loghunter-cli 0.1.0.dev0__py3-none-any.whl

tests/test_cli_help.py

@@ -0,0 +1,342 @@
+"""Per-command help, side-effect-light help short-circuit, and a few other
+parser-surface invariants that don't fit the verb-specific suites.
+
+Key promises locked here:
+  - ``loghunter <verb> --help`` / ``-h`` renders that verb's own generated help.
+  - Help fires BEFORE config load, output-registry lookup, sniff dispatch, or
+    init wizard entry.
+  - ``--help=anything`` and ``-h=anything`` are NOT help — they raise the
+    strict-parser "takes no value" error.
+  - ``loghunter conn.log`` (a real file in CWD) resolves as a path, not as an
+    unknown command.
+  - ``--output=FORMAT`` validates via the registered output handler list.
+"""
+
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+import pytest
+
+from loghunter import cli
+from loghunter.common import config as cfg
+
+
+# ── per-command help renders from the spec ───────────────────────────────────
+
+
+@pytest.mark.parametrize("verb", [
+    "", "beacon", "dns", "syslog", "scan", "duration", "aws",
+    "digest", "export", "init",
+])
+def test_render_verb_help_lists_verb_allowed_flags(verb: str) -> None:
+    """Every flag in a verb's allowed set appears in its rendered help, and
+    no flag from outside the allowed set leaks in."""
+    rendered = cli._render_verb_help(verb)
+    vs = cli._VERBS[verb]
+    for spec in cli._FLAG_LIST:
+        if spec.key in vs.allowed:
+            assert spec.long in rendered, (
+                f"{spec.long} should be in {verb!r} help"
+            )
+            if spec.short:
+                assert f"-{spec.short}" in rendered
+        else:
+            assert spec.long not in rendered, (
+                f"{spec.long} should NOT be in {verb!r} help"
+            )
+
+
+def test_render_verb_help_blob_path_never_appears() -> None:
+    """``blob_path`` is an INTERNAL routing key — must not appear in any
+    rendered help. Padding the spec/allowed-set with it would silently mint
+    an unadvertised ``--blob-path`` ([[feedback-cli-surface-discipline]])."""
+    for verb in cli._VERBS:
+        rendered = cli._render_verb_help(verb)
+        assert "blob_path" not in rendered
+        assert "--blob-path" not in rendered
+        assert "-blob-path" not in rendered
+
+
+def test_init_help_only_lists_help(capsys: pytest.CaptureFixture[str]) -> None:
+    """init's allowed set is ``{help}`` — its rendered help mentions
+    ``--help`` and nothing else from the spec."""
+    rendered = cli._render_verb_help("init")
+    assert "--help" in rendered
+    for spec in cli._FLAG_LIST:
+        if spec.key != "help":
+            assert spec.long not in rendered
+
+
+# ── side-effect-light help: no config load, no sniff, no wizard ──────────────
+
+
+def test_verb_help_does_not_load_config(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+) -> None:
+    """``<verb> --help`` short-circuits BEFORE cfg.load is called."""
+    def _exploding_load(_path=None):  # pragma: no cover — would only run on failure
+        raise RuntimeError("config must not load during help")
+
+    monkeypatch.setattr(cfg, "load", _exploding_load)
+    for argv in (
+        ["beacon", "--help"], ["beacon", "-h"],
+        ["digest", "--help"], ["digest", "-h"],
+        ["export", "--help"], ["init", "--help"],
+        ["--help"], ["-h"],
+    ):
+        rc = cli._main(argv)
+        assert rc == 0
+        capsys.readouterr()
+
+
+def test_init_help_does_not_start_wizard(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+) -> None:
+    """``init -h`` must NOT enter the wizard (no run_init call, no input())."""
