factory/lib/interns-mcp/tests/test_grep_audit.py

"""Tests for grep_audit intern — deterministic, LLM-free."""

import json

from interns_mcp.interns.base import Intern, InternResponse
from interns_mcp.interns.grep_audit import GrepAudit


def test_substring_case_sensitive(tmp_path):
    f = tmp_path / "claude.md"
    f.write_text("follow project discipline\nuse superpowers\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=["follow project discipline", "Follow Project Discipline"],
        output="json",
    )

    assert isinstance(result, InternResponse)
    data = json.loads(result.text)
    matches = data["rows"][0]["matches"]
    assert matches["follow project discipline"] is True
    assert matches["Follow Project Discipline"] is False


def test_substring_case_insensitive(tmp_path):
    f = tmp_path / "claude.md"
    f.write_text("follow project discipline\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=["FOLLOW PROJECT DISCIPLINE"],
        output="json",
        case_sensitive=False,
    )

    data = json.loads(result.text)
    assert data["rows"][0]["matches"]["FOLLOW PROJECT DISCIPLINE"] is True


def test_regex_named(tmp_path):
    f = tmp_path / "pyproject.toml"
    f.write_text('version = "0.3.0"\n', encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=[{"pattern": r'version\s*=\s*"\d+\.\d+\.\d+"', "name": "version-line", "regex": True}],
        output="json",
    )

    data = json.loads(result.text)
    assert data["rows"][0]["matches"]["version-line"] is True


def test_dict_substring(tmp_path):
    f = tmp_path / "skill.md"
    f.write_text("trigger string here\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=[{"pattern": "trigger string", "name": "trigger"}],
        output="json",
    )

    data = json.loads(result.text)
    assert data["rows"][0]["matches"]["trigger"] is True


def test_output_table(tmp_path):
    f1 = tmp_path / "a.md"
    f1.write_text("alpha beta\n", encoding="utf-8")
    f2 = tmp_path / "b.md"
    f2.write_text("gamma\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f1), str(f2)],
        patterns=["alpha", "gamma"],
        output="table",
    )

    text = result.text
    assert text.splitlines()[0].startswith("| Path |")
    assert "alpha" in text.splitlines()[0]
    assert "gamma" in text.splitlines()[0]
    # f1 has alpha but not gamma
    f1_line = [l for l in text.splitlines() if str(f1) in l][0]
    assert "✅" in f1_line
    assert "❌" in f1_line
    # f2 has gamma but not alpha
    f2_line = [l for l in text.splitlines() if str(f2) in l][0]
    assert "✅" in f2_line
    assert "❌" in f2_line


def test_output_json(tmp_path):
    f = tmp_path / "a.md"
    f.write_text("alpha\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=["alpha", "missing"],
        output="json",
    )

    data = json.loads(result.text)
    assert "rows" in data
    assert len(data["rows"]) == 1
    assert data["rows"][0]["path"] == str(f)
    assert data["rows"][0]["matches"]["alpha"] is True
    assert data["rows"][0]["matches"]["missing"] is False


def test_file_not_found_partial_result(tmp_path):
    f_real = tmp_path / "exists.md"
    f_real.write_text("hello\n", encoding="utf-8")
    f_ghost = tmp_path / "ghost.md"  # never created

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f_real), str(f_ghost)],
        patterns=["hello"],
        output="json",
    )

    data = json.loads(result.text)
    assert len(data["rows"]) == 2  # partial result, no abort
    rows_by_path = {r["path"]: r for r in data["rows"]}
    assert rows_by_path[str(f_real)]["matches"]["hello"] is True
    # Ghost row: match is None (unreadable cell), error field present
    assert rows_by_path[str(f_ghost)]["matches"]["hello"] is None
    assert "error" in rows_by_path[str(f_ghost)]


def test_file_not_found_table_renders_warning(tmp_path):
    f_ghost = tmp_path / "ghost.md"  # never created

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f_ghost)],
        patterns=["x"],
        output="table",
    )

    assert "⚠️" in result.text


def test_empty_paths():
    intern = GrepAudit()
    result = intern.run(paths=[], patterns=["x"], output="json")
    data = json.loads(result.text)
    assert data["rows"] == []
    assert result.usage["files_scanned"] == 0
    assert result.usage["matches_total"] == 0


def test_empty_patterns(tmp_path):
    f = tmp_path / "a.md"
    f.write_text("alpha\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(paths=[str(f)], patterns=[], output="json")
    data = json.loads(result.text)
    assert len(data["rows"]) == 1
    assert data["rows"][0]["matches"] == {}
    assert result.usage["patterns_evaluated"] == 0


def test_unicode_content_utf8(tmp_path):
    f = tmp_path / "ru.md"
    f.write_text("русский текст\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f)],
        patterns=["русский"],
        output="json",
    )

    data = json.loads(result.text)
    assert data["rows"][0]["matches"]["русский"] is True


def test_unicode_content_binary_garbage_does_not_crash(tmp_path):
    f = tmp_path / "binary.bin"
    f.write_bytes(b"\xff\xfe\x00\x01\x02 hello \xff")

    intern = GrepAudit()
    # errors="replace" should let us still grep for "hello"
    result = intern.run(
        paths=[str(f)],
        patterns=["hello"],
        output="json",
    )

    data = json.loads(result.text)
    assert data["rows"][0]["matches"]["hello"] is True


def test_usage_counts(tmp_path):
    f1 = tmp_path / "a.md"
    f1.write_text("alpha\n", encoding="utf-8")
    f2 = tmp_path / "b.md"
    f2.write_text("alpha beta\n", encoding="utf-8")

    intern = GrepAudit()
    result = intern.run(
        paths=[str(f1), str(f2)],
        patterns=["alpha", "beta"],
        output="json",
    )

    assert result.usage["files_scanned"] == 2
    assert result.usage["patterns_evaluated"] == 2
    # f1: alpha ✓, beta ✗  → 1
    # f2: alpha ✓, beta ✓  → 2
    assert result.usage["matches_total"] == 3


def test_endpoint_null_no_client_no_crash():
    """LLM-free intern must not require client wiring."""
    intern = GrepAudit(config={"endpoint": None})
    assert intern.client is None
    assert intern.endpoint is None


def test_base_intern_accepts_endpoint_null_config():
    """Base class survives config with explicit endpoint=null (no LLM init triggered)."""
    intern = Intern(config={"endpoint": None, "model": None})
    assert intern.endpoint is None
    assert intern.model is None
    assert intern.client is None