config.py

"""ABOUTME: Global config for CrossSectionCN.
ABOUTME: Paths, date ranges, universe filters."""
import os
from datetime import date
from pathlib import Path

from dotenv import load_dotenv

load_dotenv()

ROOT = Path(__file__).parent
DATA_RAW_VENDOR = ROOT / "pyData" / "raw_vendor"
DATA_CANONICAL = ROOT / "pyData" / "canonical"
# Backward-compatible alias: existing intermediate/factor code reads canonical tables here.
DATA_RAW = DATA_CANONICAL
DATA_INTERMEDIATE = ROOT / "pyData" / "intermediate"
DATA_FACTORS = ROOT / "pyData" / "factors"

START_DATE = date(2010, 1, 1)
END_DATE = date.today()

UNIVERSE_EXCHANGES = ["SH", "SZ"]
UNIVERSE_BOARDS = ["MAIN", "GEM", "STAR"]

FACTOR_DOC_PATH = ROOT / "factor_doc.csv"

DATA_SOURCE = os.environ.get("CNCS_DATA_SOURCE", "tushare").lower()
TUSHARE_TOKEN = os.environ.get("TUSHARE_TOKEN") or os.environ.get("TS_TOKEN")

# Dry-run / smoke: limit universe size (0 = no limit)
MAX_STOCKS = int(os.environ.get("CNCS_MAX_STOCKS", "0") or "0")

# Parallel-ish batch size for OHLCV downloads (symbols per HTTP batch call)
OHLCV_BATCH = int(os.environ.get("CNCS_OHLCV_BATCH", "25") or "25")

# Tushare official limits vary by endpoint/account; default leaves room below 500/min.
TUSHARE_MAX_REQUESTS_PER_MINUTE = int(os.environ.get("CNCS_TUSHARE_RPM", "450") or "450")

for d in (DATA_RAW_VENDOR, DATA_CANONICAL, DATA_INTERMEDIATE, DATA_FACTORS):
    d.mkdir(parents=True, exist_ok=True)