Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
Zhangir Azerbayev
commited on
Commit
·
9bdf6a3
1
Parent(s):
f293502
added some fiels
Browse files- fetch_arxiv.py +272 -0
- fetch_wiki.py +43 -29
- proof-pile.py +12 -1
- test.py +3 -2
- utils.py +46 -0
fetch_arxiv.py
ADDED
|
@@ -0,0 +1,272 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import sys
|
| 3 |
+
from pathlib import Path
|
| 4 |
+
import datetime
|
| 5 |
+
|
| 6 |
+
import tarfile
|
| 7 |
+
import xml.etree.ElementTree as ET
|
| 8 |
+
from tqdm import tqdm
|
| 9 |
+
import re
|
| 10 |
+
from itertools import chain, islice
|
| 11 |
+
import requests
|
| 12 |
+
import time
|
| 13 |
+
|
| 14 |
+
import shutil
|
| 15 |
+
|
| 16 |
+
import arxiv
|
| 17 |
+
|
| 18 |
+
import langdetect
|
| 19 |
+
from langdetect import detect
|
| 20 |
+
|
| 21 |
+
from utils import Loader as Loader
|
| 22 |
+
from utils import make_archive
|
| 23 |
+
|
| 24 |
+
def batch_loader(seq, size):
|
| 25 |
+
"""
|
| 26 |
+
Iterator that takes in a list `seq` and returns
|
| 27 |
+
chunks of size `size`
|
| 28 |
+
"""
|
| 29 |
+
return [seq[pos:pos + size] for pos in range(0, len(seq), size)]
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def _delete_files_except_pattern(path, pattern, transform = lambda x: None, verbose=False):
|
| 33 |
+
"""
|
| 34 |
+
recursively
|
| 35 |
+
"""
|
| 36 |
+
for f in os.listdir(path):
|
| 37 |
+
f_path = os.path.join(path, f)
|
| 38 |
+
if verbose:
|
| 39 |
+
print(f_path)
|
| 40 |
+
if os.path.isfile(f_path):
|
| 41 |
+
if not re.search(pattern, f):
|
| 42 |
+
os.chmod(f_path, 0o755)
|
| 43 |
+
os.remove(f_path)
|
| 44 |
+
else:
|
| 45 |
+
transform(f_path)
|
| 46 |
+
elif os.path.isdir(f_path):
|
| 47 |
+
try:
|
| 48 |
+
print(f_path)
|
| 49 |
+
except UnicodeEncodeError:
|
| 50 |
+
new_path = f_path.encode("utf-8", 'replace').decode()
|
| 51 |
+
os.system(f"mv \"{f_path}\" \"{new_path}\"")
|
| 52 |
+
f_path = new_path
|
| 53 |
+
|
| 54 |
+
_delete_files_except_pattern(f_path, pattern, transform=transform, verbose=verbose)
|
| 55 |
+
|
| 56 |
+
def _download_with_progress_bar(url):
|
| 57 |
+
response = requests.get(url, stream=True)
|
| 58 |
+
total_size_in_bytes = int(response.headers.get("content-length", 0))
|
| 59 |
+
block_size = 1024 # 1 Kibibyte
|
| 60 |
+
progress_bar = tqdm(total=total_size_in_bytes, unit="iB", unit_scale=True)
|
| 61 |
+
to_return = bytearray()
|
| 62 |
+
for data in response.iter_content(block_size):
|
| 63 |
+
progress_bar.update(len(data))
|
| 64 |
+
to_return += data
|
| 65 |
+
progress_bar.close()
|
| 66 |
+
if total_size_in_bytes != 0 and progress_bar.n != total_size_in_bytes:
|
| 67 |
+
raise AssertionError("ERROR, something went wrong")
|
| 68 |
+
