[ENH] Remove yfinance as a dependency and implement data_loader

pypfopt/data/__init__.py

@@ -0,0 +1,3 @@
+from .data_loader import load_stockdata, available_tickers, load_marketcaps
+
+__all__ = ["load_stockdata", "available_tickers", "load_marketcaps"]

pypfopt/data/data_loader.py

@@ -0,0 +1,47 @@
+import pandas as pd
+from importlib import resources
+
+
+def _load_raw_data(filename: str, **read_csv_kwargs):
+    with resources.files(__package__).joinpath(filename).open("r") as f:
+        return pd.read_csv(f, **read_csv_kwargs)
+
+
+def load_stockdata(tickers: list = None, start: str = None, end: str = None):
+
+    df = _load_raw_data("stock_prices.csv", parse_dates=["date"])
+
+    if start is not None:
+        df = df[df["date"] >= pd.to_datetime(start)]
+    if end is not None:
+        df = df[df["date"] <= pd.to_datetime(end)]
+
+    if tickers is not None:
+        cols = ["date"] + tickers
+        df = df[cols]
+
+    return df.set_index("date")
+
+
+def load_marketcaps(tickers: list = None):
+
+    df = _load_raw_data("market_caps.csv")
+
+    if tickers is not None:
+        available = set(df["ticker"])
+        invalid = set(tickers) - available
+        if invalid:
+            raise ValueError(f"Invalid tickers: {invalid}")
+
+        df = df[df["ticker"].isin(tickers)]
+
+    return dict(zip(df["ticker"], df["market_cap"]))
+
+
+def available_tickers():
+
+    df = _load_raw_data("stock_prices.csv", parse_dates=["date"])
+    cols = [c for c in df.columns if c != "date"]
+    cols.sort()
+
+    return cols

pypfopt/data/market_caps.csv

@@ -0,0 +1,34 @@
+,ticker,market_cap
+0,AAPL,3890834833408
+1,ACN,127705194496
+2,AMD,323832184832
+3,AMZN,2237050126336
+4,BAC,363737317376
+5,BLK,165749276672
+6,COST,445187457024
+7,CVS,103888756736
+8,DIS,185017466880
+9,DPZ,13497629696
+10,F,53419077632
+11,GILD,186250330112
+12,INTU,116666302464
+13,JD,38876950528
+14,JPM,802528952320
+15,KO,345075089408
+16,LUV,24963076096
+17,MA,464960028672
+18,MCD,238929035264
+19,MSFT,2962173984768
+20,NAT,1262033920
+21,NVDA,4435176587264
+22,PBI,1733088640
+23,PFE,154992394240
+24,SBUX,110238670848
+25,SPY,629947236352
+26,TGT,51424350208
+27,TM,317571497984
+28,TSLA,1513430843392
+29,UL,156296904704
+30,UNH,267158962176
+31,WMT,1013349351424
+32,XOM,642598305792

pyproject.toml

@@ -62,7 +62,6 @@ all_extras = [
 dev = [
     "pytest>=9.0.0",
     "pytest-cov>=7.0.0",
-    "yfinance>=0.2.66",
 ]
 
 # notebook tests
@@ -87,6 +86,9 @@ requires = [
 [tool.setuptools.packages.find]
 exclude = ["example", "example.*", "tests", "tests.*"]
 
+[tool.setuptools.package-data]
+pypfopt = ["data/*.csv"]
+
 [tool.ruff]
 line-length = 88
 # Keep Ruff aligned with project target version