File size: 10,176 Bytes
2ccb279 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 |
# from huggingface_hub import list_datasets, DatasetCard
# import re
# import pandas as pd
# import os
# import time
# import random
# from concurrent.futures import ThreadPoolExecutor, as_completed
# from requests.exceptions import HTTPError
# # ---------- Retry helper ----------
# def retry_load_card(dataset_id, retries=5, base_wait=60):
# """
# Try to load a dataset card with retries if 429 (rate limit) occurs.
# Uses Retry-After header if available, otherwise exponential backoff.
# """
# for attempt in range(retries):
# try:
# return DatasetCard.load(dataset_id)
# except HTTPError as e:
# if e.response is not None and e.response.status_code == 429:
# wait_time = e.response.headers.get("Retry-After")
# if wait_time is not None:
# wait_time = int(wait_time)
# else:
# wait_time = base_wait * (2 ** attempt) + random.randint(0, 10)
# print(f"[429] Rate limit hit for {dataset_id}. Sleeping {wait_time}s (attempt {attempt+1}/{retries})...")
# time.sleep(wait_time)
# continue
# else:
# raise # don't retry for other HTTP errors
# except Exception as e:
# print(f"[ERROR] {dataset_id}: {e}")
# raise
# raise RuntimeError(f"Failed to load {dataset_id} after {retries} retries.")
# # ---------- Heuristic functions with reasons ----------
# def check_card_quality(card_text, metadata, dataset_url):
# reasons = []
# length = len(card_text)
# word_count = len(card_text.split())
# if metadata is None or len(metadata) == 0:
# print(length, word_count, dataset_url)
# if length < 200:
# reasons.append("No metadata and no description")
# return "minimal", reasons, word_count
# else:
# reasons.append("No metadata but has description")
# return "minimal", reasons, word_count
# else:
# if length < 200:
# reasons.append(f"Short description (char count={length}, words={word_count})")
# return "minimal", reasons, word_count
# else:
# return "rich", reasons, word_count
# # ---------- Worker function for one dataset ----------
# def process_dataset(ds, save_dir):
# try:
# card = retry_load_card(ds.id)
# card_text = card.text or ""
# metadata = card.data.to_dict() if card.data else {}
# dataset_url = f"https://huggingface.co/datasets/{ds.id}"
# # Save README locally
# readme_path = os.path.join(save_dir, f"{ds.id.replace('/', '__')}_README.md")
# with open(readme_path, "w", encoding="utf-8") as f:
# f.write(card_text)
# category, reasons, word_count = check_card_quality(card_text, metadata, dataset_url)
# row = {
# "dataset_id": ds.id,
# "dataset_url": dataset_url,
# "downloads": getattr(ds, "downloads", None),
# "reason": "; ".join(reasons),
# "readme_path": readme_path,
# "word_count": word_count,
# "category": category,
# }
# return row
# except Exception as e:
# return {
# "dataset_id": ds.id,
# "dataset_url": f"https://huggingface.co/datasets/{ds.id}",
# "downloads": getattr(ds, "downloads", None),
# "reason": f"Failed to load card",
# "readme_path": None,
# "word_count": 0,
# "category": "minimal",
# }
# # ---------- Main ----------
# def collect_dataset_ids(limit=1000, save_dir="dataset_readmes", max_workers=16):
# minimal_results = []
# rich_results = []
# os.makedirs(save_dir, exist_ok=True)
# print(f"Fetching up to {limit} datasets (sorted by downloads)...")
# datasets = list_datasets()
# with ThreadPoolExecutor(max_workers=max_workers) as executor:
# futures = [executor.submit(process_dataset, ds, save_dir) for ds in datasets]
# for i, f in enumerate(as_completed(futures), 1):
# row = f.result()
# if row["category"] == "minimal":
# minimal_results.append(row)
# else:
# rich_results.append(row)
# return minimal_results, rich_results
# if __name__ == "__main__":
# minimal, rich = collect_dataset_ids(limit=1000, max_workers=16)
# # Save separate CSV files
# if minimal:
# pd.DataFrame(minimal).to_csv("all_minimal_dataset_cards.csv", index=False)
# if rich:
# pd.DataFrame(rich).to_csv("all_rich_dataset_cards.csv", index=False)
# print("\nSaved results to:")
# if minimal:
# print(" - minimal_dataset_cards.csv")
# if rich:
# print(" - rich_dataset_cards.csv")
# print(" - README files in ./dataset_readmes/")
# print("\nSummary:")
# print(f"Minimal: {len(minimal)}")
# print(f"Rich: {len(rich)}")
from huggingface_hub import list_datasets, DatasetCard