|
| 69 |
+
return to_return
|
| 70 |
+
|
| 71 |
+
def get_math_ids(resumption_token="init"):
|
| 72 |
+
with Loader(f"fetching metadata shard {resumption_token}..."):
|
| 73 |
+
if resumption_token=="init":
|
| 74 |
+
resp = requests.get("https://export.arxiv.org/oai2?verb=ListIdentifiers&set=math&metadataPrefix=oai_dc")
|
| 75 |
+
else:
|
| 76 |
+
time.sleep(5)
|
| 77 |
+
resp = requests.get(f"https://export.arxiv.org/oai2?verb=ListIdentifiers&resumptionToken={resumption_token}")
|
| 78 |
+
|
| 79 |
+
root = ET.fromstring(resp.content.decode("utf-8"))
|
| 80 |
+
articles = root[2]
|
| 81 |
+
|
| 82 |
+
math_ids = {}
|
| 83 |
+
for article in articles:
|
| 84 |
+
if article.tag == "{http://www.openarchives.org/OAI/2.0/}resumptionToken":
|
| 85 |
+
if article.text:
|
| 86 |
+
return math_ids | get_math_ids(resumption_token=article.text)
|
| 87 |
+
else:
|
| 88 |
+
return math_ids
|
| 89 |
+
|
| 90 |
+
db_id = article[0].text
|
| 91 |
+
eyed = db_id[db_id.rindex(":")+1:]
|
| 92 |
+
math_ids[eyed] = True
|
| 93 |
+
|
| 94 |
+
def clean_tex_file(path):
|
| 95 |
+
with open(path, encoding="utf-8") as f:
|
| 96 |
+
try:
|
| 97 |
+
src = f.read()
|
| 98 |
+
except (UnicodeDecodeError, UnicodeError):
|
| 99 |
+
print(f"Decoding error at {path} with utf-8. Trying latin-1")
|
| 100 |
+
try:
|
| 101 |
+
with open(path, encoding="latin-1") as fle:
|
| 102 |
+
src = fle.read()
|
| 103 |
+
#print("latin-1 successful\n")
|
| 104 |
+
except (UnicodeDecodeError, UnicodeError):
|
| 105 |
+
#print(f"Decoding error at {path} with latin-1. Trying utf-16")
|
| 106 |
+
try:
|
| 107 |
+
with open(path, encoding="utf-16") as fl:
|
| 108 |
+
src = fl.read()
|
| 109 |
+
#print("utf-16 successful\n")
|
| 110 |
+
except (UnicodeDecodeError, UnicodeError):
|
| 111 |
+
#print(f"Decoding error at {path} with utf-16. Trying utf-32")
|
| 112 |
+
try:
|
| 113 |
+
with open(path, encoding="utf-32") as f:
|
| 114 |
+
src = f.read()
|
| 115 |
+
except (UnicodeDecodeError, UnicodeError):
|
| 116 |
+
print(f"Decoding error at {path} with all of utf-8, 16, 32 and latin-1. Deleting this file")
|
| 117 |
+
print("This issue should only occur with a handful of quite old files. Continuing...\n")
|
| 118 |
+
return
|
| 119 |
+
|
| 120 |
+
end = re.search(r"\\end\{document\}", src)
|
| 121 |
+
if end:
|
| 122 |
+
src = src[:end.span()[1]]
|
| 123 |
+
|
| 124 |
+
bib = re.search(r"\\Refs|\\begin\{thebibliography\}", src)
|
| 125 |
+
if bib:
|
| 126 |
+
src = src[:bib.span()[0]]
|
| 127 |
+
|
| 128 |
+
os.chmod(path, 0o755)
|
| 129 |
+
with open(path, "w", encoding="utf-8") as f:
|
| 130 |
+
f.write(src)
|
| 131 |
+
|
| 132 |
+
def clean_tex_file_some_more(path):
|
| 133 |
+
with open(path) as f:
|
| 134 |
+
