Skip to content

DeepTutor: Lifelong Personalized Tutoring. https://deeptutor.info/.

8.2 KB
Raw Blame
import asyncio
import json
import logging

import pytest

from deeptutor.api.utils.task_log_stream import (
    KnowledgeTaskStreamManager,
    capture_task_logs,
    get_task_stream_manager,
)
from deeptutor.logging import PROCESS_LOG_PRIVATE_ATTR


@pytest.mark.asyncio
async def test_knowledge_task_stream_emits_process_log_sse_event():
    manager = KnowledgeTaskStreamManager()
    manager.ensure_task("task-1")
    manager.emit_log("task-1", "Indexing started")

    stream = manager.stream("task-1")
    try:
        chunk = await anext(stream)
    finally:
        await stream.aclose()

    lines = chunk.splitlines()
    header, data_line = lines[:2]
    assert header == "event: process_log"
    payload = json.loads(data_line.removeprefix("data: "))
    assert payload["type"] == "process_log"
    assert payload["message"] == "Indexing started"
    assert payload["context"]["task_id"] == "task-1"


@pytest.mark.asyncio
async def test_knowledge_task_stream_keeps_idle_sse_connection_alive():
    manager = KnowledgeTaskStreamManager()
    manager._HEARTBEAT_SECONDS = 0.01
    manager.ensure_task("task-idle")

    stream = manager.stream("task-idle")
    try:
        heartbeat = await asyncio.wait_for(anext(stream), timeout=0.2)
        assert heartbeat == ": keep-alive\n\n"

        manager.emit_log("task-idle", "Indexing resumed")
        event = await asyncio.wait_for(anext(stream), timeout=0.2)
        assert "event: process_log" in event
        assert "Indexing resumed" in event
    finally:
        await stream.aclose()


def test_knowledge_task_stream_emits_structured_failure_metadata():
    manager = KnowledgeTaskStreamManager()
    manager.ensure_task("task-failed")

    manager.emit_failed(
        "task-failed",
        "Choose a compatible chat model.",
        details="internal traceback",
        error_code="graphrag_model_incompatible",
        retryable=False,
    )

    event = list(manager._buffers["task-failed"])[-1]
    assert event["event"] == "failed"
    assert event["payload"]["detail"] == "Choose a compatible chat model."
    assert event["payload"]["details"] == "internal traceback"
    assert event["payload"]["error_code"] == "graphrag_model_incompatible"
    assert event["payload"]["retryable"] is False


def test_capture_task_logs_forwards_lightrag_non_propagating_logger():
    original_instance = KnowledgeTaskStreamManager._instance
    lightrag_logger = logging.getLogger("lightrag")
    original_handlers = list(lightrag_logger.handlers)
    original_propagate = lightrag_logger.propagate
    original_level = lightrag_logger.level
    try:
        KnowledgeTaskStreamManager._instance = KnowledgeTaskStreamManager()
        lightrag_logger.handlers = []
        lightrag_logger.propagate = False
        lightrag_logger.setLevel(logging.INFO)

        with capture_task_logs("task-native"):
            lightrag_logger.info("Chunk 1 of 1 extracted 14 Ent + 13 Rel")

        manager = get_task_stream_manager()
        events = list(manager._buffers["task-native"])
    finally:
        KnowledgeTaskStreamManager._instance = original_instance
        lightrag_logger.handlers = original_handlers
        lightrag_logger.propagate = original_propagate
        lightrag_logger.setLevel(original_level)

    assert any(
        event["event"] == "process_log"
        and event["payload"]["logger"] == "lightrag"
        and event["payload"]["message"] == "Chunk 1 of 1 extracted 14 Ent + 13 Rel"
        and event["payload"]["context"]["task_id"] == "task-native"
        for event in events
    )


def test_capture_task_logs_forwards_graphrag_propagating_logger_once():
    original_instance = KnowledgeTaskStreamManager._instance
    graphrag_logger = logging.getLogger("graphrag.api.query")
    original_handlers = list(graphrag_logger.handlers)
    original_propagate = graphrag_logger.propagate
    original_level = graphrag_logger.level
    try:
        KnowledgeTaskStreamManager._instance = KnowledgeTaskStreamManager()
        graphrag_logger.handlers = []
        graphrag_logger.propagate = True
        graphrag_logger.setLevel(logging.INFO)

