repo-scoping/tests/test_registry_service.py

import subprocess

from repo_registry.core.service import RegistryService
from repo_registry.repo_ingestion.git import GitIngestionService
from repo_registry.storage.sqlite import NotFoundError, RegistryStore


def make_service(tmp_path):
    store = RegistryStore(tmp_path / "registry.sqlite3")
    store.initialize()
    return RegistryService(store, ingestion=GitIngestionService(tmp_path / "checkouts"))


def test_manual_registry_builds_ability_map(tmp_path):
    service = make_service(tmp_path)

    repository = service.register_repository(
        name="MailRouter",
        url="https://example.com/mail-router.git",
        description="Routes incoming customer email",
    )
    ability_id = service.add_ability(
        repository.id,
        name="Business Email Routing",
        description="Route inbound messages to the right department.",
        confidence=0.92,
    )
    capability_id = service.add_capability(
        repository.id,
        ability_id,
        name="Classify Incoming Email",
        description="Classify messages into intent categories.",
        inputs=["subject", "body"],
        outputs=["intent", "confidence"],
        confidence=0.88,
    )
    service.add_feature(
        repository.id,
        capability_id,
        name="POST /api/classify-email",
        type="REST endpoint",
        location="src/routes/classify_email.py",
        confidence=0.84,
    )
    service.add_evidence(
        repository.id,
        capability_id,
        type="unit_test",
        reference="tests/test_email_classification.py",
        strength="strong",
    )

    ability_map = service.ability_map(repository.id)

    assert ability_map.repository.name == "MailRouter"
    assert ability_map.abilities[0].name == "Business Email Routing"
    capability = ability_map.abilities[0].capabilities[0]
    assert capability.name == "Classify Incoming Email"
    assert capability.inputs == ["subject", "body"]
    assert capability.features[0].location == "src/routes/classify_email.py"
    assert capability.evidence[0].strength == "strong"


def test_search_matches_approved_abilities_and_capabilities(tmp_path):
    service = make_service(tmp_path)
    repository = service.register_repository(
        name="MailRouter",
        url="https://example.com/mail-router.git",
    )
    ability_id = service.add_ability(
        repository.id,
        name="Business Email Routing",
        description="Route inbound messages.",
    )
    service.add_capability(
        repository.id,
        ability_id,
        name="Classify Incoming Email",
        description="Classify messages into intent categories.",
    )

    results = service.search("classify")

    assert len(results) == 1
    assert results[0].repository_name == "MailRouter"
    assert results[0].match_type == "capability"
    assert results[0].match_name == "Classify Incoming Email"


def test_capability_must_belong_to_repository(tmp_path):
    service = make_service(tmp_path)
    first = service.register_repository(name="First", url="https://example.com/first.git")
    second = service.register_repository(name="Second", url="https://example.com/second.git")
    ability_id = service.add_ability(first.id, name="Document Classification")

    try:
        service.add_capability(second.id, ability_id, name="Classify Document")
    except NotFoundError as exc:
        assert "ability" in str(exc)
    else:
        raise AssertionError("expected a NotFoundError")


def test_analyze_repository_records_snapshot_and_observed_facts(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Example\n", encoding="utf-8")
    (source / "requirements.txt").write_text("fastapi\n", encoding="utf-8")
    (source / "app.py").write_text(
        "from fastapi import FastAPI\n"
        "app = FastAPI()\n"
        '@app.get("/health")\n'
        "def health():\n"
        "    return {}\n",
        encoding="utf-8",
    )

    service = make_service(tmp_path)
    repository = service.register_repository(
        name="Example",
        url=str(source),
        description="A local fixture repository",
    )

    summary = service.analyze_repository(repository.id)

    assert summary.analysis_run.status == "completed"
    assert summary.snapshot is not None
    assert summary.snapshot.file_count == 3
    assert service.get_repository(repository.id).status == "analyzed"
    fact_names = {(fact.kind, fact.name, fact.path) for fact in summary.facts}
    assert ("documentation", "README", "README.md") in fact_names
    assert ("framework", "FastAPI", "requirements.txt") in fact_names
    assert ("interface", "python route decorator", "app.py") in fact_names


def test_analyze_repository_failure_is_recorded(tmp_path):
    service = make_service(tmp_path)
    repository = service.register_repository(
        name="Missing",
        url=str(tmp_path / "does-not-exist"),
    )

    summary = service.analyze_repository(repository.id)

    assert summary.analysis_run.status == "failed"
    assert summary.snapshot is None
    assert "does not exist" in (summary.analysis_run.error_message or "")
    assert service.get_repository(repository.id).status == "analysis_failed"


def test_analyze_repository_clones_git_url_before_scanning(tmp_path):
    source = tmp_path / "git-source"
    source.mkdir()
    subprocess.run(["git", "init", "-b", "main"], cwd=source, check=True)
    subprocess.run(
        ["git", "config", "user.email", "tests@example.com"],
        cwd=source,
        check=True,
    )
    subprocess.run(
        ["git", "config", "user.name", "Tests"],
        cwd=source,
        check=True,
    )
    (source / "README.md").write_text("# Git Source\n", encoding="utf-8")
    (source / "requirements.txt").write_text("pytest\n", encoding="utf-8")
    subprocess.run(["git", "add", "."], cwd=source, check=True)
    subprocess.run(["git", "commit", "-m", "initial"], cwd=source, check=True)

    service = make_service(tmp_path)
    repository = service.register_repository(name="Git Source", url=source.as_uri())

    summary = service.analyze_repository(repository.id)

    assert summary.analysis_run.status == "completed"
    assert summary.snapshot is not None
    assert str(tmp_path / "checkouts") in summary.snapshot.source_path
    fact_names = {(fact.kind, fact.name, fact.path) for fact in summary.facts}
    assert ("documentation", "README", "README.md") in fact_names
    assert ("framework", "pytest", "requirements.txt") in fact_names