text = f.read()
|
| 135 |
+
|
| 136 |
+
text = re.sub(r"(?<!\\)%.*", "", text)
|
| 137 |
+
|
| 138 |
+
match_obj = re.search(r"\\begin\{document\}", text)
|
| 139 |
+
if match_obj:
|
| 140 |
+
text = text[match_obj.span()[0]:]
|
| 141 |
+
|
| 142 |
+
match_obj = re.search(r"\\begin\{references\}", text)
|
| 143 |
+
if match_obj:
|
| 144 |
+
text = text[:match_obj.span()[0]]
|
| 145 |
+
|
| 146 |
+
text = text.strip()
|
| 147 |
+
|
| 148 |
+
os.remove(path)
|
| 149 |
+
if len(text)>280:
|
| 150 |
+
try:
|
| 151 |
+
print(path)
|
| 152 |
+
except UnicodeEncodeError:
|
| 153 |
+
path = path.encode('utf-8', 'replace').decode()
|
| 154 |
+
|
| 155 |
+
try:
|
| 156 |
+
lang = detect(text)
|
| 157 |
+
except langdetect.lang_detect_exception.LangDetectException:
|
| 158 |
+
# no linguistic features to analyze, delete
|
| 159 |
+
return
|
| 160 |
+
|
| 161 |
+
if lang=="en":
|
| 162 |
+
with open(path, "w") as f:
|
| 163 |
+
f.write(text)
|
| 164 |
+
else:
|
| 165 |
+
print("HIT NONENGLISH ARTICLE")
|
| 166 |
+
|
| 167 |
+
def process_tarball_old_scheme(tarball_name, save_dir):
|
| 168 |
+
tarball_path = os.path.join(save_dir, tarball_name)
|
| 169 |
+
os.system("tar -xf " + tarball_path + " -C " + save_dir)
|
| 170 |
+
|
| 171 |
+
last_ = tarball_name.rfind("_")
|
| 172 |
+
second_last_ = tarball_name.rfind("_", 0, last_)
|
| 173 |
+
subdir = tarball_name[second_last_+1:last_]
|
| 174 |
+
|
| 175 |
+
subpath = os.path.join(save_dir, subdir)
|
| 176 |
+
zipped_names = os.listdir(subpath)
|
| 177 |
+
|
| 178 |
+
for zipped_name in zipped_names:
|
| 179 |
+
if zipped_name[-len(".gz"):]==".gz":
|
| 180 |
+
zipped_path = os.path.join(subpath, zipped_name)
|
| 181 |
+
if re.match(r"math", zipped_name):
|
| 182 |
+
eyed = zipped_name[:-len(".gz")]
|
| 183 |
+
if tarfile.is_tarfile(zipped_path):
|
| 184 |
+
article_dir = os.path.join(subpath, eyed)
|
| 185 |
+
Path(article_dir).mkdir()
|
| 186 |
+
os.system("tar -xzf " + zipped_path + " -C " + article_dir)
|
| 187 |
+
os.remove(zipped_path)
|
| 188 |
+
else:
|
| 189 |
+
os.system("gzip -d " + zipped_path)
|
| 190 |
+
unzipped_path = os.path.join(subpath, eyed)
|
| 191 |
+
os.rename(unzipped_path, unzipped_path + ".tex")
|
| 192 |
+
else:
|
| 193 |
+
os.remove(zipped_path)
|
| 194 |
+
|
| 195 |
+
_delete_files_except_pattern(subpath, r".*\.tex", transform=clean_tex_file)
|
| 196 |
+
os.remove(tarball_path)
|
| 197 |
+
|
| 198 |
+
def process_tarball(tarball_name, save_dir, math_ids):
|
| 199 |
+
tarball_path = os.path.join(save_dir, tarball_name)
|
| 200 |
+
untar_cmd = "tar -xf " + tarball_path + " -C " + save_dir
|
| 201 |
+
os.system(untar_cmd)
|
| 202 |
+
|
| 203 |
+
last_ = tarball_name.rfind("_")
|
| 204 |
+
