CIAgent/apps/api/tests/unit/test_monitoring_smoke.py

"""Monitoring run endpoints, exercised via HTTP with CELERY_TASK_ALWAYS_EAGER
so `.delay()` runs the task synchronously in-process (see app/tasks/base.py
for why that needs a threaded asyncio bridge to work from an async route
handler). No live network: every collector's discovery/collection call is
respx-mocked."""

from __future__ import annotations

import json
import uuid
from unittest.mock import patch

import httpx
import pytest
import respx


def _register_and_login(client) -> dict[str, str]:
    email = f"user-{uuid.uuid4().hex[:12]}@example.com"
    client.post(
        "/api/v1/auth/register",
        json={"email": email, "password": "correct-horse-1", "display_name": "Test User"},
    )
    tokens = client.post(
        "/api/v1/auth/login", json={"email": email, "password": "correct-horse-1"}
    ).json()
    return {"Authorization": f"Bearer {tokens['access_token']}"}


def _create_company(client, headers, **overrides):
    payload = {
        "name": f"Run Test Co {uuid.uuid4().hex[:6]}",
        "frequency_type": "weekly",
        **overrides,
    }
    return client.post("/api/v1/companies", json=payload, headers=headers).json()


def _mock_empty_discovery():
    """Every discoverable collector type returns nothing (or a harmless
    stub), so a first run doesn't need per-source mocks for whatever
    discovery happens to find. RSS (Google News) and GOV_CONTRACT
    (USASpending) are always discovered unconditionally - see
    discovery_service.py's _PREVIEWABLE_TYPES - so their collect() calls
    need mocking here too, unlike GitHub/SEC which discover.() itself."""
    respx.get("https://api.github.com/search/users").mock(
        return_value=httpx.Response(200, text=json.dumps({"items": []}))
    )
    respx.get("https://www.sec.gov/cgi-bin/browse-edgar").mock(
        return_value=httpx.Response(
            200, text='<?xml version="1.0"?><feed xmlns="http://www.w3.org/2005/Atom"></feed>'
        )
    )
    respx.get(url__regex=r"https://news\.google\.com/rss/search.*").mock(
        return_value=httpx.Response(
            200,
            text=(
                '<?xml version="1.0"?><rss version="2.0"><channel><title>Google News</title>'
                "<item><title>Test</title><link>https://example.com/news</link>"
                "<description>Test news item</description></item></channel></rss>"
            ),
        )
    )
    respx.post("https://api.usaspending.gov/api/v2/search/spending_by_award/").mock(
        return_value=httpx.Response(200, json={"results": []})
    )


def test_run_now_executes_and_reports_final_status(client):
    headers = _register_and_login(client)
    company = _create_company(client, headers)

    with patch("socket.getaddrinfo", return_value=[(2, 1, 6, "", ("93.184.216.34", 0))]):
        with respx.mock:
            _mock_empty_discovery()
            resp = client.post(f"/api/v1/companies/{company['id']}/run", headers=headers)

    assert resp.status_code == 202
    run = resp.json()
    assert run["trigger_type"] == "manual"

    # Eager mode means the task already finished by the time .delay() returns.
    detail = client.get(f"/api/v1/runs/{run['id']}", headers=headers).json()
    assert detail["status"] == "successful"
    assert detail["completed_at"] is not None


async def test_run_now_is_idempotent_for_an_active_run(db_session, settings):
    """A run-now call while one is already queued/running returns that same
    run instead of enqueuing a duplicate. Exercised at the service layer
    directly: under CELERY_TASK_ALWAYS_EAGER a `.delay()` call runs the task
    to completion before returning, so an HTTP-level test can never observe
    an in-flight run to dedupe against."""
    from app.models.company import Company
    from app.models.enums import MonitoringRunTrigger
    from app.models.monitor_configuration import MonitorConfiguration
    from app.repositories.monitoring_run_repository import MonitoringRunRepository
    from app.services import monitoring_service

    user_id = uuid.uuid4()
    company = Company(
        id=uuid.uuid4(),
        user_id=user_id,
        name="Idempotency Co",
        slug=f"idempotency-co-{uuid.uuid4().hex[:6]}",
    )
    db_session.add(company)
    db_session.add(MonitorConfiguration(company_id=company.id))
    await db_session.commit()

