portfolio-optimiser/tests/test_scripted_cli_door_loadbearing.py

"""The offline CLI door: run the WHOLE loop over your OWN bundle with zero model calls.

Before this, an adopter without an API budget had two half-doors and no whole one:
``--live-dry-run`` takes their own bundle but STOPS before the first model call
(``run.py`` returns a ``DryRunReport``), while ``portfolio_optimiser.simulation`` runs the
complete loop but only over ITS bundle with ITS scripted answers. The seam for the missing
third case already existed — ``run_project(client_factory=...)``, the test-injection seam —
with no CLI exposure at all. ``--scripted-replies <file.json>`` is that door.

**The honesty condition is part of the feature, not decoration** (målbilde §1): a scripted run
that reads like a model run is worse than no offline mode, so the banner asserted here is
load-bearing in the same sense the dataflow is. It mirrors ``simulation.py``'s honesty banner.

Load-bearing (each blade measured by detaching exactly one thing):
1. the door drives a FULL run offline over a caller-supplied bundle (RED without the flag);
2. the run is genuinely model-free (RED if a real factory is built);
3. the banner is emitted and unmistakable (RED when detached);
4. control: no banner without the flag — so blade 3 cannot pass on a constant;
5. the two offline modes are mutually exclusive rather than one silently winning.
"""

from __future__ import annotations

import json
import shutil
from pathlib import Path

import pytest

from portfolio_optimiser import run
from portfolio_optimiser.ledger import SavingsLedger

BUNDLE_DIR = Path(__file__).resolve().parents[1] / "shared" / "examples" / "bygg-energi-mikro"

# The same scripted answers the simulation uses, re-declared here as CALLER-supplied input —
# which is the whole point of the door: these arrive as a file the adopter writes, not as a
# module constant only the shipped simulation can reach.
_VALID_PROPOSAL = (
    '{"measure":"LED-retrofit av kontorbelysning","affected_items":'
    '[{"code":"ENERGI-TOTAL-EL","quantity":300000,"unit_cost":1.0}],"claimed_saving_nok":30000}'
)
_CHECKER_APPROVE = "Tallene er innenfor feasibelt område og resonnementet holder. VERDICT: APPROVE"


@pytest.fixture(autouse=True)
def _isolate_model_env(monkeypatch: pytest.MonkeyPatch) -> None:
    """Hermetic env (mirrors ``test_live_dry_run.py``): the operator's Foundry overrides must not
    reach these assertions."""
    monkeypatch.delenv("PORTFOLIO_MODEL_MAP", raising=False)
    monkeypatch.delenv("PORTFOLIO_FOUNDRY_PROJECT_ENDPOINT", raising=False)


@pytest.fixture()
def bundle(tmp_path: Path) -> Path:
    """A throwaway COPY — the shared fixture is commons-owned and is never mutated by a test."""
    dst = tmp_path / "bundle"
    shutil.copytree(BUNDLE_DIR, dst)
    return dst


@pytest.fixture()
def replies_file(tmp_path: Path) -> Path:
    path = tmp_path / "replies.json"
    path.write_text(
        json.dumps({"proposer": _VALID_PROPOSAL, "checker": _CHECKER_APPROVE}),
        encoding="utf-8",
    )
    return path


def _argv(bundle: Path, replies_file: Path) -> list[str]:
    return [
        "BYGG-KONTOR-NORD",
        "--docs-dir",
        str(bundle),
        "--bundle-dir",
        str(bundle),
        "--scripted-replies",
        str(replies_file),
    ]


def test_scripted_door_runs_the_whole_loop_offline(bundle, replies_file, capsys) -> None:
    """Blade 1 — the door exists and completes a FULL run (not a dry-run report) over the
    caller's own bundle. RED before the flag: argparse exits 2 on an unrecognized argument."""
    rc = run.main(_argv(bundle, replies_file))
    out = capsys.readouterr().out
    assert rc == 0, out
    # A full run reports its outcome type + the candidate's verdict key; a dry-run never gets this
    # far. Since F2 this argv records NO expert verdict (no --decision/--rationale), so the line
    # says so and quotes the key one would arrive under — asserting "verdict id=" here would be
    # asserting that a run nobody reviewed minted an approval.
    assert "BYGG-KONTOR-NORD:" in out
    assert "no expert verdict given; verdict key=" in out
    assert "LIVE-DRY-RUN" not in out


def test_scripted_door_makes_no_real_client(bundle, replies_file, monkeypatch, capsys) -> None:
    """Blade 2 — genuinely model-free: the production factory must never be built. Detonates if
    the scripted factory is ignored and the run falls back to ``_default_factory``."""

    def _boom(_profile: str):  # pragma: no cover - the point is that it never runs
        raise AssertionError("a real client factory was built on the scripted path")

    monkeypatch.setattr(run, "_default_factory", _boom)
    rc = run.main(_argv(bundle, replies_file))
    assert rc == 0, capsys.readouterr().out


def test_scripted_door_says_so_unmistakably(bundle, replies_file, capsys) -> None:
    """Blade 3 — the honesty banner. A scripted run that looks like a model run is the failure
    mode this asserts against; RED the moment the banner is detached."""
    run.main(_argv(bundle, replies_file))
    out = capsys.readouterr().out.upper()
    assert "SCRIPTED" in out
    assert "NO MODEL" in out or "INGEN MODELLKALL" in out


def test_no_banner_without_the_flag(bundle, capsys) -> None:
    """Blade 4 (control) — the banner must be CAUSED by the flag, not printed unconditionally.
    Uses --live-dry-run so the control needs no model call of its own."""
    run.main(
        [
            "BYGG-KONTOR-NORD",
            "--docs-dir",
            str(bundle),
            "--bundle-dir",
            str(bundle),
            "--live-dry-run",
        ]
    )
    assert "SCRIPTED" not in capsys.readouterr().out.upper()


def test_the_two_offline_modes_are_exclusive(bundle, replies_file, capsys) -> None:
    """Blade 5 — ``--live-dry-run`` stops before the first call and ``--scripted-replies`` runs
    every call; together they are a contradiction. Refuse, never let one silently win (mirrors
    S5.3's 'refused, never ignored' partition)."""
    rc = run.main([*_argv(bundle, replies_file), "--live-dry-run"])
    assert rc == 1
    assert "refused" in capsys.readouterr().err.lower()


def test_report_mode_refuses_the_scripted_flag(tmp_path, replies_file, capsys) -> None:
    """Blade 5b — ``--report`` is mode-exclusive by ALLOWLIST; a new flag must join the refusal
    set rather than be silently dropped.

    The ledger here is VALID and saved on purpose. Measured: with a non-existent ledger path this
    test passed even after the allowlist entry was removed — the load failure refused first and
    masked the gate entirely. With a loadable ledger the report would otherwise print and return
    0, so rc 1 can only come from the mode-exclusivity check itself.
    """
    ledger_path = tmp_path / "ledger.json"
    SavingsLedger().save(str(ledger_path))
    rc = run.main(
        ["--report", "--ledger", str(ledger_path), "--scripted-replies", str(replies_file)]
    )
    assert rc == 1
    assert "mode-exclusive" in capsys.readouterr().err.lower()


def test_malformed_replies_file_is_refused(bundle, tmp_path, capsys) -> None:
    """A replies file that exists but cannot serve the run is refused with rc 1, never a
    traceback and never a partial run — the same fail-fast posture the loaders take."""
    bad = tmp_path / "bad.json"
    bad.write_text("{not json", encoding="utf-8")
    rc = run.main(_argv(bundle, bad))
    assert rc == 1
    assert "refused" in capsys.readouterr().err.lower()