rein-aharness/tests/test_assembly_checkpoint.py

"""Tests for the assembly-checkpoint approach (offline; local bare origin)."""

import subprocess
import textwrap
from pathlib import Path

import pytest

from rein_aharness import approaches as ap
from rein_aharness import assembly_checkpoint as ac
from rein_aharness.assembly_checkpoint import (AssemblyCheckpointError, ensure_web_fetcher, kind_from_title,
                                               run_assembly_checkpoint)

RUNNER = textwrap.dedent('''
    import json, sys, pathlib
    kind = sys.argv[sys.argv.index("--kind") + 1]
    cache = pathlib.Path(sys.argv[sys.argv.index("--cache") + 1])
    assert cache.is_dir() and not cache.resolve().is_relative_to(pathlib.Path.cwd().resolve())
    mode = pathlib.Path("checkpoints/MODE").read_text().strip()
    if mode == "out":
        print(json.dumps({"status": "out-of-window", "date": "2026-10-12", "kind": kind})); sys.exit(3)
    if mode == "outside":
        pathlib.Path("README.md").write_text("tampered")
    p = pathlib.Path("checkpoints/2026/2026-10-14-" + kind + ".md")
    p.parent.mkdir(parents=True, exist_ok=True); p.write_text("# record")
    print(json.dumps({"status": "ok", "date": "2026-10-14", "kind": kind, "path": str(p), "ok": 5,
                      "not_found": 4, "error": 0, "assessment": False, "phase": "setup"}))
''')


def _git(cwd, *args):
    subprocess.run(["git", "-C", str(cwd), *args], check=True, capture_output=True)


@pytest.fixture(autouse=True)
def _web_fetcher_path(tmp_path, monkeypatch):
    monkeypatch.setenv("WEB_FETCHER_PATH", str(tmp_path / "wf"))
    monkeypatch.delenv("ASSEMBLY_WEB_FETCHER_PATH", raising=False)


@pytest.fixture
def repo(tmp_path: Path) -> Path:
    origin = tmp_path / "origin.git"
    _git(tmp_path, "init", "-q", "--bare", "-b", "main", str(origin))
    work = tmp_path / "work"
    _git(tmp_path, "clone", "-q", str(origin), str(work))
    _git(work, "config", "user.email", "t@t"), _git(work, "config", "user.name", "t")
    (work / "checkpoints").mkdir()
    (work / "checkpoints" / "run_checkpoint.py").write_text(RUNNER)
    (work / "checkpoints" / "MODE").write_text("in")
    (work / "README.md").write_text("x")
    _git(work, "add", "-A"), _git(work, "commit", "-qm", "init"), _git(work, "push", "-q", "origin", "HEAD:main")
    return work


def test_kind_from_title():
    assert kind_from_title("Assembly Economics reality checkpoint (full) — steady phase") == "full"
    assert kind_from_title("anything") == "light"


def test_in_window_commits_and_pushes(repo):
    r = run_assembly_checkpoint(repo, kind="light", report_to_hub=False)
    assert r.ok and r.committed and r.pushed and r.path.endswith("2026-10-14-light.md")
    origin_head = subprocess.run(["git", "-C", str(repo), "rev-parse", "origin/main"],
                                 capture_output=True, text=True).stdout.strip()
    assert origin_head == r.origin_sha


def test_out_of_window_is_noop(repo):
    (repo / "checkpoints" / "MODE").write_text("out")
    _git(repo, "commit", "-qam", "mode out"), _git(repo, "push", "-q", "origin", "HEAD:main")
    r = run_assembly_checkpoint(repo, kind="light", report_to_hub=False)
    assert r.ok and r.out_of_window and not r.committed and not r.pushed


def test_refuses_paths_outside_checkpoints(repo):
    (repo / "checkpoints" / "MODE").write_text("outside")
    _git(repo, "commit", "-qam", "mode outside"), _git(repo, "push", "-q", "origin", "HEAD:main")
    r = run_assembly_checkpoint(repo, kind="light", report_to_hub=False)
    assert r.ok  # README change is not staged: only checkpoints/ is added
    changed = subprocess.run(["git", "-C", str(repo), "show", "--name-only", "--format=", "HEAD"],
                             capture_output=True, text=True).stdout.split()
    assert all(p.startswith("checkpoints/") for p in changed)


def test_label_selects_assembly_checkpoint():
    from rein_aharness.ops_run_client import OpsRun
    import inspect
    kw = {p: None for p in inspect.signature(OpsRun).parameters}
    for k, p in inspect.signature(OpsRun).parameters.items():
        if p.default is not inspect.Parameter.empty:
            kw[k] = p.default
    kw.update(id="x", title="Assembly Economics reality checkpoint (light) — setup phase", description="d",
              labels=["assembly-checkpoint", "research", "automated"], target_repo="assembly-economics",
              activity_definition_id="assembly-checkpoint-setup-daily")
    assert ap.select_approach(OpsRun(**kw)) == ap.APPROACH_ASSEMBLY_CHECKPOINT


def _tagged_origin(tmp_path: Path, tag: str) -> Path:
    src = tmp_path / "wf-src"
    _git(tmp_path, "init", "-q", "-b", "main", str(src))
    _git(src, "config", "user.email", "t@t"), _git(src, "config", "user.name", "t")
    (src / "README").write_text("wf")
    _git(src, "add", "-A"), _git(src, "commit", "-qm", "init"), _git(src, "tag", tag)
    return src


def test_ensure_web_fetcher_clones_pinned_tag(tmp_path, monkeypatch):
    monkeypatch.delenv("WEB_FETCHER_PATH")
    monkeypatch.setenv("ASSEMBLY_WEB_FETCHER_URL", (_tagged_origin(tmp_path, ac.WEB_FETCHER_TAG)).as_uri())
    dest = ensure_web_fetcher(tmp_path / "cache")
    assert dest == tmp_path / "cache" / f"web-fetcher-{ac.WEB_FETCHER_TAG}" and (dest / "README").exists()
    assert ensure_web_fetcher(tmp_path / "cache") == dest  # reused, re-verified


def test_ensure_web_fetcher_refuses_drifted_checkout(tmp_path, monkeypatch):
    monkeypatch.delenv("WEB_FETCHER_PATH")
    monkeypatch.setenv("ASSEMBLY_WEB_FETCHER_URL", (_tagged_origin(tmp_path, ac.WEB_FETCHER_TAG)).as_uri())
    dest = ensure_web_fetcher(tmp_path / "cache")
    _git(dest, "config", "user.email", "t@t"), _git(dest, "config", "user.name", "t")
    _git(dest, "commit", "-q", "--allow-empty", "-m", "drift")
    with pytest.raises(AssemblyCheckpointError, match="not v0.2.0"):
        ensure_web_fetcher(tmp_path / "cache")


def test_explicit_web_fetcher_path_wins(tmp_path):
    assert ensure_web_fetcher(tmp_path / "cache") == tmp_path / "wf"