learningcircuit--local-deep-research/tests/web/routes/test_context_overflow_api.py

"""
Tests for web/routes/context_overflow_api.py

Tests cover:
- get_context_overflow_metrics endpoint
- get_research_context_overflow endpoint
- Various time period filters
- Error handling
"""

import pytest
from unittest.mock import Mock
from datetime import datetime, timezone


class TestGetContextOverflowMetricsUnit:
    """Unit tests for get_context_overflow_metrics logic."""

    def test_calculate_truncation_rate(self):
        """Test truncation rate calculation."""
        # Test the calculation logic
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        requests_with_context = 100
        truncated_requests = 15

        truncation_rate = 0
        if requests_with_context > 0:
            truncation_rate = (truncated_requests / requests_with_context) * 100

        assert truncation_rate == 15.0

    def test_calculate_truncation_rate_zero_division(self):
        """Test truncation rate with zero requests."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        requests_with_context = 0
        truncated_requests = 0

        truncation_rate = 0
        if requests_with_context > 0:
            truncation_rate = (truncated_requests / requests_with_context) * 100

        assert truncation_rate == 0


class TestTimePeriodCalculation:
    """Tests for time period calculation logic."""

    def test_7d_period(self):
        """Test 7 day period calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        from datetime import timedelta

        now = datetime.now(timezone.utc)
        period = "7d"

        start_date = None
        if period != "all":
            if period == "7d":
                start_date = now - timedelta(days=7)

        assert start_date is not None
        assert (now - start_date).days == 7

    def test_30d_period(self):
        """Test 30 day period calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        from datetime import timedelta

        now = datetime.now(timezone.utc)
        period = "30d"

        start_date = None
        if period != "all":
            if period == "30d":
                start_date = now - timedelta(days=30)

        assert start_date is not None
        assert (now - start_date).days == 30

    def test_3m_period(self):
        """Test 3 month period calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        from datetime import timedelta

        now = datetime.now(timezone.utc)
        period = "3m"

        start_date = None
        if period != "all":
            if period == "3m":
                start_date = now - timedelta(days=90)

        assert start_date is not None
        assert (now - start_date).days == 90

    def test_1y_period(self):
        """Test 1 year period calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        from datetime import timedelta

        now = datetime.now(timezone.utc)
        period = "1y"

        start_date = None
        if period != "all":
            if period == "1y":
                start_date = now - timedelta(days=365)

        assert start_date is not None
        assert (now - start_date).days == 365

    def test_all_period(self):
        """Test all time period has no date filter."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        period = "all"

        start_date = None
        if period != "all":
            start_date = datetime.now(timezone.utc)

        assert start_date is None

    def test_invalid_period_defaults_to_30d(self):
        """Test that an invalid period value is normalized to 30d."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        VALID_PERIODS = {"7d", "30d", "3m", "1y", "all"}

        for invalid in ["999d", "abc", "", "10x", "<script>"]:
            period = invalid
            if period not in VALID_PERIODS:
                period = "30d"
            assert period == "30d"

    def test_valid_periods_are_not_overridden(self):
        """Test that valid period values pass through unchanged."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        VALID_PERIODS = {"7d", "30d", "3m", "1y", "all"}

        for valid in ["7d", "30d", "3m", "1y", "all"]:
            period = valid
            if period not in VALID_PERIODS:
                period = "30d"
            assert period == valid


class TestChartDataFormatting:
    """Tests for chart data formatting logic."""

    def test_format_chart_data_basic(self):
        """Test basic chart data formatting."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage = Mock()
        mock_usage.timestamp = datetime(
            2025, 1, 1, 12, 0, 0, tzinfo=timezone.utc
        )
        mock_usage.research_id = "test-research-1"
        mock_usage.prompt_tokens = 1000
        mock_usage.ollama_prompt_eval_count = None
        mock_usage.context_truncated = False
        mock_usage.tokens_truncated = 0
        mock_usage.context_limit = 4096
        mock_usage.model_name = "test-model"