second_last_ = tarball_name.rfind("_", 0, last_)
|
| 205 |
+
subdir = tarball_name[second_last_+1:last_]
|
| 206 |
+
|
| 207 |
+
subpath = os.path.join(save_dir, subdir)
|
| 208 |
+
listdir = os.listdir(subpath)
|
| 209 |
+
|
| 210 |
+
ids = [x[:-3] for x in listdir if x[-3:]==".gz"]
|
| 211 |
+
|
| 212 |
+
for eyed in ids:
|
| 213 |
+
if eyed in math_ids:
|
| 214 |
+
zipped_path = os.path.join(subpath, eyed + ".gz")
|
| 215 |
+
|
| 216 |
+
if tarfile.is_tarfile(zipped_path):
|
| 217 |
+
article_dir = os.path.join(subpath, eyed)
|
| 218 |
+
Path(article_dir).mkdir()
|
| 219 |
+
os.system("tar -xzf " + zipped_path + " -C " + article_dir)
|
| 220 |
+
os.remove(zipped_path)
|
| 221 |
+
else:
|
| 222 |
+
os.system("gzip -d " + zipped_path)
|
| 223 |
+
unzipped_path = os.path.join(subpath, eyed)
|
| 224 |
+
os.rename(unzipped_path, unzipped_path + ".tex")
|
| 225 |
+
|
| 226 |
+
_delete_files_except_pattern(subpath, r".*\.tex", transform=clean_tex_file)
|
| 227 |
+
os.remove(tarball_path)
|
| 228 |
+
|
| 229 |
+
def main():
|
| 230 |
+
"""
|
| 231 |
+
Warning: this code is *extremely* brittle
|
| 232 |
+
"""
|
| 233 |
+
math_ids = get_math_ids()
|
| 234 |
+
|
| 235 |
+
save_dir = "arxiv_1"
|
| 236 |
+
Path(save_dir).mkdir(exist_ok=True)
|
| 237 |
+
manifest_path = os.path.join(save_dir, "manifest.xml")
|
| 238 |
+
|
| 239 |
+
os.system(f"s3cmd get s3://arxiv/src/arXiv_src_manifest.xml --requester-pays {manifest_path}")
|
| 240 |
+
|
| 241 |
+
tree = ET.parse(manifest_path)
|
| 242 |
+
root = tree.getroot()
|
| 243 |
+
|
| 244 |
+
shards_and_dates = []
|
| 245 |
+
for child in root:
|
| 246 |
+
if child.tag == "file":
|
| 247 |
+
shard = child[1].text # the index of filename
|
| 248 |
+
yymm = child[9].text # the index of yymm
|
| 249 |
+
shards_and_dates.append((shard, yymm))
|
| 250 |
+
|
| 251 |
+
format_cutoff = datetime.datetime(2007, 3, 1) # arXiv switches from old to new format
|
| 252 |
+
for shard, yymm in tqdm(shards_and_dates):
|
| 253 |
+
print("SHARD: ", shard)
|
| 254 |
+
os.system(f"s3cmd get s3://arxiv/" + shard + \
|
| 255 |
+
" --requester-pays " + save_dir)
|
| 256 |
+
tarball_name=shard[shard.rindex("/")+1:]
|
| 257 |
+
|
| 258 |
+
# nb this code will stop working in 2051 ;)
|
| 259 |
+
year = int("19" + yymm[:2]) if int(yymm[:2])>50 else int("20"+yymm[:2])
|
| 260 |
+
if datetime.datetime(year, int(yymm[2:]), 1)<=format_cutoff:
|
| 261 |
+
process_tarball_old_scheme(tarball_name, save_dir)
|
| 262 |
+
else:
|
| 263 |
+
process_tarball(tarball_name, save_dir, math_ids)
|
| 264 |
+
|
| 265 |
+
os.remove(manifest_path)
|
| 266 |
+
|
| 267 |
+
if __name__=="__main__":
|
| 268 |
+
#main()
|
| 269 |
+
#_delete_files_except_pattern("arxiv_1", r".*\.tex$", transform=clean_tex_file_some_more)
