Plotting with custom benchmark_rets got error (VectorBT used undefined attribute 'obj')

[daviddwlee84]

When I tried to override the default benchmark returns #247 .

I found that VectorBT is try to broadcast it to an undefined "obj".

AttributeError: 'Portfolio' object has no attribute 'obj'

vectorbt/vectorbt/portfolio/base.py

Line 5154 in 67da37e

benchmark_rets = broadcast_to(benchmark_rets, self.obj)

Reproduction code

Simulate comparing two Portfolio results

import vectorbt as vbt
import numpy as np
import pandas as pd

close1 = pd.DataFrame(
    {
        "symbol_a": np.abs(np.random.randn(10)) * 10,
        "symbol_b": np.abs(np.random.randn(10)) * 10,
    }
)
close2 = close1[["symbol_a"]]

raw_signal1 = np.random.randn(10)
signal1_a = (raw_signal1 > 0.5) * 1 + (raw_signal1 < -0.5) * -1
signal1_b = (raw_signal1 > 0.5) * 1 + (raw_signal1 < -0.5) * -1
signal1 = pd.DataFrame({"symbol_a": signal1_a, "symbol_b": signal1_b})

raw_signal2 = pd.Series(np.random.randn(10), name="symbol_a")
signal2 = (raw_signal2 > 0.5) * 1 + (raw_signal2 < -0.5) * -1

pf1 = vbt.Portfolio.from_signals(
    close1, entries=signal1 == 1, exits=signal1 == -1, freq="T"
)
pf2 = vbt.Portfolio.from_signals(
    close2, entries=signal2 == 1, exits=signal2 == -1, freq="T"
)

# ok
print(pf1.stats(settings=dict(benchmark_rets=pf2.returns())))
# ok
print(pf1.stats(column="symbol_a", settings=dict(benchmark_rets=pf2.returns())))
# ok
print(pf2.stats(settings=dict(benchmark_rets=pf1['symbol_a'].returns())))

# ---- error ----

# works if I comment the bug line
pf1.plot(
    column="symbol_a",
    settings=dict(benchmark_rets=pf2.returns()),
    subplots=[
        "orders",
        "trade_pnl",
        "cum_returns",
        "drawdowns",
        "underwater",
        "asset_flow",
        "asset_value",
        "assets",
        "cash",
        "cash_flow",
        "gross_exposure",
        "net_exposure",
        "trades",
        "value",
    ],
)

# works if I comment the bug line
pf2.plot(
    column="symbol_a",
    settings=dict(benchmark_rets=pf1.returns()),
    subplots=[
        "orders",
        "trade_pnl",
        "cum_returns",
        "drawdowns",
        "underwater",
        "asset_flow",
        "asset_value",
        "assets",
        "cash",
        "cash_flow",
        "gross_exposure",
        "net_exposure",
        "trades",
        "value",
    ],
)

Version

• vectorbt: 0.26.1
• Python 3.8.13