        # Simulate the chart data creation logic
        ollama_used = mock_usage.ollama_prompt_eval_count
        actual_prompt = ollama_used or mock_usage.prompt_tokens
        tokens_truncated = mock_usage.tokens_truncated or 0
        original_tokens = (
            actual_prompt + tokens_truncated
            if mock_usage.context_truncated
            else actual_prompt
        )

        chart_item = {
            "timestamp": mock_usage.timestamp.isoformat(),
            "research_id": mock_usage.research_id,
            "prompt_tokens": mock_usage.prompt_tokens,
            "ollama_prompt_tokens": ollama_used,
            "original_prompt_tokens": original_tokens,
            "context_limit": mock_usage.context_limit,
            "truncated": bool(mock_usage.context_truncated),
            "tokens_truncated": tokens_truncated,
            "model": mock_usage.model_name,
        }

        assert chart_item["prompt_tokens"] == 1000
        assert chart_item["original_prompt_tokens"] == 1000
        assert chart_item["truncated"] is False

    def test_format_chart_data_with_truncation(self):
        """Test chart data formatting with truncation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage = Mock()
        mock_usage.timestamp = datetime(
            2025, 1, 1, 12, 0, 0, tzinfo=timezone.utc
        )
        mock_usage.research_id = "test-research-1"
        mock_usage.prompt_tokens = 5000
        mock_usage.ollama_prompt_eval_count = 4000
        mock_usage.context_truncated = True
        mock_usage.tokens_truncated = 1000
        mock_usage.context_limit = 4096
        mock_usage.model_name = "test-model"

        ollama_used = mock_usage.ollama_prompt_eval_count
        actual_prompt = ollama_used or mock_usage.prompt_tokens
        tokens_truncated = mock_usage.tokens_truncated or 0
        original_tokens = (
            actual_prompt + tokens_truncated
            if mock_usage.context_truncated
            else actual_prompt
        )

        chart_item = {
            "timestamp": mock_usage.timestamp.isoformat(),
            "research_id": mock_usage.research_id,
            "prompt_tokens": mock_usage.prompt_tokens,
            "ollama_prompt_tokens": ollama_used,
            "original_prompt_tokens": original_tokens,
            "context_limit": mock_usage.context_limit,
            "truncated": bool(mock_usage.context_truncated),
            "tokens_truncated": tokens_truncated,
            "model": mock_usage.model_name,
        }

        assert chart_item["truncated"] is True
        assert chart_item["tokens_truncated"] == 1000
        assert chart_item["original_prompt_tokens"] == 5000  # 4000 + 1000


class TestPhaseStatsCalculation:
    """Tests for phase stats calculation logic."""

    def test_calculate_phase_stats_single_phase(self):
        """Test phase stats for single phase."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage = Mock()
        mock_usage.research_phase = "analysis"
        mock_usage.prompt_tokens = 1000
        mock_usage.completion_tokens = 200
        mock_usage.total_tokens = 1200
        mock_usage.context_truncated = False

        token_usage = [mock_usage]

        phase_stats = {}
        for req in token_usage:
            phase = req.research_phase or "unknown"
            if phase not in phase_stats:
                phase_stats[phase] = {
                    "count": 0,
                    "prompt_tokens": 0,
                    "completion_tokens": 0,
                    "total_tokens": 0,
                    "truncated_count": 0,
                }
            phase_stats[phase]["count"] += 1
            phase_stats[phase]["prompt_tokens"] += req.prompt_tokens or 0
            phase_stats[phase]["completion_tokens"] += (
                req.completion_tokens or 0
            )
            phase_stats[phase]["total_tokens"] += req.total_tokens or 0
            if req.context_truncated:
                phase_stats[phase]["truncated_count"] += 1

        assert "analysis" in phase_stats
        assert phase_stats["analysis"]["count"] == 1
        assert phase_stats["analysis"]["total_tokens"] == 1200

    def test_calculate_phase_stats_multiple_phases(self):
        """Test phase stats for multiple phases."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage1 = Mock()
        mock_usage1.research_phase = "analysis"
        mock_usage1.prompt_tokens = 1000
        mock_usage1.completion_tokens = 200
        mock_usage1.total_tokens = 1200
        mock_usage1.context_truncated = False

        mock_usage2 = Mock()
        mock_usage2.research_phase = "synthesis"
        mock_usage2.prompt_tokens = 2000
        mock_usage2.completion_tokens = 400
        mock_usage2.total_tokens = 2400
        mock_usage2.context_truncated = True

        token_usage = [mock_usage1, mock_usage2]

