import json
import logging
import subprocess

from repo_registry.core.logging import LOGGER_NAME
from repo_registry.core.service import RegistryService
from repo_registry.llm_extraction import ExtractedAbility, ExtractedCapability
from repo_registry.repo_ingestion.git import GitIngestionService
from repo_registry.semantic import HashingEmbeddingProvider
from repo_registry.storage.sqlite import NotFoundError, RegistryStore
from tests.fixtures import (
    write_empty_repo,
    write_javascript_typescript_package_repo,
    write_misleading_docs_repo,
    write_python_cli_repo,
    write_readme_only_repo,
)


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


class FakeLLMExtractor:
    def __init__(self, abilities):
        self.abilities = abilities
        self.calls = []

    def extract(self, repository, chunks):
        self.calls.append((repository, chunks))
        return self.abilities


class FailingLLMExtractor:
    def extract(self, repository, chunks):
        raise RuntimeError("provider unavailable")


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_manual_registry_updates_and_deletes_approved_entries(tmp_path):
    service = make_service(tmp_path)
    repository = service.register_repository(
        name="Manual",
        url="https://example.com/manual.git",
        description="Manual registry fixture.",
    )
    ability_id = service.add_ability(repository.id, name="Original Ability")
    capability_id = service.add_capability(
        repository.id,
        ability_id,
        name="Original Capability",
    )
    feature_id = service.add_feature(
        repository.id,
        capability_id,
        name="Original Feature",
        type="API",
    )
    evidence_id = service.add_evidence(
        repository.id,
        capability_id,
        type="test",
        reference="tests/test_original.py",
    )

    service.update_ability(repository.id, ability_id, name="Updated Ability")
    service.update_capability(
        repository.id,
        capability_id,
        name="Updated Capability",
        inputs=["request"],
        outputs=["response"],
    )
    service.update_feature(repository.id, feature_id, location="src/api.py")
    ability_map = service.update_evidence(
        repository.id,
        evidence_id,
        strength="strong",
    )

    ability = ability_map.abilities[0]
    capability = ability.capabilities[0]
    assert ability.name == "Updated Ability"
    assert capability.name == "Updated Capability"
    assert capability.inputs == ["request"]
    assert capability.outputs == ["response"]
    assert capability.features[0].location == "src/api.py"
    assert capability.evidence[0].strength == "strong"

    service.delete_feature(repository.id, feature_id)
    service.delete_evidence(repository.id, evidence_id)
    ability_map = service.delete_capability(repository.id, capability_id)
    assert ability_map.abilities[0].capabilities == []

    ability_map = service.delete_ability(repository.id, ability_id)
    assert ability_map.abilities == []


def test_repository_update_and_delete(tmp_path):
    service = make_service(tmp_path)
    repository = service.register_repository(
        name="Original",
        url="https://example.com/original.git",
        description="Original description.",
    )
    ability_id = service.add_ability(repository.id, name="Original Ability")

    updated = service.update_repository(
        repository.id,
        name="Updated",
        description="Updated description.",
        branch="develop",
    )

    assert updated.name == "Updated"
    assert updated.description == "Updated description."
    assert updated.branch == "develop"
    assert service.ability_map(repository.id).abilities[0].id == ability_id

    service.delete_repository(repository.id)

    try:
        service.get_repository(repository.id)
    except NotFoundError as exc:
        assert "repository" in str(exc)
    else:
        raise AssertionError("expected a NotFoundError")


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",
        description="Manual test repository.",
    )
    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"

    abilities = service.list_abilities()
    capabilities = service.list_capabilities()

    assert abilities[0].repository_name == "MailRouter"
    assert abilities[0].name == "Business Email Routing"
    assert capabilities[0].ability_name == "Business Email Routing"
    assert capabilities[0].name == "Classify Incoming Email"


