meta" / "files" / "common-licenses",
40        help="Write licenses to directory DEST. Default is %(default)s",
41    )
42
43    args = parser.parse_args()
44
45    if args.version:
46        version = f"v{args.version}"
47    else:
48        # Fetch the latest release
49        req = urllib.request.Request(
50            "https://api.github.com/repos/spdx/license-list-data/releases/latest"
51        )
52        req.add_header("X-GitHub-Api-Version", "2022-11-28")
53        req.add_header("Accept", "application/vnd.github+json")
54        with urllib.request.urlopen(req) as response:
55            data = json.load(response)
56            version = data["tag_name"]
57
58    print(f"Pulling SPDX license list version {version}")
59    req = urllib.request.Request(
60        f"https://raw.githubusercontent.com/spdx/license-list-data/{version}/json/licenses.json"
61    )
62    with urllib.request.urlopen(req) as response:
63        spdx_licenses = json.load(response)
64
65    with (TOP_DIR / "meta" / "files" / "spdx-licenses.json").open("w") as f:
66        json.dump(spdx_licenses, f, sort_keys=True, indent=2)
67
68    total_count = len(spdx_licenses["licenses"])
69    updated = 0
70    for idx, lic in enumerate(spdx_licenses["licenses"]):
71        lic_id = lic["licenseId"]
72
73        print(f"[{idx + 1} of {total_count}] ", end="")
74
75        dest_license_file = args.dest / lic_id
76        if dest_license_file.is_file() and not args.overwrite:
77            print(f"Skipping {lic_id} since it already exists")
78            continue
79
80        print(f"Fetching {lic_id}... ", end="", flush=True)
81
82        req = urllib.request.Request(lic["detailsUrl"])
83        with urllib.request.urlopen(req) as response:
84            lic_data = json.load(response)
85
86        if lic_data["isDeprecatedLicenseId"] and not args.deprecated:
87            print("Skipping (deprecated)")
88            continue
89
90        with dest_license_file.open("w") as f:
91            f.write(lic_data["licenseText"])
92        updated += 1
93        print("done")
94
95    print(f"Updated {updated} licenses")
96
97    return 0
98
99
100if __name__ == "__main__":
101    sys.exit(main())
102