|
| 270 |
+
for f in tqdm(os.listdir("arxiv")):
|
| 271 |
+
f_path = os.path.join("arxiv", f)
|
| 272 |
+
make_archive(f_path)
|
fetch_wiki.py
CHANGED
|
@@ -6,8 +6,14 @@ import re
|
|
| 6 |
import pypandoc
|
| 7 |
import json
|
| 8 |
from pathlib import Path
|
|
|
|
| 9 |
from fetch_books_and_formal import _download_with_progress_bar
|
| 10 |
-
from
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
def page_titles_of_category(cat_page):
|
| 13 |
"""
|
|
@@ -43,8 +49,11 @@ PROOFWIKI_URL = (
|
|
| 43 |
"https://zenodo.org/record/4902289/files/naturalproofs_proofwiki.json?download=1"
|
| 44 |
)
|
| 45 |
def proofwiki(testing=False):
|
|
|
|
| 46 |
save_dir = "wiki/proofwiki"
|
|
|
|
| 47 |
Path(save_dir).mkdir(parents=True, exist_ok=True)
|
|
|
|
| 48 |
|
| 49 |
if testing:
|
| 50 |
with open("naturalproofs/proofwiki.json") as f:
|
|
@@ -55,48 +64,53 @@ def proofwiki(testing=False):
|
|
| 55 |
struct = json.loads(resp.decode("utf-8"))
|
| 56 |
print("DONE DOWNLOADING PROOFWIKI")
|
| 57 |
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
thm_string += (
|
| 65 |
-
"Tags: " + ", ".join(thm["categories"]).replace("/", ": ") + "\n\n"
|
| 66 |
-
)
|
| 67 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
thm_string += (
|
| 69 |
-
"\\begin{
|
| 70 |
-
+ "\n".join(
|
| 71 |
-
+ "\n\\end{
|
| 72 |
)
|
| 73 |
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
)
|
| 80 |
-
|
| 81 |
-
thms_list.append(thm_string.strip())
|
| 82 |
-
|
| 83 |
-
with open(os.path.join(save_dir, f"shard_{i}.txt"), "w") as f:
|
| 84 |
-
f.write("<|endoftext|>\n".join(thms_list))
|
| 85 |
|
| 86 |
defn_strings = []
|
| 87 |
for defn in struct["dataset"]["definitions"]:
|
| 88 |
if defn["contents"]:
|
| 89 |
-
|
| 90 |
"\\begin{definition}["
|
| 91 |
+ defn["label"]
|
| 92 |
+ "]\n"
|
| 93 |
+ "\n".join(defn["contents"])
|
| 94 |
+ "\n\\end{definition}").strip()
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
|
| 98 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 99 |
|
| 100 |
if __name__=="__main__":
|
| 101 |
#wikipedia()
|
| 102 |
proofwiki()
|
|
|
|
|
|
|
|
|
| 6 |
import pypandoc
|
| 7 |
import json
|
| 8 |
from pathlib import Path
|
| 9 |
+
|
| 10 |
from fetch_books_and_formal import _download_with_progress_bar
|
| 11 |
+
from utils import make_archive
|
| 12 |
+
|
| 13 |
+
import random
|
| 14 |
+
|
| 15 |
+
random.seed(20)
|
| 16 |
+
|
| 17 |
|
| 18 |
def page_titles_of_category(cat_page):
|
| 19 |
"""
|
|
|
|
| 49 |
"https://zenodo.org/record/4902289/files/naturalproofs_proofwiki.json?download=1"
|
| 50 |
)
|
| 51 |
def proofwiki(testing=False):
|
| 52 |
+
VAL_RATE = 0.025
|
| 53 |
save_dir = "wiki/proofwiki"
|
| 54 |
+
val_dir = "wiki/proofwiki_val"
