File size: 5,417 Bytes
c8beaf3 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 | #!/usr/bin/env python3
"""Read-only verification of fixed Hugging Face dependency objects."""
from __future__ import annotations
import argparse
import csv
import hashlib
import json
import urllib.parse
import urllib.request
from pathlib import Path
def read_tsv(path: Path) -> list[dict[str, str]]:
with path.open("r", encoding="utf-8", newline="") as handle:
return list(csv.DictReader(handle, delimiter="\t"))
def fetch_tree(repo: str, revision: str) -> dict[str, dict[str, object]]:
encoded = urllib.parse.quote(repo, safe="/")
rev = urllib.parse.quote(revision, safe="")
url = f"https://huggingface.co/api/datasets/{encoded}/tree/{rev}?recursive=true&expand=true"
with urllib.request.urlopen(url, timeout=60) as response:
rows = json.load(response)
return {str(row["path"]): row for row in rows if row.get("type") == "file"}
def full_remote_sha(repo: str, revision: str, path: str) -> str:
url = (
f"https://huggingface.co/datasets/{repo}/resolve/{revision}/"
f"{urllib.parse.quote(path)}?download=true"
)
digest = hashlib.sha256()
with urllib.request.urlopen(url, timeout=60) as response:
while True:
chunk = response.read(8 * 1024 * 1024)
if not chunk:
break
digest.update(chunk)
return digest.hexdigest()
def range_remote_sha(repo: str, revision: str, path: str, start: int, end: int) -> str:
url = (
f"https://huggingface.co/datasets/{repo}/resolve/{revision}/"
f"{urllib.parse.quote(path)}?download=true"
)
request = urllib.request.Request(url, headers={"Range": f"bytes={start}-{end}"})
with urllib.request.urlopen(request, timeout=60) as response:
if response.status != 206:
raise ValueError(f"server did not honor byte range: HTTP {response.status}")
payload = response.read()
if len(payload) != end - start + 1:
raise ValueError(f"range length mismatch: {len(payload)} != {end - start + 1}")
return hashlib.sha256(payload).hexdigest()
def main() -> int:
default = Path(__file__).resolve().parents[1] / "dependencies" / "DEPENDENCIES.tsv"
parser = argparse.ArgumentParser()
parser.add_argument("--dependencies", type=Path, default=default)
parser.add_argument("--output", type=Path)
args = parser.parse_args()
rows = read_tsv(args.dependencies)
trees: dict[tuple[str, str], dict[str, dict[str, object]]] = {}
results: list[dict[str, object]] = []
failed = 0
for row in rows:
key = (row["source_repo"], row["source_revision"])
if key not in trees:
trees[key] = fetch_tree(*key)
remote = trees[key].get(row["repository_path"])
status = "PASS"
method = "lfs_oid"
detail = ""
if remote is None:
status, detail = "FAIL", "path missing"
elif int(remote.get("size", -1)) != int(row["bytes"]):
status, detail = "FAIL", f"size {remote.get('size')} != {row['bytes']}"
elif remote.get("lfs"):
oid = str(remote["lfs"].get("oid", ""))
if oid != row["sha256"]:
status, detail = "FAIL", f"LFS oid {oid} != expected SHA-256"
elif int(row["bytes"]) <= 20_000_000:
method = "full_download_sha256"
actual = full_remote_sha(row["source_repo"], row["source_revision"], row["repository_path"])
if actual != row["sha256"]:
status, detail = "FAIL", f"SHA-256 {actual} != expected"
elif row.get("range_bytes") and row.get("range_first_sha256") and row.get("range_last_sha256"):
method = "fixed_range_sha256"
width = int(row["range_bytes"])
total = int(row["bytes"])
try:
first = range_remote_sha(
row["source_repo"], row["source_revision"], row["repository_path"], 0, width - 1
)
last = range_remote_sha(
row["source_repo"],
row["source_revision"],
row["repository_path"],
total - width,
total - 1,
)
if first != row["range_first_sha256"] or last != row["range_last_sha256"]:
status, detail = "FAIL", "fixed-range SHA-256 mismatch"
except Exception as exc:
status, detail = "UNRESOLVED", str(exc)
else:
method = "metadata_only"
status, detail = "UNRESOLVED", "no LFS oid and no fixed-range receipt"
if status == "FAIL":
failed += 1
results.append(
{
"component": row["component"],
"path": row["repository_path"],
"status": status,
"method": method,
"detail": detail,
}
)
receipt = {
"status": "PASS" if failed == 0 else "FAIL",
"checked": len(results),
"failed": failed,
"results": results,
"service_commitment": "none; this is a manually runnable verification capability",
}
text = json.dumps(receipt, indent=2, sort_keys=True) + "\n"
if args.output:
args.output.write_text(text, encoding="utf-8")
print(text, end="")
return 0 if failed == 0 else 1
if __name__ == "__main__":
raise SystemExit(main())
|