#!/usr/bin/env python3
"""Preserve complete notices for the actual locked runtime/build crate graph."""
import argparse
import hashlib
import json
from pathlib import Path
import re
import subprocess

REVIEWED = {'GPL-3.0-only', 'MIT', 'MIT OR Apache-2.0', 'Apache-2.0 OR MIT',
            'MIT/Apache-2.0', 'Unlicense OR MIT', 'Unlicense/MIT', 'MPL-2.0',
            'Apache-2.0', 'Apache-2.0 OR BSL-1.0',
            'Apache-2.0 WITH LLVM-exception OR Apache-2.0 OR MIT',
            'BSD-2-Clause OR Apache-2.0 OR MIT',
            'Zlib OR Apache-2.0 OR MIT', 'MIT OR Apache-2.0 OR Zlib',
            '(MIT OR Apache-2.0) AND Unicode-3.0'}
NOTICE_NAME = re.compile(r'^(licen[cs]e|copying|notice|unlicense|copyright)([-.]|$)', re.I)
p = argparse.ArgumentParser(description=__doc__)
p.add_argument('destination', type=Path)
p.add_argument('--revision', required=True)
a = p.parse_args()
root = Path(__file__).resolve().parent.parent
host = next(x[6:] for x in subprocess.check_output(['rustc', '-vV'], text=True).splitlines() if x.startswith('host: '))
metadata = json.loads(subprocess.check_output(['cargo', 'metadata', '--locked', '--format-version', '1', '--filter-platform', host], cwd=root, text=True))
# Cargo metadata includes dev-unified/optional edges; use Cargo's actual normal
# and build tree to select the shipped graph instead of guessing feature edges.
tree = subprocess.check_output(['cargo', 'tree', '--locked', '-p', 'sized', '--target', host, '--edges', 'normal,build', '--prefix', 'none', '--format', '{p}'], cwd=root, text=True)
selected = {tuple(re.match(r'^(\S+) v(\S+)', line).groups()) for line in tree.splitlines()}
packages = sorted([x for x in metadata['packages'] if (x['name'], x['version']) in selected], key=lambda x: (x['name'], x['version']))
assert len(packages) == len(selected), 'Ambiguous or missing package identity'
texts, inventory = {}, []
for package in packages:
    if package['license'] not in REVIEWED:
        raise ValueError(f"Unreviewed licence: {package['name']} {package['license']}")
    directory = Path(package['manifest_path']).parent
    files = sorted(f for f in directory.iterdir() if f.is_file() and NOTICE_NAME.match(f.name))
    if not files:
        raise ValueError(f"Missing licence text: {package['name']}")
    notices = []
    for f in files:
        data = f.read_bytes(); text = data.decode('utf-8'); assert text.strip()
        digest = hashlib.sha256(data).hexdigest()
        texts.setdefault(digest, {'labels': [], 'text': text})['labels'].append(f"{package['name']} {package['version']} / {f.name}")
        notices.append({'file': f.name, 'sha256': digest})
    inventory.append({'name': package['name'], 'version': package['version'], 'license': package['license'], 'notices': notices})
header = ['Sized - Licences and credits', '', f'Source revision: {a.revision}',
          f'Target: {host}', '', 'Sized and its core are GPL-3.0-only, without warranty.',
          'Matching source, including the pinned core and dependency sources, is available at:',
          'https://gamertan.com/projects/sized/ and the matching Gitea release.',
          'The colored 2.2.0 source remains under MPL-2.0 and is additionally distributed',
          'under GPL-3.0-only as part of this Larger Work under MPL section 3.3.',
          'All original notices are preserved. See docs/RELEASE-LICENSING.md in the source.', '',
          'Runtime and build dependencies (development-only dependencies are excluded):']
header += [f"  {x['name']} {x['version']} - {x['license']}" for x in inventory]
for x in texts.values():
    header += ['', '='*72, '\n'.join(x['labels']), '='*72, x['text']]
a.destination.mkdir(parents=True, exist_ok=True)
(a.destination/'LICENCES.txt').write_text('\n'.join(header)+'\n')
(a.destination/'DEPENDENCIES.json').write_text(json.dumps({'revision': a.revision, 'target': host, 'packages': inventory}, indent=2)+'\n')
print(f'Packaged complete notices for {len(inventory)} runtime/build crates ({host}).')