def test_search_matches_features_and_evidence_with_context(tmp_path):
    service = make_service(tmp_path)
    repository = service.register_repository(
        name="MailRouter",
        url="https://example.com/mail-router-feature.git",
        description="Manual test repository.",
    )
    ability_id = service.add_ability(repository.id, name="Business Email Routing")
    capability_id = service.add_capability(
        repository.id,
        ability_id,
        name="Classify Incoming Email",
    )
    service.add_feature(
        repository.id,
        capability_id,
        name="POST /api/classify-email",
        type="REST endpoint",
        location="src/routes/classify_email.py",
    )
    service.add_evidence(
        repository.id,
        capability_id,
        type="unit_test",
        reference="tests/test_email_classification.py",
        strength="strong",
    )

    feature_results = service.search("classify_email")
    evidence_results = service.search("unit_test")

    assert feature_results[0].match_type == "feature"
    assert feature_results[0].matched_field == "location"
    assert feature_results[0].ability_name == "Business Email Routing"
    assert feature_results[0].capability_name == "Classify Incoming Email"
    assert feature_results[0].source_reference == "src/routes/classify_email.py"
    assert evidence_results[0].match_type == "evidence"
    assert evidence_results[0].evidence_level == "strong"
    assert evidence_results[0].confidence == 0.9


def test_search_filters_by_status_language_and_framework(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Filterable\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="Filterable", url=str(source))
    summary = service.analyze_repository(repository.id)
    service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    results = service.search(
        "repository",
        status="indexed",
        language="Python",
        framework="FastAPI",
        ability="Support Filterable",
        capability="Repository Structure",
    )
    wrong_language_results = service.search(
        "repository",
        status="indexed",
        language="TypeScript",
        framework="FastAPI",
    )
    wrong_capability_results = service.search(
        "repository",
        status="indexed",
        language="Python",
        framework="FastAPI",
        capability="Email Routing",
    )

    assert results
    assert {result.repository_name for result in results} == {"Filterable"}
    assert wrong_language_results == []
    assert wrong_capability_results == []


def test_fixture_breadth_readme_only_repo_stays_conservative(tmp_path):
    source = write_readme_only_repo(tmp_path)
    service = make_service(tmp_path)
    repository = service.register_repository(name="Readme Only", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)

    assert summary.analysis_run.status == "completed"
    assert graph.abilities[0].confidence == 0.45
    assert graph.abilities[0].capabilities == []
    assert service.ability_map(repository.id).abilities == []


def test_fixture_breadth_python_cli_repo_extracts_reviewable_cli_claims(tmp_path):
    source = write_python_cli_repo(tmp_path)
    service = make_service(tmp_path)
    repository = service.register_repository(name="Python CLI", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)

    capability = graph.abilities[0].capabilities[0]
    assert summary.analysis_run.status == "completed"
    assert capability.name == "Expose Repository Interface"
    assert capability.features[0].type == "CLI"
    assert capability.features[0].name.startswith("CLI command surface:")
    assert capability.evidence[0].reference == "tests/test_cli.py"
    assert service.ability_map(repository.id).abilities == []


def test_fixture_breadth_javascript_typescript_package_extracts_structure_and_api(tmp_path):
    source = write_javascript_typescript_package_repo(tmp_path)
    service = make_service(tmp_path)
    repository = service.register_repository(name="JS TS Package", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)

    fact_names = {(fact.kind, fact.name, fact.path) for fact in summary.facts}
    capability_names = {
        capability.name
        for ability in graph.abilities
        for capability in ability.capabilities
    }
    feature_types = {
        feature.type
        for ability in graph.abilities
        for capability in ability.capabilities
        for feature in capability.features
    }
    assert ("language", "TypeScript", "") in fact_names
    assert ("framework", "React", "package.json") in fact_names
    assert ("framework", "Vitest", "package.json") in fact_names
    assert "Expose Repository Interface" in capability_names
    assert "Describe Repository Structure" in capability_names
    assert "API" in feature_types
    assert service.ability_map(repository.id).abilities == []


def test_fixture_breadth_misleading_docs_do_not_become_approved_truth(tmp_path):
    source = write_misleading_docs_repo(tmp_path)
    service = make_service(tmp_path)
    repository = service.register_repository(name="Misleading Docs", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    ability_map = service.ability_map(repository.id)

    assert summary.analysis_run.status == "completed"
    assert graph.abilities[0].confidence == 0.45
    assert graph.abilities[0].capabilities == []
    assert ability_map.abilities == []