|
| 55 |
Path(save_dir).mkdir(parents=True, exist_ok=True)
|
| 56 |
+
Path(val_dir).mkdir(parents=True, exist_ok=True)
|
| 57 |
|
| 58 |
if testing:
|
| 59 |
with open("naturalproofs/proofwiki.json") as f:
|
|
|
|
| 64 |
struct = json.loads(resp.decode("utf-8"))
|
| 65 |
print("DONE DOWNLOADING PROOFWIKI")
|
| 66 |
|
| 67 |
+
for i, thm in enumerate(struct["dataset"]["theorems"]):
|
| 68 |
+
if thm["contents"]:
|
| 69 |
+
thm_string = "\\section{" + thm["label"] + "}\n"
|
| 70 |
+
thm_string += (
|
| 71 |
+
"Tags: " + ", ".join(thm["categories"]).replace("/", ": ") + "\n\n"
|
| 72 |
+
)
|
|
|
|
|
|
|
|
|
|
| 73 |
|
| 74 |
+
thm_string += (
|
| 75 |
+
"\\begin{theorem}\n"
|
| 76 |
+
+ "\n".join(thm["contents"])
|
| 77 |
+
+ "\n\\end{theorem}\n\n"
|
| 78 |
+
)
|
| 79 |
+
|
| 80 |
+
for proof in thm["proofs"]:
|
| 81 |
thm_string += (
|
| 82 |
+
"\\begin{proof}\n"
|
| 83 |
+
+ "\n".join(proof["contents"])
|
| 84 |
+
+ "\n\\end{proof}\n\n"
|
| 85 |
)
|
| 86 |
|
| 87 |
+
if random.random()>VAL_RATE:
|
| 88 |
+
with open(os.path.join(save_dir, f"""thm_{thm["id"]}.txt"""), "w") as f:
|
| 89 |
+
f.write(thm_string)
|
| 90 |
+
else:
|
| 91 |
+
with open(os.path.join(val_dir, f"""thm_{thm["id"]}.txt"""), "w") as f:
|
| 92 |
+
f.write(thm_string)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
|
| 94 |
defn_strings = []
|
| 95 |
for defn in struct["dataset"]["definitions"]:
|
| 96 |
if defn["contents"]:
|
| 97 |
+
defn_string = (
|
| 98 |
"\\begin{definition}["
|
| 99 |
+ defn["label"]
|
| 100 |
+ "]\n"
|
| 101 |
+ "\n".join(defn["contents"])
|
| 102 |
+ "\n\\end{definition}").strip()
|
| 103 |
+
|
| 104 |
+
if random.random()>VAL_RATE:
|
| 105 |
+
with open(os.path.join(save_dir, f"""def_{defn["id"]}.txt"""), "w") as f:
|
| 106 |
+
f.write(defn_string)
|
| 107 |
+
else:
|
| 108 |
+
with open(os.path.join(val_dir, f"""def_{defn["id"]}.txt"""), "w") as f:
|
| 109 |
+
f.write(defn_string)
|
| 110 |
+
|
| 111 |
|
| 112 |
if __name__=="__main__":
|
| 113 |
#wikipedia()
|
| 114 |
proofwiki()
|
| 115 |
+
make_archive("wiki/proofwiki")
|
| 116 |
+
make_archive("wiki/proofwiki_val")
|
proof-pile.py
CHANGED
|
@@ -75,6 +75,7 @@ class ProofPile(datasets.GeneratorBasedBuilder):
|
|
| 75 |
# data = datasets.load_dataset('my_dataset', 'first_domain')
|
| 76 |
# data = datasets.load_dataset('my_dataset', 'second_domain')
|
| 77 |
BUILDER_CONFIGS = [
|
|
|
|
| 78 |
datasets.BuilderConfig(name="books", version=VERSION, description="Open source math textbooks"),
|
| 79 |
datasets.BuilderConfig(name="formal", version=VERSION, description="Formal math libraries"),
|
| 80 |
datasets.BuilderConfig(name="stack-exchange", version=VERSION, description="math overflow and math stack exchange"),