        phase_stats = {}
        for req in token_usage:
            phase = req.research_phase or "unknown"
            if phase not in phase_stats:
                phase_stats[phase] = {
                    "count": 0,
                    "prompt_tokens": 0,
                    "completion_tokens": 0,
                    "total_tokens": 0,
                    "truncated_count": 0,
                }
            phase_stats[phase]["count"] += 1
            phase_stats[phase]["prompt_tokens"] += req.prompt_tokens or 0
            phase_stats[phase]["completion_tokens"] += (
                req.completion_tokens or 0
            )
            phase_stats[phase]["total_tokens"] += req.total_tokens or 0
            if req.context_truncated:
                phase_stats[phase]["truncated_count"] += 1

        assert "analysis" in phase_stats
        assert "synthesis" in phase_stats
        assert phase_stats["synthesis"]["truncated_count"] == 1

    def test_calculate_phase_stats_unknown_phase(self):
        """Test phase stats handles unknown phase."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage = Mock()
        mock_usage.research_phase = None
        mock_usage.prompt_tokens = 500
        mock_usage.completion_tokens = 100
        mock_usage.total_tokens = 600
        mock_usage.context_truncated = False

        token_usage = [mock_usage]

        phase_stats = {}
        for req in token_usage:
            phase = req.research_phase or "unknown"
            if phase not in phase_stats:
                phase_stats[phase] = {
                    "count": 0,
                    "prompt_tokens": 0,
                    "completion_tokens": 0,
                    "total_tokens": 0,
                    "truncated_count": 0,
                }
            phase_stats[phase]["count"] += 1
            phase_stats[phase]["prompt_tokens"] += req.prompt_tokens or 0
            phase_stats[phase]["completion_tokens"] += (
                req.completion_tokens or 0
            )
            phase_stats[phase]["total_tokens"] += req.total_tokens or 0
            if req.context_truncated:
                phase_stats[phase]["truncated_count"] += 1

        assert "unknown" in phase_stats


class TestOverviewCalculation:
    """Tests for overview metrics calculation."""

    def test_calculate_overview_basic(self):
        """Test basic overview calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage = Mock()
        mock_usage.total_tokens = 1200
        mock_usage.prompt_tokens = 1000
        mock_usage.completion_tokens = 200
        mock_usage.context_limit = 4096
        mock_usage.context_truncated = False
        mock_usage.tokens_truncated = 0

        token_usage = [mock_usage]

        total_tokens = sum(req.total_tokens or 0 for req in token_usage)
        total_prompt = sum(req.prompt_tokens or 0 for req in token_usage)
        total_completion = sum(
            req.completion_tokens or 0 for req in token_usage
        )
        context_limit = next(
            (req.context_limit for req in token_usage if req.context_limit),
            None,
        )
        truncated_requests = [
            req for req in token_usage if req.context_truncated
        ]
        max_tokens_used = max((req.prompt_tokens or 0) for req in token_usage)

        assert total_tokens == 1200
        assert total_prompt == 1000
        assert total_completion == 200
        assert context_limit == 4096
        assert len(truncated_requests) == 0
        assert max_tokens_used == 1000

    def test_calculate_overview_with_truncation(self):
        """Test overview with truncation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usage1 = Mock()
        mock_usage1.total_tokens = 1200
        mock_usage1.prompt_tokens = 1000
        mock_usage1.completion_tokens = 200
        mock_usage1.context_limit = 4096
        mock_usage1.context_truncated = False
        mock_usage1.tokens_truncated = 0

        mock_usage2 = Mock()
        mock_usage2.total_tokens = 5200
        mock_usage2.prompt_tokens = 5000
        mock_usage2.completion_tokens = 200
        mock_usage2.context_limit = 4096
        mock_usage2.context_truncated = True
        mock_usage2.tokens_truncated = 1000

        token_usage = [mock_usage1, mock_usage2]

        truncated_requests = [
            req for req in token_usage if req.context_truncated
        ]
        tokens_lost = sum(
            req.tokens_truncated or 0 for req in truncated_requests
        )

        assert len(truncated_requests) == 1
        assert tokens_lost == 1000


class TestModelStatsFormatting:
    """Tests for model stats formatting."""