def test_fixture_breadth_empty_repo_produces_no_candidate_claims(tmp_path):
    source = write_empty_repo(tmp_path)
    service = make_service(tmp_path)
    repository = service.register_repository(name="Empty Repo", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)

    assert summary.analysis_run.status == "completed"
    assert summary.snapshot is not None
    assert summary.snapshot.file_count == 0
    assert summary.facts == []
    assert graph.abilities == []
    assert service.ability_map(repository.id).abilities == []


def test_semantic_search_adds_hybrid_matches_without_changing_text_default(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text(
        "# Queue Worker\n\nHandles postponed customer jobs.\n",
        encoding="utf-8",
    )

    text_service = make_service(tmp_path)
    repository = text_service.register_repository(
        name="Queue Worker",
        url=str(source),
        description="Processes deferred customer work.",
    )
    ability_id = text_service.add_ability(
        repository.id,
        name="Background Job Processing",
        description="Run deferred work outside request handling.",
        confidence=0.8,
    )
    capability_id = text_service.add_capability(
        repository.id,
        ability_id,
        name="Process Customer Tasks",
        description="Execute queued customer tasks asynchronously.",
        confidence=0.7,
    )
    text_service.add_feature(
        repository.id,
        capability_id,
        name="worker task loop",
        type="background worker",
        location="worker.py",
        confidence=0.6,
    )
    text_service.analyze_repository(repository.id)

    assert text_service.search("customer queued") == []

    semantic_service = RegistryService(
        text_service.store,
        ingestion=GitIngestionService(tmp_path / "checkouts"),
        embedding_provider=HashingEmbeddingProvider(),
    )
    results = semantic_service.search("customer queued")

    assert results
    assert results[0].match_type in {"capability", "content_chunk"}
    assert results[0].matched_field == "semantic"
    assert results[0].vector_score > 0
    assert results[0].hybrid_score >= results[0].vector_score * 0.35
    assert any(result.match_type == "content_chunk" for result in results)


def test_register_repository_imports_metadata_when_name_is_omitted(tmp_path):
    source = tmp_path / "metadata-source"
    source.mkdir()
    (source / "pyproject.toml").write_text(
        '[project]\nname = "metadata-source"\ndescription = "Imported description."\n',
        encoding="utf-8",
    )

    service = make_service(tmp_path)

    repository = service.register_repository(url=str(source))

    assert repository.name == "metadata-source"
    assert repository.description == "Imported description."


def test_operational_logging_records_analysis_and_review_events(tmp_path, caplog):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Logged\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)

    with caplog.at_level(logging.INFO, logger=LOGGER_NAME):
        repository = service.register_repository(name="Logged", url=str(source))
        summary = service.analyze_repository(repository.id)
        service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    events = [json.loads(record.message)["event"] for record in caplog.records]
    assert "repository_registered" in events
    assert "analysis_started" in events
    assert "analysis_completed" in events
    assert "review_decision_recorded" in events


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",
        description="Manual first repository.",
    )
    second = service.register_repository(
        name="Second",
        url="https://example.com/second.git",
        description="Manual second repository.",
    )
    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
    chunks = service.list_content_chunks(repository.id, summary.analysis_run.id)
    chunk_sources = {(chunk.kind, chunk.path) for chunk in chunks}
    assert ("documentation", "README.md") in chunk_sources
    assert ("manifest", "requirements.txt") in chunk_sources
    assert ("interface", "app.py") in chunk_sources

    candidate_graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    assert candidate_graph.repository.name == "Example"
    assert candidate_graph.abilities
    assert "Example" in candidate_graph.abilities[0].description
    assert "@app.get" in candidate_graph.abilities[0].capabilities[0].description
    assert candidate_graph.abilities[0].capabilities[0].features[0].name == "GET /health"
    capability_names = {
        capability.name
        for ability in candidate_graph.abilities
        for capability in ability.capabilities
    }
    assert "Expose Repository Interface" in capability_names


