Datasets:
File size: 6,880 Bytes
7a088d4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 | #!/usr/bin/env python3
"""Re-read every serialized Kanripo record and verify it against raw sources."""
import argparse
from collections import Counter, defaultdict
import hashlib
import json
from pathlib import Path
from kanripo_clean import RULES, RESIDUAL, clean, load_inputs, sha, source_path
from kanripo_profile import dump
def records(path):
with path.open(encoding='utf-8') as f:
for line in f:
yield json.loads(line)
def file_hash(path):
h = hashlib.sha256()
with path.open('rb') as f:
for block in iter(lambda: f.read(1024*1024), b''):
h.update(block)
return h.hexdigest()
def check(condition, message):
if not condition:
raise ValueError(message)
def verify(output):
output = output.resolve()
manifest = json.loads((output / 'manifest.json').read_text())
check(manifest['complete'], 'incomplete build')
check(manifest['rules'] == RULES, 'rules changed')
check(manifest['script_sha256'] == file_hash(Path(__file__).with_name('kanripo_clean.py')), 'implementation changed')
root, audit = Path(manifest['root']), Path(manifest['audit'])
catalog, inventory, metadata, inventory_data = load_inputs(root, audit)
check(sha(metadata) == manifest['source_manifest_sha256'], 'metadata mismatch')
check(sha(inventory_data) == manifest['inventory_sha256'], 'inventory changed')
for name, digest in manifest['output_hashes'].items():
check(file_hash(output / name) == digest, f'output changed: {name}')
file_rows = list(records(output / 'files.jsonl'))
check(len(file_rows) == len(inventory), 'file coverage mismatch')
docs = iter(records(output / 'documents.jsonl'))
review = {r['id']: r for r in records(output / 'review.jsonl')}
edits = {r['id']: r for r in records(output / 'edits.jsonl')}
totals, removed, dynasties = Counter(), Counter(), Counter()
exact = defaultdict(list)
works, review_works = set(), set()
for i, (entry, f) in enumerate(zip(inventory, file_rows)):
rid = entry['id']
check(f['raw_path'] == entry['raw_path'], 'file order/coverage mismatch')
row = catalog[rid]
check(row['status'] == 'complete' and row['dynasty'] not in {'清', '淸'}, 'ineligible work')
raw = source_path(root, entry).read_bytes()
check(sha(raw) == entry['raw_sha256'] == f['raw_sha256'], 'raw hash mismatch')
original = raw.decode('utf-8')
text, counts, issues = clean(original)
digest = sha(text.encode('utf-8'))
expected_status = 'review' if issues else ('included' if text.strip() else 'empty')
check(f['status'] == expected_status, 'status mismatch')
check(f['original_chars'] == len(original) and f['clean_chars'] == len(text), 'length mismatch')
check(f['removed'] == counts and f['clean_sha256'] == digest, 'edit audit mismatch')
check((f['id'] in edits) == bool(counts), 'edit coverage mismatch')
if counts:
check(edits[f['id']] == f, 'edit record mismatch')
totals['input_files'] += 1
totals['original_chars'] += len(original)
totals[expected_status + '_files'] += 1
removed.update(counts)
if issues:
check(f['id'] in review and review[f['id']]['issues'] == issues, 'review mismatch')
review_works.add(rid)
elif expected_status == 'included':
doc = next(docs, None)
check(doc is not None and doc['id'] == f['id'], 'document order/coverage mismatch')
check(doc['text'] == text and doc['chars'] == len(text) and doc['clean_sha256'] == digest, 'serialized text mismatch')
for key in ['raw_path', 'raw_sha256', 'dynasty', 'label', 'source', 'work_id', 'classification_source', 'dynasty_raw']:
check(doc[key] == f[key], 'document provenance mismatch')
check(not RESIDUAL.search(doc['text']) and '¶' not in doc['text'], 'management syntax survived')
exact[digest].append(doc['id'])
works.add(rid)
totals['clean_chars'] += len(text)
dynasties[row['dynasty']] += len(text)
if (i+1) % 10000 == 0:
print(f'verified {i+1}/{len(inventory)}', flush=True)
check(next(docs, None) is None, 'extra documents')
check(set(review) == {r['id'] for r in file_rows if r['status'] == 'review'}, 'review coverage')
check(set(edits) == {r['id'] for r in file_rows if r['removed']}, 'edits coverage')
body_works = {r['id'] for r in inventory}
expected_excluded = {}
for row in catalog.values():
reasons = []
if row['dynasty'] in {'清', '淸'}:
reasons.append('qing_user_excluded')
totals['qing_catalog_works'] += 1
if row['status'] == 'complete':
totals['qing_complete_works'] += 1
if row['status'] != 'complete':
reasons.append('download_not_complete')
totals['incomplete_catalog_works'] += 1
elif row['dynasty'] not in {'清', '淸'} and row['id'] not in body_works:
reasons.append('no_body_files')
totals['no_body_works'] += 1
if reasons:
expected_excluded[row['id']] = reasons
exclusions = list(records(output / 'excluded.jsonl'))
check(len(exclusions) == len(expected_excluded), 'exclusion coverage')
check({r['id']: r['reasons'] for r in exclusions} == expected_excluded, 'exclusion reasons')
totals['excluded_catalog_works'] = len(exclusions)
expected_duplicates = [dict(sha256=h, ids=ids) for h, ids in exact.items() if len(ids) > 1]
check(list(records(output / 'duplicates.jsonl')) == expected_duplicates, 'duplicate group mismatch')
totals.update(included_works=len(works), review_works=len(review_works), duplicate_groups=len(expected_duplicates))
report = json.loads((output / 'report.json').read_text())
check(totals == Counter(report['totals']), 'total mismatch')
check(removed == Counter(report['removed']), 'removal count mismatch')
check(dynasties == Counter(report['chars_by_dynasty']), 'dynasty character counts mismatch')
result = dict(verified=True, verified_files=len(inventory), included_files=totals['included_files'],
clean_chars=totals['clean_chars'], documents_sha256=file_hash(output / 'documents.jsonl'),
checks=['raw hashes', 'full serialized text and provenance', 'Qing exclusion and file coverage',
'management marker absence', 'review and empty statuses', 'totals and duplicates'])
dump(output / 'verification.json', result)
print(json.dumps(result, ensure_ascii=False), flush=True)
return result
if __name__ == '__main__':
p = argparse.ArgumentParser(description=__doc__)
p.add_argument('output', type=Path)
a = p.parse_args()
verify(a.output)
|