|
|
@@ -119,9 +120,19 @@ class ProofPile(datasets.GeneratorBasedBuilder):
|
|
| 119 |
with open("splits.json") as f:
|
| 120 |
splits = json.load(f)
|
| 121 |
|
| 122 |
-
self.archived_configs = ["stack-exchange", "math-dataset", "wiki"]
|
| 123 |
|
| 124 |
if self.config.name in self.archived_configs:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 125 |
if self.config.name=="stack-exchange":
|
| 126 |
train_paths = [os.path.join("./stack-exchange", x) for x in ["math_overflow.tar.gz",
|
| 127 |
"math_stack_exchange.tar.gz"]]
|
|
|
|
| 75 |
# data = datasets.load_dataset('my_dataset', 'first_domain')
|
| 76 |
# data = datasets.load_dataset('my_dataset', 'second_domain')
|
| 77 |
BUILDER_CONFIGS = [
|
| 78 |
+
datasets.BuilderConfig(name="arxiv", version=VERSION, description="All of English arxiv.math up to 03/22"),
|
| 79 |
datasets.BuilderConfig(name="books", version=VERSION, description="Open source math textbooks"),
|
| 80 |
datasets.BuilderConfig(name="formal", version=VERSION, description="Formal math libraries"),
|
| 81 |
datasets.BuilderConfig(name="stack-exchange", version=VERSION, description="math overflow and math stack exchange"),
|
|
|
|
| 120 |
with open("splits.json") as f:
|
| 121 |
splits = json.load(f)
|
| 122 |
|
| 123 |
+
self.archived_configs = ["arxiv", "stack-exchange", "math-dataset", "wiki"]
|
| 124 |
|
| 125 |
if self.config.name in self.archived_configs:
|
| 126 |
+
if self.config.name=="arxiv":
|
| 127 |
+
train_paths = []
|
| 128 |
+
val_paths = []
|
| 129 |
+
for f in os.listdir("arxiv"):
|
| 130 |
+
f_path = os.path.join("./arxiv", f)
|
| 131 |
+
# validation set is june of years divisible by 4
|
| 132 |
+
if int(f[1])%4==0 and int(f[3])==6:
|
| 133 |
+
val_paths.append(f_path)
|
| 134 |
+
else:
|
| 135 |
+
train_paths.append(f_path)
|
| 136 |
if self.config.name=="stack-exchange":
|
| 137 |
train_paths = [os.path.join("./stack-exchange", x) for x in ["math_overflow.tar.gz",
|
| 138 |
"math_stack_exchange.tar.gz"]]
|
test.py
CHANGED
|
@@ -2,6 +2,9 @@ from datasets import load_dataset
|
|
| 2 |
from itertools import islice
|
| 3 |
import sys
|
| 4 |
|
|
|
|
|
|
|
|
|
|
| 5 |
"""
|
| 6 |
dataset = load_dataset("./proof-pile.py", "books")
|
| 7 |
print("BOOKS")
|
|
@@ -20,11 +23,9 @@ for x in dataset["train"]:
|
|
| 20 |
print(x)
|
| 21 |
break
|
| 22 |
print(dataset)
|
| 23 |
-
"""
|
| 24 |
dataset = load_dataset("./proof-pile.py", "wiki")
|
| 25 |
print("WIKI")
|
| 26 |
print(dataset)
|
| 27 |
-
"""
|
| 28 |
dataset = load_dataset("./proof-pile.py", "math-dataset", download_mode='force_redownload')
|
| 29 |
print("MATH DATASET")
|
| 30 |
print(dataset)
|
|
|
|
| 2 |
from itertools import islice
|
| 3 |
import sys
|
| 4 |
|