def test_analyze_repository_can_use_optional_llm_extractor(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text(
        "# Email Router\nRoutes incoming customer email.\n",
        encoding="utf-8",
    )
    store = RegistryStore(tmp_path / "registry.sqlite3")
    store.initialize()
    extractor = FakeLLMExtractor(
        [
            ExtractedAbility(
                name="Business Email Routing",
                description="Route incoming messages.",
                source_paths=["README.md"],
                capabilities=[
                    ExtractedCapability(
                        name="Classify Incoming Email",
                        description="Classify messages by intent.",
                        source_paths=["README.md"],
                    )
                ],
            )
        ]
    )
    service = RegistryService(
        store,
        ingestion=GitIngestionService(tmp_path / "checkouts"),
        llm_extractor=extractor,
    )
    repository = service.register_repository(name="Email Router", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    decisions = service.list_review_decisions(repository.id, summary.analysis_run.id)

    assert extractor.calls
    assert extractor.calls[0][1]
    assert graph.abilities[0].name == "Business Email Routing"
    assert graph.abilities[0].capabilities[0].name == "Classify Incoming Email"
    assert graph.abilities[0].source_refs[0].path == "README.md"
    assert decisions[0].action == "llm_extraction_used"
    assert "1 candidate ability" in decisions[0].notes


def test_analyze_repository_falls_back_when_optional_llm_extractor_returns_no_candidates(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Fallback\n", encoding="utf-8")
    store = RegistryStore(tmp_path / "registry.sqlite3")
    store.initialize()
    service = RegistryService(
        store,
        ingestion=GitIngestionService(tmp_path / "checkouts"),
        llm_extractor=FakeLLMExtractor([]),
    )
    repository = service.register_repository(name="Fallback", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)

    assert graph.abilities[0].name == "Support Fallback"


def test_analyze_repository_records_llm_failure_and_falls_back(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Failing LLM\n", encoding="utf-8")
    store = RegistryStore(tmp_path / "registry.sqlite3")
    store.initialize()
    service = RegistryService(
        store,
        ingestion=GitIngestionService(tmp_path / "checkouts"),
        llm_extractor=FailingLLMExtractor(),
    )
    repository = service.register_repository(name="Failing LLM", url=str(source))

    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    decisions = service.list_review_decisions(repository.id, summary.analysis_run.id)

    assert summary.analysis_run.status == "completed"
    assert graph.abilities[0].name == "Support Failing LLM"
    assert decisions[0].action == "llm_extraction_failed"
    assert "provider unavailable" in decisions[0].notes


def test_approve_candidate_graph_publishes_ability_map_once(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Example\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",
    )
    (source / "cli.py").write_text(
        "import click\n\n"
        "@click.command()\n"
        "def health():\n"
        "    click.echo('ok')\n",
        encoding="utf-8",
    )

    service = make_service(tmp_path)
    repository = service.register_repository(name="Example", url=str(source))
    summary = service.analyze_repository(repository.id)

    ability_map = service.approve_candidate_graph(
        repository.id,
        summary.analysis_run.id,
        notes="Looks good for the first pass.",
    )
    second_approval = service.approve_candidate_graph(
        repository.id,
        summary.analysis_run.id,
    )

    assert service.get_repository(repository.id).status == "indexed"
    assert len(ability_map.abilities) == 1
    assert len(second_approval.abilities) == 1
    assert ability_map.abilities[0].name == "Support Example"
    assert ability_map.abilities[0].capabilities[0].features[0].location == "app.py"
    assert ability_map.abilities[0].capabilities[0].features[0].source_refs
    assert ability_map.abilities[0].capabilities[0].features[0].source_refs[0].line == 3
    assert ability_map.abilities[0].capabilities[0].evidence[0].source_refs

    candidate_graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    assert candidate_graph.abilities[0].status == "approved"
    decisions = service.list_review_decisions(repository.id, summary.analysis_run.id)
    assert decisions[0].action == "approve_candidate_graph"
    assert decisions[0].notes == "Looks good for the first pass."