import re
import pandas as pd
import os
import time
import random
from concurrent.futures import ThreadPoolExecutor, as_completed
from requests.exceptions import HTTPError
# # ---------- Retry helper ----------
# def retry_load_card(dataset_id, retries=5, base_wait=60):
# for attempt in range(retries):
# try:
# return DatasetCard.load(dataset_id)
# except HTTPError as e:
# if e.response is not None and e.response.status_code == 429:
# wait_time = e.response.headers.get("Retry-After")
# if wait_time is not None:
# wait_time = int(wait_time)
# else:
# wait_time = base_wait * (2 ** attempt) + random.randint(0, 10)
# print(f"[429] Rate limit hit for {dataset_id}. Sleeping {wait_time}s (attempt {attempt+1}/{retries})...")
# time.sleep(wait_time)
# continue
# else:
# raise
# except Exception as e:
# print(f"[ERROR] {dataset_id}: {e}")
# raise
# raise RuntimeError(f"Failed to load {dataset_id} after {retries} retries.")
# ---------- Heuristic functions with reasons ----------
def check_card_quality(card_text, metadata, dataset_url):
reasons = []
length = len(card_text)
word_count = len(card_text.split())
if metadata is None or len(metadata) == 0:
print(length, word_count, dataset_url)
if length < 200:
reasons.append("No metadata and no description")
return "minimal", reasons, word_count
else:
reasons.append("No metadata but has description")
return "minimal", reasons, word_count
else:
if length < 200:
reasons.append(f"Short description (char count={length}, words={word_count})")
return "minimal", reasons, word_count
else:
return "rich", reasons, word_count
# ---------- Worker function for one dataset ----------
def process_dataset(ds, save_dir):
try:
card = DatasetCard.load(ds.id)
card_text = card.text or ""
metadata = card.data.to_dict() if card.data else {}
dataset_url = f"https://huggingface.co/datasets/{ds.id}"
# Save README locally
readme_path = os.path.join(save_dir, f"{ds.id.replace('/', '__')}_README.md")
with open(readme_path, "w", encoding="utf-8") as f:
f.write(card_text)
category, reasons, word_count = check_card_quality(card_text, metadata, dataset_url)
row = {
"dataset_id": ds.id,
"dataset_url": dataset_url,
"downloads": getattr(ds, "downloads", None),
"author": metadata.get("author", None),
"license": metadata.get("license", None),
"tags": ", ".join(metadata.get("tags", [])) if metadata.get("tags") else None,
"task_categories": ", ".join(metadata.get("task_categories", [])) if metadata.get("task_categories") else None,
"last_modified": getattr(ds, "lastModified", None),
"reason": "; ".join(reasons),
"readme_path": readme_path,
"word_count": word_count,
"category": category,
}
return row
except Exception as e:
return {
"dataset_id": ds.id,
"dataset_url": f"https://huggingface.co/datasets/{ds.id}",
"downloads": getattr(ds, "downloads", None),
"author": None,
"license": None,
"tags": None,
"task_categories": None,
"last_modified": None,
"reason": "Failed to load card",
"readme_path": None,
"word_count": 0,
"category": "minimal",
}
# ---------- Main ----------
def collect_dataset_ids(save_dir="dataset_readmes", max_workers=16):
minimal_results = []
rich_results = []
os.makedirs(save_dir, exist_ok=True)
datasets = list_datasets()
with ThreadPoolExecutor(max_workers=max_workers) as executor:
futures = [executor.submit(process_dataset, ds, save_dir) for ds in datasets]
for i, f in enumerate(as_completed(futures), 1):
row = f.result()
if row["category"] == "minimal":
minimal_results.append(row)
else:
rich_results.append(row)
return minimal_results, rich_results
if __name__ == "__main__":
minimal, rich = collect_dataset_ids(limit=100, max_workers=16)
# Save separate CSV files
if minimal:
pd.DataFrame(minimal).to_csv("all_minimal_dataset_cards.csv", index=False)
if rich:
pd.DataFrame(rich).to_csv("all_rich_dataset_cards.csv", index=False)
print("\nSaved results to:")
if minimal:
print(" - minimal_dataset_cards.csv")
if rich:
print(" - rich_dataset_cards.csv")
print(" - README files in ./dataset_readmes/")
print("\nSummary:")
print(f"Minimal: {len(minimal)}")
print(f"Rich: {len(rich)}")
|