Datasets:
File size: 3,351 Bytes
5237fbc | 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 | """Convert every structure in dataset_划分.csv to MatText text representations.
Run inside venv-rep. Output: reps.parquet with one row per structure.
Light representations and SLICES are computed separately so a SLICES failure
does not lose the others. Each structure gets a wall-clock limit.
"""
import os, sys, csv, signal, time, json
os.environ.setdefault("CUDA_VISIBLE_DEVICES", "")
os.environ.setdefault("TF_CPP_MIN_LOG_LEVEL", "3")
for v in ("OMP_NUM_THREADS", "MKL_NUM_THREADS", "OPENBLAS_NUM_THREADS", "TF_NUM_INTRAOP_THREADS", "TF_NUM_INTEROP_THREADS"):
os.environ.setdefault(v, "1")
import warnings; warnings.filterwarnings("ignore")
from multiprocessing import Pool
ROOT = sys.argv[1]
OUT = sys.argv[2]
NPROC = int(sys.argv[3]) if len(sys.argv) > 3 else 32
LIGHT = ["composition", "cif_symmetrized", "cif_p1", "atom_sequences", "atom_sequences_plusplus", "crystal_text_llm", "zmatrix"]
HEAVY = ["slices", "local_env"]
LIMIT = {"slices": 180, "local_env": 120}
class TO(Exception): pass
def _alarm(*a): raise TO()
def resolve(rel):
d, f = rel.split("/", 1)
for pre in ("P-", "N-", ""):
p = os.path.join(ROOT, pre + d, f)
if os.path.exists(p): return p
return None
def work(row):
from pymatgen.core import Structure
from xtal2txt.core import TextRep
out = dict(id=row["id"], label=int(row["label"]), set=row["set"], source_dir=row["source_dir"])
err = {}
p = resolve(row["cif_path"])
try:
s = Structure.from_file(p)
out.update(formula=s.composition.reduced_formula, nsites=len(s), nelements=len(s.composition.elements),
is_ordered=bool(s.is_ordered))
try: out["spacegroup"] = int(s.get_space_group_info(symprec=0.1)[1])
except Exception: out["spacegroup"] = None
tr = TextRep.from_input(s)
except Exception as e:
out["error"] = json.dumps({"load": repr(e)[:200]}); return out
signal.signal(signal.SIGALRM, _alarm)
for rep in LIGHT + HEAVY:
signal.alarm(LIMIT.get(rep, 60))
try:
v = tr.get_requested_text_reps([rep])[rep]
out[rep] = v if isinstance(v, str) and v else None
if out[rep] is None: err[rep] = "empty"
except TO:
out[rep] = None; err[rep] = "timeout"
except Exception as e:
out[rep] = None; err[rep] = repr(e)[:200]
finally:
signal.alarm(0)
out["error"] = json.dumps(err) if err else None
return out
if __name__ == "__main__":
rows = list(csv.DictReader(open(os.path.join(ROOT, "dataset_划分.csv"), encoding="utf-8-sig")))
if len(sys.argv) > 4: rows = rows[: int(sys.argv[4])]
t0 = time.time(); res = []
with Pool(NPROC, maxtasksperchild=200) as pool:
for i, r in enumerate(pool.imap_unordered(work, rows, chunksize=4)):
res.append(r)
if (i + 1) % 500 == 0: print(f"{i+1}/{len(rows)} {time.time()-t0:.0f}s", flush=True)
import pandas as pd
df = pd.DataFrame(res)
order = {r["id"]: i for i, r in enumerate(rows)}
df = df.sort_values("id", key=lambda c: c.map(order)).reset_index(drop=True)
df.to_parquet(OUT)
print("rows", len(df), "time", f"{time.time()-t0:.0f}s")
for rep in LIGHT + HEAVY:
if rep in df: print(f"{rep:26s} ok={df[rep].notna().sum()} missing={df[rep].isna().sum()}")
|