import io import logging import lz4.frame import os import requests import shutil import zipfile import time from env import BASE_DIR from pandas.tseries.offsets import BDay logger = logging.getLogger(__name__) def convertToNone(v): return v if v else None def download_cds_data(payload, workdate, report="FIXED_COUPON"): payload.update({"report": report}) r = requests.post("https://www.markit.com/export.jsp", params=payload) content = io.BytesIO(r.content) save_dir = BASE_DIR / "Tranche_data" / "CDS" / f"{workdate:%Y}" if not save_dir.exists(): save_dir.mkdir() suffix = "fixed" if report == "FIXED_COUPON" else "parspread" csv_file = save_dir / f"{workdate}_{suffix}.csv.lz4" try: with zipfile.ZipFile(content) as z: with lz4.frame.open(csv_file, "wb") as f2: for f in z.namelist(): if f.endswith("csv"): f1 = z.open(f) if report == "FIXED_COUPON": next(f1) next(f1) shutil.copyfileobj(f1, f2) f1.close() except zipfile.BadZipFile: logger.error(content.getvalue().decode().strip()) def download_composite_data( payload, workdate, historical=False, reports=("COMPOSITES", "TRANCHE_COMPOSITES") ): # if historical, we want to maintain the invariant mtime(f)== payload['date'] + BDay(1) if historical: ts = (workdate + BDay(1)).timestamp() for report in reports: for family in ["CDX", "ITRAXX-EUROPE"]: payload.update({"family": family, "report": report}) retry = 0 while retry < 10: r = requests.post("https://www.markit.com/export.jsp", params=payload) try: with zipfile.ZipFile(io.BytesIO(r.content)) as z: for f in z.namelist(): if f.endswith("csv"): path = z.extract( f, path=os.path.join( os.environ["BASE_DIR"], "Tranche_data", "Composite_reports", ), ) if historical: os.utime(path, (ts, ts)) except zipfile.BadZipfile: logger.error(r.content.decode().strip()) time.sleep(5) retry += 1 continue else: break