    run_repo = MonitoringRunRepository(db_session)
    existing = await run_repo.create(
        company_id=company.id, trigger_type=MonitoringRunTrigger.SCHEDULED
    )
    await db_session.commit()

    with patch("app.tasks.collection.run_monitoring.delay") as mock_delay:
        result = await monitoring_service.enqueue_run_now(db_session, settings, user_id, company.id)

    assert result.id == existing.id
    mock_delay.assert_not_called()


async def test_run_now_enforces_daily_manual_run_cap(db_session, settings):
    from app.core.errors import RateLimitedError
    from app.models.company import Company
    from app.models.enums import MonitoringRunStatus, MonitoringRunTrigger
    from app.models.monitor_configuration import MonitorConfiguration
    from app.repositories.monitoring_run_repository import MonitoringRunRepository
    from app.services import monitoring_service

    user_id = uuid.uuid4()
    company = Company(
        id=uuid.uuid4(),
        user_id=user_id,
        name="Rate Limited Co",
        slug=f"rate-limited-co-{uuid.uuid4().hex[:6]}",
    )
    db_session.add(company)
    db_session.add(MonitorConfiguration(company_id=company.id))
    await db_session.commit()

    run_repo = MonitoringRunRepository(db_session)
    for _ in range(2):
        run = await run_repo.create(company_id=company.id, trigger_type=MonitoringRunTrigger.MANUAL)
        await run_repo.mark_finished(run, status=MonitoringRunStatus.SUCCESSFUL, error_summary=None)
    await db_session.commit()

    capped_settings = settings.model_copy(update={"max_manual_runs_per_day": 2})

    with pytest.raises(RateLimitedError):
        await monitoring_service.enqueue_run_now(db_session, capped_settings, user_id, company.id)


def test_run_history_and_single_run_scoped_to_owner(client):
    owner_headers = _register_and_login(client)
    other_headers = _register_and_login(client)
    company = _create_company(client, owner_headers)

    with patch("socket.getaddrinfo", return_value=[(2, 1, 6, "", ("93.184.216.34", 0))]):
        with respx.mock:
            _mock_empty_discovery()
            run = client.post(
                f"/api/v1/companies/{company['id']}/run", headers=owner_headers
            ).json()

    assert (
        client.get(f"/api/v1/companies/{company['id']}/runs", headers=other_headers).status_code
        == 404
    )
    assert client.get(f"/api/v1/runs/{run['id']}", headers=other_headers).status_code == 404
    assert client.get(f"/api/v1/runs/{run['id']}", headers=owner_headers).status_code == 200


def test_successful_first_run_generates_a_baseline_report(client):
    headers = _register_and_login(client)
    company = _create_company(client, headers)
    client.post(
        f"/api/v1/companies/{company['id']}/sources",
        json={
            "source_type": "custom_url",
            "name": "Pricing",
            "base_url": "https://example.com/pricing",
        },
        headers=headers,
    )

    with patch("socket.getaddrinfo", return_value=[(2, 1, 6, "", ("93.184.216.34", 0))]):
        with respx.mock:
            _mock_empty_discovery()
            respx.get("https://example.com/robots.txt").mock(return_value=httpx.Response(404))
            respx.get("https://example.com/pricing").mock(
                return_value=httpx.Response(
                    200,
                    html="<html><head><title>Pricing</title></head><body>"
                    "<article><h1>Pricing</h1><p>Plans start at $10/month.</p></article>"
                    "</body></html>",
                )
            )
            client.post(f"/api/v1/companies/{company['id']}/run", headers=headers)

    reports = client.get(f"/api/v1/companies/{company['id']}/reports", headers=headers).json()
    assert len(reports) == 1
    assert reports[0]["report_type"] == "baseline"
    assert reports[0]["model_provider"] == "mock"


def test_manual_run_does_not_disrupt_next_scheduled_run(client):
    headers = _register_and_login(client)
    company = _create_company(client, headers)
    original_next_run = company["monitor_configuration"]["next_run"]

    with patch("socket.getaddrinfo", return_value=[(2, 1, 6, "", ("93.184.216.34", 0))]):
        with respx.mock:
            _mock_empty_discovery()
            client.post(f"/api/v1/companies/{company['id']}/run", headers=headers)

    refreshed = client.get(f"/api/v1/companies/{company['id']}", headers=headers).json()
    assert refreshed["monitor_configuration"]["next_run"] == original_next_run
    assert refreshed["monitor_configuration"]["last_run"] is not None