+    called = {"wizard": False}
+
+    def _spy_run_init():  # pragma: no cover — would only run on failure
+        called["wizard"] = True
+
+    monkeypatch.setattr("loghunter.cli_init.run_init", _spy_run_init)
+
+    rc = cli._main(["init", "-h"])
+
+    assert rc == 0
+    assert called["wizard"] is False
+    out = capsys.readouterr().out
+    assert "Usage: loghunter init" in out
+
+
+def test_digest_help_does_not_sniff(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+    tmp_path: Path,
+) -> None:
+    """``digest --help`` must NOT call sniff_format_detailed even when a
+    positional is also passed."""
+    called = {"sniffed": False}
+
+    def _spy_sniff(_path):  # pragma: no cover — would only run on failure
+        called["sniffed"] = True
+        raise RuntimeError("sniff must not run during help")
+
+    monkeypatch.setattr(
+        "loghunter.common.loader.sniff_format_detailed", _spy_sniff,
+    )
+    pretend = tmp_path / "anything.log"
+    pretend.write_text("placeholder\n", encoding="utf-8")
+
+    rc = cli._main(["digest", str(pretend), "--help"])
+
+    assert rc == 0
+    assert called["sniffed"] is False
+
+
+def test_help_with_output_bogus_short_circuits_before_registry(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+) -> None:
+    """``--help --output=bogus`` shows usage; the output registry is NEVER
+    consulted — the help short-circuit wins."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+
+    def _exploding_get_handler(_name):  # pragma: no cover
+        raise RuntimeError("registry must not run during help")
+
+    monkeypatch.setattr("loghunter.cli.get_handler", _exploding_get_handler)
+
+    rc = cli._main(["beacon", "--help", "--output=bogus"])
+
+    assert rc == 0
+
+
+# ── --help=anything / -h=anything are NOT help ───────────────────────────────
+
+
+def test_help_with_value_raises_takes_no_value(
+    capsys: pytest.CaptureFixture[str],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """``--help=foo`` is a value-on-bool error from the strict parser, not
+    a help short-circuit."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    with pytest.raises(SystemExit):
+        cli.main(["--help=foo"])
+    err = capsys.readouterr().err
+    assert "--help (-h) takes no value" in err
+
+
+def test_short_help_with_value_raises_takes_no_value(
+    capsys: pytest.CaptureFixture[str],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    with pytest.raises(SystemExit):
+        cli.main(["-h=foo"])
+    err = capsys.readouterr().err
+    assert "--help (-h) takes no value" in err
+
+
+# ── loghunter conn.log (bare filename in CWD) ────────────────────────────────
+
+
+def test_bare_filename_in_cwd_routes_as_analyze_path(
+    monkeypatch: pytest.MonkeyPatch,
+    tmp_path: Path,
+) -> None:
+    """A token that exists on disk routes to the analyze path (not 'unknown
+    command') even when it lacks the path-shape prefixes."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    monkeypatch.chdir(tmp_path)
+
+    captured: dict[str, object] = {}
+
+    def fake_run(**kwargs: object) -> None:
+        captured.update(kwargs)
+
+    monkeypatch.setattr("loghunter.runner.run", fake_run)
+
+    (tmp_path / "conn.log").write_text("", encoding="utf-8")
+
+    cli._main(["conn.log"])
+
+    # CLI now passes raw strings; the resolver owns Path conversion.
+    assert captured.get("zeek_dir") == "conn.log"
+
+
+# ── --output=FORMAT validation via the registry ──────────────────────────────
+
+
+def test_unknown_output_format_raises_with_available_list(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+    tmp_path: Path,
+) -> None:
+    """``--output=bogus`` raises a CLI-formatted error with the registry's
+    live available-format list, not a hardcoded one."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    zeek_dir = tmp_path / "zeek"
+    zeek_dir.mkdir()
+    (zeek_dir / "conn.log").write_text("", encoding="utf-8")
+
+    with pytest.raises(SystemExit) as exc:
+        cli.main([f"--zeek-dir={zeek_dir}", "--output=bogus"])
+
+    assert exc.value.code == 1
+    err = capsys.readouterr().err
+    assert "Unknown output format 'bogus'." in err
+    assert "Available formats:" in err
+    # Built-in handlers must surface
+    for fmt in ("text", "json", "csv", "html"):
+        assert fmt in err
+
+
+def test_digest_unknown_output_format_uses_same_registry_error(
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+    tmp_path: Path,
+) -> None:
+    """Digest validates --output via the same registry — uniform error voice.
