ewoooc/tests/test_external_mcp_rag_integration_service.py

import json
import subprocess
import sys


def test_external_mcp_rag_readback_reports_partial_truth():
    from services.external_mcp_rag_integration_service import (
        POLICY,
        build_external_mcp_rag_integration_readback,
    )

    payload = build_external_mcp_rag_integration_readback()
    inventory = {item["id"]: item for item in payload["inventory"]}

    assert payload["policy"] == POLICY
    assert payload["status"] == "partially_integrated"
    assert payload["completion"]["all_absorbed"] is False
    assert payload["completion"]["unresolved_count"] >= 1
    assert "不是全部完成" in payload["answer_to_owner"]
    assert inventory["rag.pixelrag.visual_evidence"]["status"] == "integrated_phase1_visual_receipts"
    assert inventory["rag.qwen3_vl_embedding"]["status"] == "deferred_until_ollama_first_verified"
    assert inventory["rag.faiss_visual_index"]["status"] == "rejected_for_production_without_adr"
    assert inventory["mcp.gemini_grounding"]["status"] == "disabled_by_default_fallback_only"
    assert payload["controlled_apply"]["db_write"] is False
    assert payload["controlled_apply"]["network_call"] is False


def test_external_mcp_rag_readback_can_filter_family():
    from services.external_mcp_rag_integration_service import (
        build_external_mcp_rag_integration_readback,
    )

    payload = build_external_mcp_rag_integration_readback(target_family="external_rag")

    assert payload["completion"]["total_capabilities"] == 3
    assert {item["family"] for item in payload["inventory"]} == {"external_rag"}


def test_external_mcp_rag_cli_outputs_machine_readable_json():
    completed = subprocess.run(
        [
            sys.executable,
            "scripts/ops/report_external_mcp_rag_integration.py",
            "--family",
            "external_rag",
        ],
        capture_output=True,
        check=False,
        text=True,
    )

    assert completed.returncode == 0
    payload = json.loads(completed.stdout)
    assert payload["success"] is True
    assert payload["completion"]["total_capabilities"] == 3
    assert payload["inventory"][0]["family"] == "external_rag"


def test_external_mcp_rag_ai_automation_route_returns_readback():
    from flask import Flask
    from routes import system_public_routes as routes

    app = Flask(__name__)
    with app.test_request_context(
        "/api/ai-automation/external-mcp-rag-integration?family=external_mcp"
    ):
        response = routes.ai_automation_external_mcp_rag_integration_api.__wrapped__()
    payload = response.get_json()

    assert payload["policy"] == "read_only_external_mcp_rag_integration_readback_v1"
    assert payload["success"] is True
    assert {item["family"] for item in payload["inventory"]} == {"external_mcp"}
    assert payload["runtime"]["mcp"]["caller_count"] >= 1


def test_public_federation_receipt_preserves_two_product_identities(monkeypatch):
    from services import external_mcp_rag_integration_service as service

    monkeypatch.setattr(service, "_mcp_runtime_snapshot", lambda: {
        "enabled": True,
        "servers": {"firecrawl": "internal", "postgres": "internal"},
        "caller_count": 2,
        "tool_registry": {
            "collector": {"firecrawl": ["scrape_url"]},
            "diagnostics": {"postgres": ["query"]},
        },
    })
    monkeypatch.setattr(service, "_rag_runtime_snapshot", lambda: {
        "enabled": True,
        "vector_store": "pgvector",
        "embedding_model": "bge-m3:latest",
        "embedding_dim": 1024,
        "embedding_signature": "signature-v1",
    })
    monkeypatch.setattr(service, "_pixelrag_snapshot", lambda: {
        "enabled": True,
        "platform_count": 2,
        "platforms": ["shopee", "coupang"],
        "visual_rag_stage": "phase1_visual_evidence_receipts",
    })

    payload = service.build_public_mcp_federation_receipt(
        runtime_version="V10.796",
        observed_at="2026-07-15T10:00:00+08:00",
    )
    receipts = {item["product_id"]: item for item in payload["receipts"]}

    assert payload["receipt_count"] == 2
    assert set(receipts) == {"ewoooc", "momo-pro-system"}
    assert receipts["ewoooc"]["canonical_repo"] == "wooo/ewoooc"
    assert receipts["momo-pro-system"]["canonical_repo"] == "wooo/momo-pro-system"
    assert payload["status"] == "partial_degraded"
    assert "rag_embedding_model_not_immutable" in payload["blockers"]
    assert receipts["ewoooc"]["runtime"]["mcp"]["tool_count"] == 2
    assert receipts["ewoooc"]["operation_boundaries"]["db_write_allowed"] is False
    assert receipts["ewoooc"]["operation_boundaries"]["secret_read_allowed"] is False
    for receipt in receipts.values():
        assert (
            service.compute_public_federation_receipt_fingerprint(receipt)
            == receipt["integrity"]["normalized_fingerprint_sha256"]
        )


def test_public_federation_receipt_does_not_expose_internal_endpoints(monkeypatch):
    from services import external_mcp_rag_integration_service as service

    monkeypatch.setattr(service, "_mcp_runtime_snapshot", lambda: {
        "enabled": True,
        "servers": {"firecrawl": "http://192.168.0.188:3002"},
        "caller_count": 1,
        "tool_registry": {"collector": {"firecrawl": ["scrape_url"]}},
    })
    payload = service.build_public_mcp_federation_receipt(
        runtime_version="V10.796",
        observed_at="2026-07-15T10:00:00+08:00",
    )
    serialized = json.dumps(payload, ensure_ascii=False)

    assert "192.168." not in serialized
    assert "http://" not in serialized
    assert "https://" not in serialized
    assert "scrape_url" not in serialized


def test_public_federation_route_is_anonymous_aggregate_only(monkeypatch):
    from flask import Flask
    from routes import system_public_routes as routes
    from services import external_mcp_rag_integration_service as service
    from services.http_access_policy_service import INTENTIONALLY_PUBLIC_ENDPOINTS

    monkeypatch.setattr(service, "build_public_mcp_federation_receipt", lambda **_: {
        "success": True,
        "schema_version": service.PUBLIC_FEDERATION_SCHEMA_VERSION,
        "policy": service.PUBLIC_FEDERATION_POLICY,
        "status": "ready",
        "receipt_count": 2,
        "receipts": [
            {"product_id": "ewoooc"},
            {"product_id": "momo-pro-system"},
        ],
    })
    app = Flask(__name__)
    app.register_blueprint(routes.system_public_bp)

    response = app.test_client().get("/api/public/mcp-federation-readback")
    payload = response.get_json()

    assert response.status_code == 200
    assert payload["receipt_count"] == 2
    assert {item["product_id"] for item in payload["receipts"]} == {
        "ewoooc",
        "momo-pro-system",
    }
    assert (
        "system_public.public_mcp_federation_readback_api"
        in INTENTIONALLY_PUBLIC_ENDPOINTS
    )