def test_analysis_run_diff_keeps_approved_map_stable_until_change_approval(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Example\n", encoding="utf-8")
    app_file = source / "app.py"
    app_file.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))
    first_summary = service.analyze_repository(repository.id)
    approved_before = service.approve_candidate_graph(
        repository.id,
        first_summary.analysis_run.id,
    )

    app_file.write_text(
        "from fastapi import FastAPI\n"
        "app = FastAPI()\n"
        '@app.get("/ready")\n'
        "def ready():\n"
        "    return {}\n",
        encoding="utf-8",
    )
    second_summary = service.analyze_repository(repository.id)
    approved_after_analysis = service.ability_map(repository.id)
    diff = service.diff_analysis_runs(
        repository.id,
        first_summary.analysis_run.id,
        second_summary.analysis_run.id,
    )

    assert approved_after_analysis.abilities[0].capabilities[0].features[0].name == (
        approved_before.abilities[0].capabilities[0].features[0].name
    )
    assert any(item.item_type == "feature" for item in diff.candidates.added)
    assert any(item.item_type == "feature" for item in diff.candidates.removed)
    assert any(item.item_type == "feature" for item in diff.approved_entries.added)
    assert any(item.item_type == "feature" for item in diff.approved_entries.removed)

    approved_after_review = service.approve_analysis_run_changes(
        repository.id,
        second_summary.analysis_run.id,
        notes="Accept route change.",
    )

    assert approved_after_review.abilities[0].capabilities[0].features[0].name == (
        "GET /ready"
    )
    decisions = service.list_review_decisions(
        repository.id,
        second_summary.analysis_run.id,
    )
    assert decisions[0].action == "approve_analysis_run_changes"
    assert decisions[0].notes == "Accept route change."


def test_reject_candidate_ability_excludes_it_from_approval(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Rejectable\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="Rejectable", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    candidate = graph.abilities[0]

    rejected_graph = service.reject_candidate_ability(
        repository.id,
        summary.analysis_run.id,
        candidate.id,
        notes="Too generic.",
    )
    ability_map = service.approve_candidate_graph(
        repository.id,
        summary.analysis_run.id,
    )

    assert service.get_repository(repository.id).status == "reviewing"
    assert rejected_graph.abilities[0].status == "rejected"
    assert rejected_graph.abilities[0].capabilities[0].status == "rejected"
    assert rejected_graph.abilities[0].capabilities[0].features[0].status == "rejected"
    assert ability_map.abilities == []


def test_edit_candidate_graph_values_before_approval(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Editable\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="Editable", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    candidate_ability = graph.abilities[0]
    candidate_capability = candidate_ability.capabilities[0]

    service.edit_candidate_ability(
        repository.id,
        summary.analysis_run.id,
        candidate_ability.id,
        name="Service Health Monitoring",
        description="Expose health state for operational monitoring.",
        confidence=0.91,
        notes="Curator renamed the generic ability.",
    )
    service.edit_candidate_capability(
        repository.id,
        summary.analysis_run.id,
        candidate_capability.id,
        name="Report HTTP Health",
        description="Return a lightweight health response over HTTP.",
        confidence=0.87,
    )

    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    assert service.get_repository(repository.id).status == "indexed"
    assert ability_map.abilities[0].name == "Service Health Monitoring"
    assert ability_map.abilities[0].description == (
        "Expose health state for operational monitoring."
    )
    assert ability_map.abilities[0].confidence == 0.91
    assert ability_map.abilities[0].capabilities[0].name == "Report HTTP Health"
    assert ability_map.abilities[0].capabilities[0].confidence == 0.87


def test_reject_candidate_capability_excludes_it_from_approval(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Capability Reject\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="Capability Reject", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    candidate_capability = graph.abilities[0].capabilities[0]

    rejected_graph = service.reject_candidate_capability(
        repository.id,
        summary.analysis_run.id,
        candidate_capability.id,
        notes="Interface is not relevant.",
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    assert rejected_graph.abilities[0].capabilities[0].status == "rejected"
    assert rejected_graph.abilities[0].capabilities[0].features[0].status == "rejected"
    approved_capability_names = {
        capability.name for capability in ability_map.abilities[0].capabilities
    }
    assert candidate_capability.name not in approved_capability_names


def test_reject_candidate_feature_and_evidence_excludes_only_those_leaves(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Leaf Reject\n", encoding="utf-8")
    (source / "tests").mkdir()
    (source / "tests" / "test_health.py").write_text(
        "def test_health(): pass\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="Leaf Reject", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    capability = graph.abilities[0].capabilities[0]