+    Registry check happens BEFORE digest's text-only rail, so ``--output=bogus``
+    reports 'Unknown output format', not 'currently supports only text'."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    log = tmp_path / "x.log"
+    log.write_text("", encoding="utf-8")
+
+    with pytest.raises(SystemExit) as exc:
+        cli.main(["digest", str(log), "--output=bogus"])
+
+    assert exc.value.code == 1
+    err = capsys.readouterr().err
+    assert "Unknown output format 'bogus'." in err
+
+
+# ── export positionals come from the parser ──────────────────────────────────
+
+
+def test_export_positionals_come_from_parser(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """``loghunter export splunk q1 q2`` consumes positionals from the
+    parser's ``paths`` list, not by re-scraping raw args."""
+    captured: dict = {}
+
+    def fake_run_export(**kwargs):
+        captured.update(kwargs)
+
+    monkeypatch.setattr("loghunter.exporters.run_export", fake_run_export)
+    monkeypatch.setattr(cfg, "load", lambda _=None: {
+        "export": {"splunk": {"host": "192.0.2.20", "port": 8089,
+                              "query": {"q1": {"spl": "x"}, "q2": {"spl": "y"}}}},
+    })
+
+    cli.main(["export", "splunk", "q1", "q2"])
+
+    assert captured["backend"] == "splunk"
+    assert captured["query_names"] == ["q1", "q2"]
+
+
+# ── digest combination guard (preserved) ─────────────────────────────────────
+
+
+def test_digest_path_plus_zeek_dir_still_rejected(
+    monkeypatch: pytest.MonkeyPatch,
+    tmp_path: Path,
+) -> None:
+    """``digest PATH --zeek-dir=…`` is rejected (positional self-routes via
+    sniff). Bare digest with --zeek-dir is allowed — that's the bare-conn
+    config-driven path."""
+    monkeypatch.setattr(cfg, "load", lambda _=None: {"loghunter": {}})
+    log = tmp_path / "x.log"
+    log.write_text("", encoding="utf-8")
+    with pytest.raises(ValueError, match="--zeek-dir is not valid alongside"):
+        cli._main(["digest", str(log), "--zeek-dir=/x"])
+
+
+# ── bare short-form value flag mentions both spellings ───────────────────────
+
+
+def test_bare_short_value_flag_short_lead_message(
+    capsys: pytest.CaptureFixture[str],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """A bare short value flag (``-o``) raises the actionable error mentioning
+    both ``-o=…`` and ``--out=…``."""
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    with pytest.raises(SystemExit) as exc:
+        cli.main(["-o"])
+    assert exc.value.code == 1
+    err = capsys.readouterr().err
+    assert "loghunter: --out (-o) needs a value: -o=… or --out=…" in err
+
+
+# ── --detect/-d on single-detector verbs raises wrong-verb ───────────────────
+
+
+def test_detect_on_single_detector_verb_raises_wrong_verb(
+    capsys: pytest.CaptureFixture[str],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    with pytest.raises(SystemExit) as exc:
+        cli.main(["beacon", "--detect=all"])
+    assert exc.value.code == 1
+    err = capsys.readouterr().err
+    assert "--detect (-d) is not valid for beacon" in err
+
+
+def test_short_detect_on_single_detector_verb_raises_wrong_verb(
+    capsys: pytest.CaptureFixture[str],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    monkeypatch.setattr(cfg, "SEARCH_PATHS", [])
+    with pytest.raises(SystemExit) as exc:
+        cli.main(["beacon", "-d=all"])
+    assert exc.value.code == 1
+    err = capsys.readouterr().err
+    assert "-d (--detect) is not valid for beacon" in err