repo-scoping/tests/test_candidate_graph.py

from repo_registry.candidate_graph.generator import CandidateGraphGenerator
from repo_registry.core.models import ContentChunk, ObservedFact, Repository


def fact(id, kind, name, path="", value=""):
    return ObservedFact(
        id=id,
        repository_id=1,
        analysis_run_id=1,
        snapshot_id=1,
        kind=kind,
        path=path,
        name=name,
        value=value,
        metadata={},
    )


def chunk(id, kind, path, text, start_line=1, end_line=1):
    return ContentChunk(
        id=id,
        repository_id=1,
        analysis_run_id=1,
        snapshot_id=1,
        path=path,
        kind=kind,
        start_line=start_line,
        end_line=end_line,
        text=text,
    )


def test_candidate_generator_builds_purpose_seed_from_observed_facts():
    repository = Repository(
        id=1,
        name="MailRouter",
        url="/tmp/mail-router",
        description="Routes incoming customer email to the right team.",
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        fact(2, "interface", "python route decorator", "app.py", '@app.post("/classify")'),
        fact(3, "test", "test_app.py", "tests/test_app.py"),
        fact(4, "framework", "FastAPI", "requirements.txt"),
    ]

    graph = CandidateGraphGenerator().generate(repository, facts)

    assert len(graph) == 1
    ability = graph[0]
    assert ability.name == "Route Incoming Customer Email To The Right Team"
    assert "Usefulness" not in ability.name
    assert ability.source_refs[0].path == "README.md"
    interface_capability = ability.capabilities[0]
    assert interface_capability.name == "Expose Repository Interface"
    assert interface_capability.confidence == 0.75
    assert interface_capability.inputs == ["HTTP request"]
    assert interface_capability.outputs == ["HTTP response"]
    assert interface_capability.features[0].type == "API"
    assert interface_capability.features[0].name == "POST /classify"
    assert interface_capability.features[0].location == "app.py"
    assert interface_capability.evidence[0].strength == "strong"


def test_candidate_generator_enriches_descriptions_from_content_chunks():
    repository = Repository(
        id=1,
        name="MailRouter",
        url="/tmp/mail-router",
        description=None,
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        fact(2, "interface", "python route decorator", "app.py", '@app.post("/classify")'),
    ]
    chunks = [
        chunk(
            1,
            "documentation",
            "README.md",
            "# MailRouter\nRoutes incoming customer email to the right team.",
            end_line=2,
        ),
        chunk(
            2,
            "interface",
            "app.py",
            '@app.post("/classify")\ndef classify_email():\n    return {}',
            start_line=5,
            end_line=7,
        ),
    ]

    graph = CandidateGraphGenerator().generate(repository, facts, chunks)

    assert graph[0].name == "Route Incoming Customer Email To The Right Team"
    assert "MailRouter. Routes incoming customer email" in graph[0].description
    assert '@app.post("/classify")' in graph[0].capabilities[0].description


def test_candidate_confidence_scoring_stays_conservative_for_weak_facts():
    repository = Repository(
        id=1,
        name="WeakDocs",
        url="/tmp/weak-docs",
        description=None,
        branch="main",
        status="analyzed",
    )

    graph = CandidateGraphGenerator().generate(
        repository,
        [fact(1, "documentation", "README", "README.md")],
    )

    assert graph[0].confidence == 0.45
    assert graph[0].capabilities == []


def test_candidate_confidence_scoring_increases_with_supporting_facts():
    repository = Repository(
        id=1,
        name="Supported",
        url="/tmp/supported",
        description=None,
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        fact(2, "interface", "python route decorator", "app.py", '@app.get("/health")'),
        fact(3, "test", "test_app.py", "tests/test_app.py"),
        fact(4, "example", "client.py", "examples/client.py"),
        fact(5, "framework", "FastAPI", "requirements.txt"),
        fact(6, "language", "Python"),
        fact(7, "manifest", "requirements.txt", "requirements.txt"),
    ]

    graph = CandidateGraphGenerator().generate(repository, facts)

    assert graph[0].confidence == 1.0
    assert graph[0].capabilities[0].confidence == 0.85
    assert graph[0].capabilities[1].confidence == 0.75


def test_candidate_generator_names_cli_features_from_nearby_function():
    repository = Repository(
        id=1,
        name="CliTool",
        url="/tmp/cli-tool",
        description=None,
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        ObservedFact(
            id=2,
            repository_id=1,
            analysis_run_id=1,
            snapshot_id=1,
            kind="interface",
            path="cli.py",
            name="python CLI command decorator",
            value="@click.command()",
            metadata={"line": 3},
        ),
    ]
    chunks = [
        chunk(
            1,
            "interface",
            "cli.py",
            "@click.command()\ndef import_repositories():\n    pass",
            start_line=3,
            end_line=5,
        )
    ]

    graph = CandidateGraphGenerator().generate(repository, facts, chunks)

    feature = graph[0].capabilities[0].features[0]
    capability = graph[0].capabilities[0]
    assert feature.type == "CLI"
    assert feature.name == "CLI command import_repositories"
    assert capability.inputs == ["CLI arguments"]
    assert capability.outputs == ["command output"]


def test_candidate_generator_uses_generic_io_for_unknown_interfaces():
    repository = Repository(
        id=1,
        name="UnknownInterface",
        url="/tmp/unknown-interface",
        description=None,
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        fact(2, "interface", "possible surface", "src/plugin.py"),
    ]

    graph = CandidateGraphGenerator().generate(repository, facts)

    capability = graph[0].capabilities[0]
    assert capability.inputs == ["caller input"]
    assert capability.outputs == ["callable interface result"]


def test_candidate_generator_groups_many_interface_facts_into_behavioral_features():
    repository = Repository(
        id=1,
        name="Registry",
        url="/tmp/registry",
        description=None,
        branch="main",
        status="analyzed",
    )
    facts = [
        fact(1, "documentation", "README", "README.md"),
        fact(2, "interface", "python route decorator", "src/api.py", '@app.get("/repos")'),
        fact(3, "interface", "python route decorator", "src/api.py", '@app.post("/repos")'),
        fact(
            4,
            "interface",
            "python route decorator",
            "src/api.py",
            '@app.post("/repos/{repository_id}/analysis-runs")',
        ),
        fact(5, "test", "test_api.py", "tests/test_api.py"),
    ]

    graph = CandidateGraphGenerator().generate(repository, facts)

    capability = graph[0].capabilities[0]
    assert len(capability.features) == 1
    feature = capability.features[0]
    assert feature.name == (
        "HTTP API surface: GET /repos, POST /repos, POST /repos/{repository_id}/analysis-runs"
    )
    assert feature.type == "API"
    assert feature.location == "src/api.py"
    assert len(feature.source_refs) == 3