    service.reject_candidate_feature(
        repository.id,
        summary.analysis_run.id,
        capability.features[0].id,
        notes="Feature is incidental.",
    )
    service.reject_candidate_evidence(
        repository.id,
        summary.analysis_run.id,
        capability.evidence[0].id,
        notes="Evidence is too weak.",
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    approved_capability = ability_map.abilities[0].capabilities[0]
    assert approved_capability.name == capability.name
    assert approved_capability.features == []
    assert len(approved_capability.evidence) == len(capability.evidence) - 1


def test_relink_candidate_capability_to_another_ability_before_approval(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Relink Capability\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="Relink Capability", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    capability = graph.abilities[0].capabilities[0]
    with service.store.connect() as connection:
        cursor = connection.execute(
            """
            INSERT INTO candidate_abilities
              (repository_id, analysis_run_id, name, description, confidence)
            VALUES (?, ?, ?, ?, ?)
            """,
            (
                repository.id,
                summary.analysis_run.id,
                "Operations Visibility",
                "Curator-created target ability.",
                0.72,
            ),
        )
        target_ability_id = int(cursor.lastrowid)

    relinked_graph = service.relink_candidate_capability(
        repository.id,
        summary.analysis_run.id,
        capability.id,
        target_ability_id=target_ability_id,
        notes="Move interface under the operational ability.",
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    target_candidate = [
        ability for ability in relinked_graph.abilities if ability.id == target_ability_id
    ][0]
    assert target_candidate.capabilities[0].id == capability.id
    approved_target = [
        ability for ability in ability_map.abilities if ability.name == "Operations Visibility"
    ][0]
    assert approved_target.capabilities[0].name == capability.name


def test_relink_candidate_feature_and_evidence_to_another_capability(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Relink Leaves\n", encoding="utf-8")
    (source / "requirements.txt").write_text("fastapi\n", encoding="utf-8")
    (source / "tests").mkdir()
    (source / "tests" / "test_health.py").write_text(
        "def test_health(): pass\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="Relink Leaves", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    source_capability = graph.abilities[0].capabilities[0]
    target_capability = graph.abilities[0].capabilities[1]
    feature = source_capability.features[0]
    evidence = source_capability.evidence[0]

    service.relink_candidate_feature(
        repository.id,
        summary.analysis_run.id,
        feature.id,
        target_capability_id=target_capability.id,
    )
    service.relink_candidate_evidence(
        repository.id,
        summary.analysis_run.id,
        evidence.id,
        target_capability_id=target_capability.id,
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    approved_capabilities = {
        capability.name: capability for capability in ability_map.abilities[0].capabilities
    }
    assert approved_capabilities[source_capability.name].features == []
    assert feature.name in {
        item.name for item in approved_capabilities[target_capability.name].features
    }
    assert evidence.reference in {
        item.reference for item in approved_capabilities[target_capability.name].evidence
    }


def test_merge_candidate_ability_moves_capabilities_to_target(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Merge Ability\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="Merge Ability", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    source_ability = graph.abilities[0]
    with service.store.connect() as connection:
        cursor = connection.execute(
            """
            INSERT INTO candidate_abilities
              (repository_id, analysis_run_id, name, description, confidence)
            VALUES (?, ?, ?, ?, ?)
            """,
            (
                repository.id,
                summary.analysis_run.id,
                "Merged Operational Ability",
                "Preferred duplicate ability.",
                0.83,
            ),
        )
        target_ability_id = int(cursor.lastrowid)

    graph = service.merge_candidate_ability(
        repository.id,
        summary.analysis_run.id,
        source_ability.id,
        target_ability_id=target_ability_id,
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    merged_source = [ability for ability in graph.abilities if ability.id == source_ability.id][0]
    target = [ability for ability in graph.abilities if ability.id == target_ability_id][0]
    assert merged_source.status == "merged"
    assert target.capabilities
    assert [ability.name for ability in ability_map.abilities] == [
        "Merged Operational Ability"
    ]
    assert ability_map.abilities[0].capabilities


def test_merge_candidate_capability_moves_children_to_target(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Merge Capability\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="Merge Capability", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    source_capability = graph.abilities[0].capabilities[0]
    target_capability = graph.abilities[0].capabilities[1]