    def test_format_model_stats(self):
        """Test model stats formatting."""
        # Mock a SQLAlchemy result row
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_stat = Mock()
        mock_stat.model_name = "test-model"
        mock_stat.model_provider = "test-provider"
        mock_stat.total_requests = 100
        mock_stat.truncated_count = 10
        mock_stat.avg_context_limit = 4096.0

        model_stats = [mock_stat]

        formatted = []
        for stat in model_stats:
            formatted.append(
                {
                    "model": stat.model_name,
                    "provider": stat.model_provider,
                    "total_requests": stat.total_requests,
                    "truncated_count": int(stat.truncated_count or 0),
                    "truncation_rate": round(
                        (stat.truncated_count or 0) / stat.total_requests * 100,
                        2,
                    )
                    if stat.total_requests > 0
                    else 0,
                    "avg_context_limit": round(stat.avg_context_limit, 0)
                    if stat.avg_context_limit
                    else None,
                }
            )

        assert formatted[0]["model"] == "test-model"
        assert formatted[0]["truncation_rate"] == 10.0
        assert formatted[0]["avg_context_limit"] == 4096

    def test_format_model_stats_with_none_values(self):
        """Test model stats with None values."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_stat = Mock()
        mock_stat.model_name = "test-model"
        mock_stat.model_provider = None
        mock_stat.total_requests = 50
        mock_stat.truncated_count = None
        mock_stat.avg_context_limit = None

        model_stats = [mock_stat]

        formatted = []
        for stat in model_stats:
            formatted.append(
                {
                    "model": stat.model_name,
                    "provider": stat.model_provider,
                    "total_requests": stat.total_requests,
                    "truncated_count": int(stat.truncated_count or 0),
                    "truncation_rate": round(
                        (stat.truncated_count or 0) / stat.total_requests * 100,
                        2,
                    )
                    if stat.total_requests > 0
                    else 0,
                    "avg_context_limit": round(stat.avg_context_limit, 0)
                    if stat.avg_context_limit
                    else None,
                }
            )

        assert formatted[0]["truncated_count"] == 0
        assert formatted[0]["avg_context_limit"] is None


class TestContextOverflowApiRoutes:
    """Tests for context overflow API routes."""

    def test_metrics_url_with_suffix_is_not_registered(self):
        """The metrics endpoint is mounted at /api/context-overflow (no
        /metrics suffix). A GET to /api/context-overflow/metrics
        therefore must 404 — pinning this so a future URL prefix change
        can't silently move the route."""
        from flask import Flask
        from local_deep_research.web.routes.context_overflow_api import (
            context_overflow_bp,
        )

        app = Flask(__name__)
        app.config["SECRET_KEY"] = "test-secret"
        app.register_blueprint(context_overflow_bp)

        with app.test_client() as client:
            response = client.get("/api/context-overflow/metrics")
            assert response.status_code == 404, response.status_code

    def test_research_context_overflow_route_requires_auth(self):
        """The research-context-overflow route is wrapped in
        @login_required; with no auth set up, the decorator must
        short-circuit with 401 (not 200, not 404). Note: the route
        URL is /api/research/<id>/context-overflow (not the inverse
        ordering that the previous test name suggested)."""
        from flask import Flask
        from local_deep_research.web.routes.context_overflow_api import (
            context_overflow_bp,
        )

        app = Flask(__name__)
        app.config["SECRET_KEY"] = "test-secret"
        app.register_blueprint(context_overflow_bp)

        with app.test_client() as client:
            response = client.get("/api/research/123/context-overflow")
            assert response.status_code == 401, response.status_code


class TestContextOverflowBlueprintImport:
    """Tests for context overflow API blueprint import."""

    def test_blueprint_exists(self):
        """Test that context overflow API blueprint exists."""
        from local_deep_research.web.routes.context_overflow_api import (
            context_overflow_bp,
        )

        assert context_overflow_bp is not None
        assert context_overflow_bp.name == "context_overflow_api"


class TestContextUtilizationCalculation:
    """Tests for context utilization calculation."""

    def test_calculate_context_utilization_percentage(self):
        """Test context utilization percentage calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        prompt_tokens = 3000
        context_limit = 4096

        utilization = (prompt_tokens / context_limit) * 100

        assert utilization == pytest.approx(73.24, rel=0.01)

    def test_calculate_context_utilization_at_limit(self):
        """Test context utilization at 100%."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        prompt_tokens = 4096
        context_limit = 4096

        utilization = (prompt_tokens / context_limit) * 100

        assert utilization == 100.0

    def test_calculate_context_utilization_over_limit(self):
        """Test context utilization over 100% (truncation case)."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        prompt_tokens = 5000
        context_limit = 4096

        utilization = (prompt_tokens / context_limit) * 100

        assert utilization > 100.0
        assert utilization == pytest.approx(122.07, rel=0.01)


class TestAverageCalculations:
    """Tests for average context calculations."""

