ancient-scripts-datasets / verify_hf.py
Alvin
Add complete dataset: all sources, metadata, scripts, docs, and phylo enrichment
26786e3
import sys, os
sys.stdout.reconfigure(encoding='utf-8')
sys.stderr.reconfigure(encoding='utf-8')
from huggingface_hub import HfApi
from pathlib import Path
api = HfApi()
repo = "Nacryos/ancient-scripts-datasets"
local_base = Path("C:/Users/alvin/hf-ancient-scripts")
checks = [
("docs", 20),
("scripts", 37),
("cognate_pipeline", 91),
("data/validation", 58),
("data/training/validation", 53),
("data/training/audit_trails", 27),
("data/training/language_profiles", 1110),
("data/training/raw", 14),
("data/cited_sources", 5),
]
all_pass = True
for path, expected in checks:
local_dir = local_base / path
# Count local files (exclude __pycache__, .pyc, .git)
local_count = 0
local_names = set()
for root, dirs, files in os.walk(local_dir):
dirs[:] = [d for d in dirs if d not in ('__pycache__', '.git')]
for f in files:
if not f.endswith('.pyc'):
local_count += 1
rel = os.path.relpath(os.path.join(root, f), local_dir).replace("\\", "/")
local_names.add(rel)
# Count HF files
try:
hf_items = list(api.list_repo_tree(repo, path_in_repo=path, repo_type="dataset", recursive=True))
hf_names = set()
for item in hf_items:
if hasattr(item, 'size'):
rel = item.path.replace(path + "/", "", 1)
hf_names.add(rel)
hf_count = len(hf_names)
except Exception as e:
hf_count = 0
hf_names = set()
print(f"ERROR listing {path}: {e}")
missing = local_names - hf_names
extra = hf_names - local_names
status = "PASS" if local_count == hf_count and not missing else "FAIL"
if status == "FAIL":
all_pass = False
print(f"{path:40s} local={local_count:5d} hf={hf_count:5d} expected={expected:5d} {status}")
if missing and len(missing) <= 10:
for f in sorted(missing):
print(f" MISSING on HF: {f}")
elif missing:
print(f" MISSING on HF: {len(missing)} files (too many to list)")
if extra and len(extra) <= 10:
for f in sorted(extra):
print(f" EXTRA on HF: {f}")
elif extra:
print(f" EXTRA on HF: {len(extra)} files")
# Verify Source column in languages.tsv metadata
print(f"\n{'='*60}")
print("METADATA PROVENANCE CHECK:")
from huggingface_hub import hf_hub_download
lang_path = hf_hub_download(repo_id=repo, filename="data/training/metadata/languages.tsv", repo_type="dataset")
with open(lang_path, "r", encoding="utf-8") as f:
lines = f.readlines()
header = lines[0].strip().split("\t")
print(f" languages.tsv columns: {header}")
print(f" languages.tsv entries: {len(lines)-1}")
# Check that Sources column exists
if "Sources" in header:
src_idx = header.index("Sources")
sources_present = sum(1 for line in lines[1:] if line.strip().split("\t")[src_idx].strip())
print(f" Entries with Sources: {sources_present}/{len(lines)-1}")
if sources_present > 0:
print(f" STATUS: PASS")
else:
print(f" STATUS: FAIL - no source data")
else:
print(f" STATUS: FAIL - no Sources column")
# Also check Source column in a sample lexicon
print(f"\n{'='*60}")
print("LEXICON PROVENANCE CHECK (Source column in TSVs):")
for iso in ["hit", "ave", "ine-pro", "xlc", "txb"]:
lex_path = hf_hub_download(repo_id=repo, filename=f"data/training/lexicons/{iso}.tsv", repo_type="dataset")
with open(lex_path, "r", encoding="utf-8") as f:
lex_lines = f.readlines()
lex_header = lex_lines[0].strip().split("\t")
if "Source" in lex_header:
src_idx = lex_header.index("Source")
sources = set()
for line in lex_lines[1:]:
parts = line.strip().split("\t")
if len(parts) > src_idx:
sources.add(parts[src_idx])
print(f" {iso:10s} entries={len(lex_lines)-1:5d} sources={sources}")
else:
print(f" {iso:10s} FAIL - no Source column")
print(f"\n{'='*60}")
if all_pass:
print("OVERALL: PASS")
else:
print("OVERALL: FAIL")