llm-ingestion-okf/tests/test_folder_of_bundles.py

"""One folder, every bundle under it: the command line's own door (v1.1 F).

The server has read a FOLDER of bundles since `okf mcp --root`; the generic
skill read one bundle at a time, the one it was pointed at, and could not see
the others. These tests hold the two command-line entries the skill now uses
-- `okf card <folder>` and `okf consume <folder>` -- to the server's OWN
functions: one source, two doors. Both corpora are invented here.
"""

from __future__ import annotations

import json
import subprocess
import sys
from pathlib import Path

import pytest

from llm_ingestion_okf import mcp_server
from llm_ingestion_okf.cli import build

BAKERY = {
    "surdeig.md": (
        "# Surdeig\n\n"
        "## Heving\n\n"
        "Surdeigen hever i tolv timer ved romtemperatur før den formes.\n\n"
        "## Steking\n\n"
        "Brødet stekes i førti minutter på to hundre og tretti grader.\n"
    ),
}
GARDEN = {
    "tomater.md": (
        "# Tomater\n\n"
        "## Vanning\n\n"
        "Tomatene vannes hver morgen, og jorda skal aldri tørke helt ut.\n\n"
        "## Oppbinding\n\n"
        "Plantene bindes opp til en stokk når de er tretti centimeter høye.\n"
    ),
}


def _bundle(tmp_path: Path, folder: Path, name: str, documents: dict[str, str]) -> Path:
    source = tmp_path / f"src-{name}"
    source.mkdir()
    for file_name, text in documents.items():
        (source / file_name).write_text(text, encoding="utf-8")
    target = folder / name
    build(source, target, bundle_id=name, okf_version="0.2")
    return target


@pytest.fixture(scope="module")
def folder(tmp_path_factory: pytest.TempPathFactory) -> Path:
    tmp_path = tmp_path_factory.mktemp("folder-of-bundles")
    root = tmp_path / "samlinger"
    root.mkdir()
    _bundle(tmp_path, root, "bakeri", BAKERY)
    _bundle(tmp_path, root, "hage", GARDEN)
    return root


def _okf(*argv: str) -> subprocess.CompletedProcess[str]:
    return subprocess.run(
        [sys.executable, "-m", "llm_ingestion_okf.cli", *argv],
        capture_output=True,
        text=True,
        check=False,
    )


def _surface(folder: Path) -> mcp_server.Surface:
    return mcp_server.build_surface(bundle=None, roots=[folder])


# --- F1: the overview -------------------------------------------------------


def test_the_card_of_a_folder_names_every_bundle_under_it(folder: Path) -> None:
    run = _okf("card", str(folder))
    assert run.returncode == 0, run.stderr
    overview = json.loads(run.stdout)
    assert [card["bundle_id"] for card in overview["cards"]] == ["bakeri", "hage"]
    assert [entry["bundle_id"] for entry in overview["bundles"]] == ["bakeri", "hage"]
    assert overview["unreadable"] == []
    assert overview["shape"] == "one-to-many"
    # Each card carries the map, which is what the working method reads first.
    assert all(card["map"] for card in overview["cards"])


def test_the_card_of_a_folder_is_the_servers_own_listing_and_description(folder: Path) -> None:
    """One source: the bytes the command prints are the server's two replies."""
    surface = _surface(folder)
    listing = mcp_server.call_list(surface, {})
    described = mcp_server.call_describe(surface, {})
    overview = json.loads(_okf("card", str(folder)).stdout)
    assert overview["bundles"] == listing["bundles"]
    assert overview["unreadable"] == listing["unreadable"]
    assert overview["cards"] == described["cards"]


def test_the_card_of_one_bundle_is_unchanged(folder: Path) -> None:
    """Pointed at one bundle, the command prints that bundle's card, as before."""
    bundle = folder / "bakeri"
    run = _okf("card", str(bundle))
    assert run.returncode == 0, run.stderr
    card = json.loads(run.stdout)
    assert card == mcp_server.card(bundle.resolve(), profile=mcp_server.okf_consume.DEFAULT_PROFILE)
    assert "cards" not in card


def test_a_folder_holding_no_bundle_is_refused_rather_than_empty(tmp_path: Path) -> None:
    empty = tmp_path / "tom"
    empty.mkdir()
    run = _okf("card", str(empty))
    assert run.returncode == 1
    assert "bundle_none_served" in run.stderr


def test_a_broken_bundle_under_the_folder_is_reported(folder: Path, tmp_path: Path) -> None:
    root = tmp_path / "med-feil"
    root.mkdir()
    for name in ("bakeri", "hage"):
        (root / name).symlink_to(folder / name)  # never followed: not listed
    broken = root / "odelagt"
    broken.mkdir()
    (broken / "index.md").write_text("---\ntitle: x\n---\n", encoding="utf-8")
    real = root / "ekte"
    real.mkdir()
    for source in (folder / "hage").rglob("*"):
        target = real / source.relative_to(folder / "hage")
        if source.is_dir():
            target.mkdir(parents=True, exist_ok=True)
        else:
            target.write_bytes(source.read_bytes())
    overview = json.loads(_okf("card", str(root)).stdout)
    assert [card["bundle_id"] for card in overview["cards"]] == ["hage"]
    assert overview["unreadable"] == [
        {"directory": "odelagt", "reason": "index.md declares no bundle_id"}
    ]