fix: name the missing price CSV instead of raising a polars traceback

book/marimo/notebooks/optimize.py

@@ -12,6 +12,7 @@
 from __future__ import annotations
 
 import argparse
+import sys
 import warnings
 from collections.abc import Callable
 from functools import cache
@@ -292,8 +293,15 @@ def optimize(key: str, *, n_trials: int, seed: int) -> optuna.Study:
     return study
 
 
-def main(argv: list[str] | None = None) -> None:
-    """Parse command-line arguments and run the requested Optuna study/studies."""
+def main(argv: list[str] | None = None) -> int:
+    """Parse command-line arguments and run the requested Optuna study/studies.
+
+    Returns the process exit code: ``0`` on success, ``1`` when the price data the
+    experiments read is missing. That case is caught here rather than allowed to
+    propagate because it is the one failure a user hits through no fault of the
+    search — a setup problem deserving a one-line message naming the file, not a
+    traceback out of the loader.
+    """
     parser = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
     parser.add_argument(
         "--experiment",
@@ -315,10 +323,15 @@ def main(argv: list[str] | None = None) -> None:
     if not args.verbose:
         optuna.logging.set_verbosity(optuna.logging.WARNING)
     keys = list(EXPERIMENTS.keys()) if args.experiment == "all" else [args.experiment]
-    for key in keys:
-        n_trials = args.trials if args.trials is not None else DEFAULT_TRIALS[key]
-        optimize(key, n_trials=n_trials, seed=args.seed)
+    try:
+        for key in keys:
+            n_trials = args.trials if args.trials is not None else DEFAULT_TRIALS[key]
+            optimize(key, n_trials=n_trials, seed=args.seed)
+    except FileNotFoundError as error:
+        print(f"error: {error}", file=sys.stderr)
+        return 1
+    return 0
 
 
 if __name__ == "__main__":
-    main()
+    sys.exit(main())

book/marimo/notebooks/preamble.py

@@ -23,6 +23,9 @@
 #: Directory holding the marimo notebooks (this file's own directory).
 NOTEBOOK_DIR = Path(__file__).resolve().parent
 
+#: Price file every notebook reads, from the ``public/`` directory beside it.
+PRICES_CSV = "Prices_hashed.csv"
+
 
 def load_notebook(name: str) -> dict[str, Any]:
     """Execute sibling notebook ``name`` (e.g. ``"Experiment1.py"``) and return its namespace.
@@ -56,8 +59,21 @@ def load_prices(notebook_file: str) -> pl.DataFrame:
         Datetime(time_unit='ns', time_zone=None)
         >>> set(prices.drop(date_col).dtypes) == {pl.Float64}
         True
+
+    An absent CSV is checked for here rather than left to ``pl.read_csv``, so the
+    error names the file that was looked for — the file ships with the repository,
+    so its absence is a setup problem with an obvious remedy:
+
+        >>> try:
+        ...     load_prices(str(NOTEBOOK_DIR / "elsewhere" / "preamble.py"))
+        ... except FileNotFoundError as error:
+        ...     PRICES_CSV in str(error)
+        True
     """
-    path = Path(notebook_file).parent / "public" / "Prices_hashed.csv"
+    path = Path(notebook_file).parent / "public" / PRICES_CSV
+    if not path.is_file():
+        msg = f"Price data not found: {path} — it ships with the repository; check the checkout is complete."
+        raise FileNotFoundError(msg)
     dframe = pl.read_csv(str(path), try_parse_dates=True)
     dframe = dframe.with_columns(pl.col(date_col).cast(pl.Datetime("ns")))
     dframe = dframe.with_columns([pl.col(col).cast(pl.Float64) for col in dframe.columns if col != date_col])

tests/test_optimize.py

@@ -96,13 +96,37 @@ def _raise(*_args, **_kwargs):
 
 
 def test_main_runs_single_experiment(capsys):
-    """The CLI entry point runs one experiment and prints its summary."""
-    optimize["main"](["--experiment", "1", "--trials", "1"])
+    """The CLI entry point runs one experiment, prints its summary and reports success."""
+    assert optimize["main"](["--experiment", "1", "--trials", "1"]) == 0
     out = capsys.readouterr().out
     assert "Experiment 1" in out
     assert "best Sharpe" in out
 
 
+def test_main_reports_missing_price_data_without_a_traceback(capsys, monkeypatch):
+    """Absent price data exits non-zero with a message naming the file, not a traceback.
+
+    ``load_prices`` raises ``FileNotFoundError`` when the CSV is gone; ``main`` is the
+    only entry point a user drives directly, so it must turn that into a one-line
+    diagnostic. Patching the module-global ``optimize`` (the same ``__globals__``
+    technique the singular-matrix and zero-baseline tests use) raises the error from
+    inside the loop without deleting the file the rest of the suite reads.
+    """
+
+    def _raise(*_args, **_kwargs):
+        """Stand in for the study runner, failing the way a missing price CSV does."""
+        msg = "Price data not found: /nowhere/public/Prices_hashed.csv"
+        raise FileNotFoundError(msg)
+
+    main = optimize["main"]
+    monkeypatch.setitem(main.__globals__, "optimize", _raise)
+
+    assert main(["--experiment", "1", "--trials", "1"]) == 1
+    captured = capsys.readouterr()
+    assert "Prices_hashed.csv" in captured.err
+    assert "Traceback" not in captured.err
+
+
 def test_default_trials_cover_every_experiment():
     """DEFAULT_TRIALS has an entry for each registered experiment (no KeyError in main)."""
     assert set(optimize["DEFAULT_TRIALS"]) == set(optimize["EXPERIMENTS"])

tests/test_preamble.py

@@ -12,6 +12,7 @@
 preamble = load_notebook("preamble.py")
 load_prices = preamble["load_prices"]
 date_col = preamble["date_col"]
+PRICES_CSV = preamble["PRICES_CSV"]
 
 
 @pytest.fixture(scope="module")
@@ -58,6 +59,18 @@ def test_load_prices_non_date_columns_are_float64(prices_df):
             assert prices_df[col].dtype == pl.Float64, f"Column {col!r} is not Float64"
 
 
+def test_load_prices_missing_csv_names_the_expected_path(tmp_path):
+    """A missing price CSV raises FileNotFoundError naming the file that was looked for.
+
+    ``tmp_path`` stands in for a caller sitting outside the notebook directory, so
+    the ``public/`` sibling the loader resolves does not exist. The guard must fire
+    before ``pl.read_csv``, whose own error names neither the expected location nor
+    the fact that the file ships with the repository.
+    """
+    with pytest.raises(FileNotFoundError, match=PRICES_CSV):
+        load_prices(str(tmp_path / "Experiment1.py"))
+
+
 def test_load_prices_interpolation_applied(prices_df):
     """load_prices applies interpolation to the raw data."""
     from jquantstats import interpolate