    graph = service.merge_candidate_capability(
        repository.id,
        summary.analysis_run.id,
        source_capability.id,
        target_capability_id=target_capability.id,
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    merged_source = [
        capability
        for ability in graph.abilities
        for capability in ability.capabilities
        if capability.id == source_capability.id
    ][0]
    target = [
        capability
        for ability in graph.abilities
        for capability in ability.capabilities
        if capability.id == target_capability.id
    ][0]
    assert merged_source.status == "merged"
    assert target.features
    assert [capability.name for capability in ability_map.abilities[0].capabilities] == [
        target_capability.name
    ]
    assert ability_map.abilities[0].capabilities[0].features


def test_merge_candidate_feature_and_evidence_omits_duplicate_leaves(tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Merge Leaves\n", encoding="utf-8")
    (source / "tests").mkdir()
    (source / "tests" / "test_health.py").write_text(
        "def test_health(): pass\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"
        '@app.get("/ready")\n'
        "def ready():\n"
        "    return {}\n",
        encoding="utf-8",
    )
    (source / "cli.py").write_text(
        "import click\n\n"
        "@click.command()\n"
        "def health():\n"
        "    click.echo('ok')\n",
        encoding="utf-8",
    )

    service = make_service(tmp_path)
    repository = service.register_repository(name="Merge Leaves", url=str(source))
    summary = service.analyze_repository(repository.id)
    graph = service.candidate_graph(repository.id, summary.analysis_run.id)
    capability = graph.abilities[0].capabilities[0]

    service.merge_candidate_feature(
        repository.id,
        summary.analysis_run.id,
        capability.features[1].id,
        target_feature_id=capability.features[0].id,
    )
    service.merge_candidate_evidence(
        repository.id,
        summary.analysis_run.id,
        capability.evidence[1].id,
        target_evidence_id=capability.evidence[0].id,
    )
    ability_map = service.approve_candidate_graph(repository.id, summary.analysis_run.id)

    approved_capability = ability_map.abilities[0].capabilities[0]
    assert len(approved_capability.features) == len(capability.features) - 1
    assert len(approved_capability.evidence) == len(capability.evidence) - 1


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"),
        description="Manual missing repository.",
    )

    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


def test_analyze_repository_can_use_cached_checkout_without_fetching(tmp_path, monkeypatch):
    service = make_service(tmp_path)
    url = "https://example.com/private/repo.git"
    cached = tmp_path / "checkouts" / "repo-b5d250ec3c59"
    cached.mkdir(parents=True)
    (cached / "README.md").write_text("# Cached Repo\n", encoding="utf-8")

    def fail_run_git(*args, **kwargs):
        raise AssertionError("cached analysis should not run git")

    monkeypatch.setattr(service.ingestion, "_run_git", fail_run_git)
    repository = service.register_repository(
        name="Cached",
        url=url,
        description="Already cloned.",
    )

    summary = service.analyze_repository(
        repository.id,
        use_cached_checkout=True,
    )

    assert summary.analysis_run.status == "completed"
    assert summary.snapshot is not None
    assert str(cached) == summary.snapshot.source_path
    assert ("documentation", "README", "README.md") in {
        (fact.kind, fact.name, fact.path) for fact in summary.facts
    }


def test_operational_logging_records_analysis_and_review_events(caplog, tmp_path):
    source = tmp_path / "repo"
    source.mkdir()
    (source / "README.md").write_text("# Logged Service\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)
    caplog.set_level(logging.INFO, logger=LOGGER_NAME)

    repository = service.register_repository(name="Logged", url=str(source))
    summary = service.analyze_repository(repository.id)
    service.approve_candidate_graph(
        repository.id,
        summary.analysis_run.id,
        notes="Logged approval.",
    )

    payloads = [
        json.loads(record.message)
        for record in caplog.records
        if record.name == LOGGER_NAME
    ]
    events = {payload["event"] for payload in payloads}

    assert "repository_registered" in events
    assert "analysis_started" in events
    assert "analysis_completed" in events
    assert "review_decision_recorded" in events
    assert all(payload["repository_id"] == repository.id for payload in payloads)