        with capture_task_logs("task-graphrag"):
            graphrag_logger.info("GraphRAG local search selected 3 text units")

        manager = get_task_stream_manager()
        events = list(manager._buffers["task-graphrag"])
    finally:
        KnowledgeTaskStreamManager._instance = original_instance
        graphrag_logger.handlers = original_handlers
        graphrag_logger.propagate = original_propagate
        graphrag_logger.setLevel(original_level)

    matches = [
        event
        for event in events
        if event["event"] == "process_log"
        and event["payload"]["logger"] == "graphrag.api.query"
        and event["payload"]["message"] == "GraphRAG local search selected 3 text units"
        and event["payload"]["context"]["task_id"] == "task-graphrag"
    ]
    assert len(matches) == 1


def test_capture_task_logs_excludes_private_non_propagating_library_diagnostics():
    original_instance = KnowledgeTaskStreamManager._instance
    graphrag_logger = logging.getLogger("graphrag")
    original_handlers = list(graphrag_logger.handlers)
    original_propagate = graphrag_logger.propagate
    original_level = graphrag_logger.level
    try:
        KnowledgeTaskStreamManager._instance = KnowledgeTaskStreamManager()
        graphrag_logger.handlers = []
        graphrag_logger.propagate = False
        graphrag_logger.setLevel(logging.INFO)

        with capture_task_logs("task-private"):
            graphrag_logger.error(
                "Stack trace contains sk-secret-must-not-leak",
                extra={PROCESS_LOG_PRIVATE_ATTR: True},
            )

        manager = get_task_stream_manager()
        events = list(manager._buffers["task-private"])
    finally:
        KnowledgeTaskStreamManager._instance = original_instance
        graphrag_logger.handlers = original_handlers
        graphrag_logger.propagate = original_propagate
        graphrag_logger.setLevel(original_level)

    assert events == []


def test_capture_task_logs_keeps_user_stages_and_drops_runtime_noise():
    original_instance = KnowledgeTaskStreamManager._instance
    loggers = {
        name: logging.getLogger(name)
        for name in (
            "root",
            "asyncio",
            "deeptutor.knowledge.progress_tracker",
            "deeptutor.services.rag.pipelines.pageindex.pipeline",
        )
    }
    original_levels = {name: logger.level for name, logger in loggers.items()}
    try:
        KnowledgeTaskStreamManager._instance = KnowledgeTaskStreamManager()
        for logger in loggers.values():
            logger.setLevel(logging.INFO)
        with capture_task_logs("task-curated"):
            loggers["root"].error("Request timed out")
            loggers["asyncio"].error("Event loop is closed")
            loggers["deeptutor.knowledge.progress_tracker"].info("duplicate progress")
            loggers["deeptutor.services.rag.pipelines.pageindex.pipeline"].info(
                "PageIndex: submitting manual.pdf"
            )
        events = list(get_task_stream_manager()._buffers["task-curated"])
    finally:
        KnowledgeTaskStreamManager._instance = original_instance
        for name, logger in loggers.items():
            logger.setLevel(original_levels[name])

    messages = [event["payload"]["message"] for event in events]
    assert messages == ["PageIndex: submitting manual.pdf"]


def test_completed_task_buffers_are_bounded_and_restore_terminal_event():
    manager = KnowledgeTaskStreamManager()
    manager._MAX_RETAINED_TASKS = 3

    for index in range(8):
        task_id = f"task-{index}"
        manager.ensure_task(task_id)
        manager.emit_log(task_id, "x" * 100)
        manager.emit_complete(task_id)

    assert manager.retained_task_count() == 3
    assert len(manager._terminal_tombstones) == 5

    manager.ensure_task("task-0")
    restored = list(manager._buffers["task-0"])
    assert restored[-1]["event"] == "complete"


def test_task_buffer_has_approximate_byte_ceiling():
    manager = KnowledgeTaskStreamManager()
    manager._MAX_BYTES_PER_TASK = 2_000
    manager.ensure_task("large-task")

    for _ in range(20):
        manager.emit_log("large-task", "x" * 500)

    assert manager._buffer_bytes["large-task"] <= manager._MAX_BYTES_PER_TASK
    assert len(manager._buffers["large-task"]) < 20