    def test_calculate_average_prompt_tokens(self):
        """Test average prompt tokens calculation."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usages = [
            Mock(prompt_tokens=1000),
            Mock(prompt_tokens=2000),
            Mock(prompt_tokens=3000),
        ]

        total = sum(u.prompt_tokens for u in mock_usages)
        average = total / len(mock_usages)

        assert average == 2000.0

    def test_calculate_average_with_empty_list(self):
        """Test average calculation with empty list."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usages = []

        total = sum(getattr(u, "prompt_tokens", 0) for u in mock_usages)
        average = total / len(mock_usages) if mock_usages else 0

        assert average == 0


class TestResearchIdExtraction:
    """Tests for research ID extraction logic."""

    def test_extract_unique_research_ids(self):
        """Test extracting unique research IDs from usages."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usages = [
            Mock(research_id="research1"),
            Mock(research_id="research2"),
            Mock(research_id="research1"),  # Duplicate
            Mock(research_id="research3"),
        ]

        unique_ids = list(set(u.research_id for u in mock_usages))

        assert len(unique_ids) == 3
        assert "research1" in unique_ids
        assert "research2" in unique_ids
        assert "research3" in unique_ids

    def test_extract_research_ids_with_none(self):
        """Test extracting research IDs with None values."""
        # audit: PUNCHLIST reviewed 2026-05 — issue resolved by prior PR (recommendation: DELETE).
        mock_usages = [
            Mock(research_id="research1"),
            Mock(research_id=None),
            Mock(research_id="research2"),
        ]

        unique_ids = list(
            set(u.research_id for u in mock_usages if u.research_id)
        )

        assert len(unique_ids) == 2
        assert None not in unique_ids


class TestTokenStatsAggregation:
    """Tests for token statistics aggregation."""

    def test_aggregate_total_tokens_by_model(self):
        """Test aggregating total tokens by model."""
        mock_usages = [
            Mock(model_name="gpt-4", total_tokens=1000),
            Mock(model_name="gpt-4", total_tokens=2000),
            Mock(model_name="claude-3", total_tokens=1500),
        ]

        model_totals = {}
        for usage in mock_usages:
            model = usage.model_name
            if model not in model_totals:
                model_totals[model] = 0
            model_totals[model] += usage.total_tokens

        assert model_totals["gpt-4"] == 3000
        assert model_totals["claude-3"] == 1500

    def test_aggregate_truncated_requests_by_model(self):
        """Test aggregating truncated requests by model."""
        mock_usages = [
            Mock(model_name="gpt-4", context_truncated=True),
            Mock(model_name="gpt-4", context_truncated=False),
            Mock(model_name="claude-3", context_truncated=True),
            Mock(model_name="claude-3", context_truncated=True),
        ]

        model_truncated = {}
        for usage in mock_usages:
            model = usage.model_name
            if model not in model_truncated:
                model_truncated[model] = 0
            if usage.context_truncated:
                model_truncated[model] += 1

        assert model_truncated["gpt-4"] == 1
        assert model_truncated["claude-3"] == 2


class TestModelTokenStatsFields:
    """Tests for model_token_stats fields added in context-overflow rework."""

    def test_min_prompt_field_present(self):
        """min_prompt is computed from func.min(prompt_tokens)."""
        mock_row = Mock()
        mock_row.model_name = "llama3"
        mock_row.model_provider = "ollama"
        mock_row.total_requests = 5
        mock_row.total_tokens = 10000
        mock_row.min_prompt = 200
        mock_row.avg_prompt = 1500
        mock_row.max_prompt = 3000
        mock_row.avg_response_time_ms = 1200

        entry = {
            "model": mock_row.model_name,
            "provider": mock_row.model_provider,
            "total_requests": mock_row.total_requests,
            "total_tokens": int(mock_row.total_tokens or 0),
            "min_prompt": int(mock_row.min_prompt or 0),
            "avg_prompt": round(mock_row.avg_prompt or 0, 0),
            "max_prompt": int(mock_row.max_prompt or 0),
            "avg_response_time_ms": round(
                mock_row.avg_response_time_ms or 0, 0
            ),
        }

        assert entry["min_prompt"] == 200

    def test_min_prompt_defaults_to_zero(self):
        """min_prompt falls back to 0 when the row value is None."""
        mock_row = Mock()
        mock_row.model_name = "llama3"
        mock_row.model_provider = "ollama"
        mock_row.total_requests = 3
        mock_row.total_tokens = 5000
        mock_row.min_prompt = None
        mock_row.avg_prompt = 800
        mock_row.max_prompt = 2000
        mock_row.avg_response_time_ms = None