| 5 |
+
dataset = load_dataset("./proof-pile.py", "arxiv")
|
| 6 |
+
print('ARXIV')
|
| 7 |
+
print(dataset)
|
| 8 |
"""
|
| 9 |
dataset = load_dataset("./proof-pile.py", "books")
|
| 10 |
print("BOOKS")
|
|
|
|
| 23 |
print(x)
|
| 24 |
break
|
| 25 |
print(dataset)
|
|
|
|
| 26 |
dataset = load_dataset("./proof-pile.py", "wiki")
|
| 27 |
print("WIKI")
|
| 28 |
print(dataset)
|
|
|
|
| 29 |
dataset = load_dataset("./proof-pile.py", "math-dataset", download_mode='force_redownload')
|
| 30 |
print("MATH DATASET")
|
| 31 |
print(dataset)
|
utils.py
CHANGED
|
@@ -1,7 +1,53 @@
|
|
| 1 |
import os
|
| 2 |
import tarfile
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
|
| 4 |
def make_archive(path):
|
| 5 |
with tarfile.open(path + ".tar.gz", "w:gz") as tar:
|
| 6 |
tar.add(path, arcname=os.path.sep)
|
| 7 |
os.system(f"rm -r {path}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import os
|
| 2 |
import tarfile
|
| 3 |
+
from itertools import cycle
|
| 4 |
+
from shutil import get_terminal_size
|
| 5 |
+
from threading import Thread
|
| 6 |
+
from time import sleep
|
| 7 |
|
| 8 |
def make_archive(path):
|
| 9 |
with tarfile.open(path + ".tar.gz", "w:gz") as tar:
|
| 10 |
tar.add(path, arcname=os.path.sep)
|
| 11 |
os.system(f"rm -r {path}")
|
| 12 |
+
|
| 13 |
+
class Loader:
|
| 14 |
+
def __init__(self, desc="Loading...", end="Done!", timeout=0.1):
|
| 15 |
+
"""
|
| 16 |
+
A loader-like context manager
|
| 17 |
+
|
| 18 |
+
Args:
|
| 19 |
+
desc (str, optional): The loader's description. Defaults to "Loading...".
|
| 20 |
+
end (str, optional): Final print. Defaults to "Done!".
|
| 21 |
+
timeout (float, optional): Sleep time between prints. Defaults to 0.1.
|
| 22 |
+
"""
|
| 23 |
+
self.desc = desc
|
| 24 |
+
self.end = end
|
| 25 |
+
self.timeout = timeout
|
| 26 |
+
|
| 27 |
+
self._thread = Thread(target=self._animate, daemon=True)
|
| 28 |
+
self.steps = ["⢿", "⣻", "⣽", "⣾", "⣷", "⣯", "⣟", "⡿"]
|
| 29 |
+
self.done = False
|
| 30 |
+
|
| 31 |
+
def start(self):
|
| 32 |
+
self._thread.start()
|
| 33 |
+
return self
|
| 34 |
+
|
| 35 |
+
def _animate(self):
|
| 36 |
+
for c in cycle(self.steps):
|
| 37 |
+
if self.done:
|
| 38 |
+
break
|
| 39 |
+
print(f"\r{self.desc} {c}", flush=True, end="")
|
| 40 |
+
sleep(self.timeout)
|
| 41 |
+
|
| 42 |
+
def __enter__(self):
|
| 43 |
+
self.start()
|
| 44 |
+
|
| 45 |
+
def stop(self):
|
| 46 |
+
self.done = True
|
| 47 |
+
cols = get_terminal_size((80, 20)).columns
|
| 48 |
+
print("\r" + " " * cols, end="", flush=True)
|
| 49 |
+
print(f"\r{self.end}", flush=True)
|
| 50 |
+
|
| 51 |
+
def __exit__(self, exc_type, exc_value, tb):
|
| 52 |
+
# handle exceptions with those variables ^
|
| 53 |
+
self.stop()
|