        entry = {
            "model": mock_row.model_name,
            "provider": mock_row.model_provider,
            "total_requests": mock_row.total_requests,
            "total_tokens": int(mock_row.total_tokens or 0),
            "min_prompt": int(mock_row.min_prompt or 0),
            "avg_prompt": round(mock_row.avg_prompt or 0, 0),
            "max_prompt": int(mock_row.max_prompt or 0),
            "avg_response_time_ms": round(
                mock_row.avg_response_time_ms or 0, 0
            ),
        }

        assert entry["min_prompt"] == 0
        assert entry["avg_response_time_ms"] == 0


class TestChartDataNewFields:
    """Tests for new chart_data fields: provider, research_phase, response_time_ms."""

    def test_chart_data_includes_provider(self):
        """chart_data items include the provider field."""
        mock_usage = Mock()
        mock_usage.timestamp = datetime(
            2025, 1, 1, 12, 0, 0, tzinfo=timezone.utc
        )
        mock_usage.research_id = "test-123"
        mock_usage.prompt_tokens = 1000
        mock_usage.completion_tokens = 200
        mock_usage.ollama_prompt_eval_count = None
        mock_usage.context_truncated = False
        mock_usage.tokens_truncated = 0
        mock_usage.context_limit = 4096
        mock_usage.model_name = "llama3"
        mock_usage.model_provider = "ollama"
        mock_usage.research_phase = "analysis"
        mock_usage.response_time_ms = 1500

        ollama_used = mock_usage.ollama_prompt_eval_count
        actual_prompt = ollama_used or mock_usage.prompt_tokens
        tokens_truncated = mock_usage.tokens_truncated or 0
        original_tokens = (
            actual_prompt + tokens_truncated
            if mock_usage.context_truncated
            else actual_prompt
        )

        chart_item = {
            "timestamp": mock_usage.timestamp.isoformat(),
            "research_id": mock_usage.research_id,
            "prompt_tokens": mock_usage.prompt_tokens,
            "completion_tokens": mock_usage.completion_tokens,
            "ollama_prompt_tokens": ollama_used,
            "original_prompt_tokens": original_tokens,
            "context_limit": mock_usage.context_limit,
            "truncated": bool(mock_usage.context_truncated),
            "tokens_truncated": tokens_truncated,
            "model": mock_usage.model_name,
            "provider": mock_usage.model_provider,
            "research_phase": mock_usage.research_phase,
            "response_time_ms": mock_usage.response_time_ms,
        }

        assert chart_item["provider"] == "ollama"
        assert chart_item["research_phase"] == "analysis"
        assert chart_item["response_time_ms"] == 1500

    def test_chart_data_fields_handle_none(self):
        """chart_data new fields default correctly when None."""
        mock_usage = Mock()
        mock_usage.timestamp = datetime(
            2025, 1, 1, 12, 0, 0, tzinfo=timezone.utc
        )
        mock_usage.research_id = "test-456"
        mock_usage.prompt_tokens = 500
        mock_usage.completion_tokens = 100
        mock_usage.ollama_prompt_eval_count = None
        mock_usage.context_truncated = False
        mock_usage.tokens_truncated = 0
        mock_usage.context_limit = None
        mock_usage.model_name = "gpt-4"
        mock_usage.model_provider = None
        mock_usage.research_phase = None
        mock_usage.response_time_ms = None

        chart_item = {
            "model": mock_usage.model_name,
            "provider": mock_usage.model_provider,
            "research_phase": mock_usage.research_phase,
            "response_time_ms": mock_usage.response_time_ms,
        }

        assert chart_item["provider"] is None
        assert chart_item["research_phase"] is None
        assert chart_item["response_time_ms"] is None


class TestCurrentContextWindow:
    """Tests for current_context_window field in API response."""

    def test_current_context_window_in_response(self):
        """API response includes current_context_window from SettingsManager."""
        # Simulate the response construction logic
        mock_context_window = 8192

        response = {
            "status": "success",
            "overview": {},
            "token_summary": {},
            "pagination": {},
            "current_context_window": mock_context_window,
        }

        assert response["current_context_window"] == 8192

    def test_current_context_window_none_when_unset(self):
        """current_context_window is None when setting is not configured."""
        response = {
            "status": "success",
            "overview": {},
            "token_summary": {},
            "pagination": {},
            "current_context_window": None,
        }

        assert response["current_context_window"] is None