Skip to content

Scrapy, a fast high-level web crawling & scraping framework for Python.

20.7 KB
Raw Blame
from __future__ import annotations

import asyncio
from typing import TYPE_CHECKING, Any
from unittest import mock

import pytest
from twisted.internet.defer import CancelledError, Deferred, DeferredList
from twisted.python import failure

from scrapy import Spider, signals
from scrapy.downloadermiddlewares.robotstxt import RobotsTxtMiddleware
from scrapy.exceptions import CannotResolveHostError, IgnoreRequest, NotConfigured
from scrapy.http import Request, Response, TextResponse
from scrapy.http.request import NO_CALLBACK
from scrapy.settings import Settings
from scrapy.utils.asyncio import call_later, sleep
from scrapy.utils.defer import deferred_from_coro, maybe_deferred_to_future
from scrapy.utils.misc import build_from_crawler
from scrapy.utils.test import get_crawler
from tests.utils.decorators import coroutine_test
from tests.utils.robotstxt import rerp_available

if TYPE_CHECKING:
    from collections.abc import Iterable

    from tests.mockserver.http import MockServer


class TestRobotsTxtMiddleware:
    def setup_method(self) -> None:
        self.crawler: mock.MagicMock = mock.MagicMock()
        self.crawler.settings = Settings()
        self.crawler.engine.download_async = mock.AsyncMock()
        self.crawler.signals.send_catch_log_async = mock.AsyncMock(return_value=[])

    def teardown_method(self):
        del self.crawler

    def test_robotstxt_settings(self):
        self.crawler.settings = Settings()
        self.crawler.settings.set("USER_AGENT", "CustomAgent")
        with pytest.raises(NotConfigured):
            build_from_crawler(RobotsTxtMiddleware, self.crawler)

    def _get_successful_crawler(self) -> mock.MagicMock:
        crawler = self.crawler
        crawler.settings.set("ROBOTSTXT_OBEY", True)
        ROBOTS = """
User-Agent: *
Disallow: /admin/
Disallow: /static/
# taken from https://en.wikipedia.org/robots.txt
Disallow: /wiki/K%C3%A4ytt%C3%A4j%C3%A4:
Disallow: /wiki/Käyttäjä:
User-Agent: UnicödeBöt
Disallow: /some/randome/page.html
""".encode()
        response = TextResponse("http://site.local/robots.txt", body=ROBOTS)

        async def return_response(request: Request) -> Response:
            deferred: Deferred[Response] = Deferred()
            call_later(0, deferred.callback, response)
            return await maybe_deferred_to_future(deferred)

        crawler.engine.download_async.side_effect = return_response
        return crawler

    @coroutine_test
    async def test_robotstxt(self):
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        self.assertRobotsTxtRequested("http://site.local")
        await self.assertIgnored(Request("http://site.local/admin/main"), middleware)
        await self.assertIgnored(Request("http://site.local/static/"), middleware)
        await self.assertIgnored(
            Request("http://site.local/wiki/K%C3%A4ytt%C3%A4j%C3%A4:"), middleware
        )
        await self.assertIgnored(
            Request("http://site.local/wiki/Käyttäjä:"), middleware
        )

    @coroutine_test
    async def test_robotstxt_emits_robots_parsed_signal(self):
        crawler = self._get_successful_crawler()
        middleware = RobotsTxtMiddleware(crawler)
        request = Request("http://site.local/allowed")
        await self.assertNotIgnored(request, middleware)
        calls = [
            kwargs
            for _, kwargs in crawler.signals.send_catch_log_async.call_args_list
            if kwargs.get("signal") is signals.robots_parsed
        ]
        assert len(calls) == 1
        assert calls[0]["request"] is request
        assert calls[0]["robotparser"] is not None
        assert calls[0]["response"].url == "http://site.local/robots.txt"

    @coroutine_test
    async def test_robotstxt_multiple_reqs(self) -> None:
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        d1 = deferred_from_coro(
            self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        )
        d2 = deferred_from_coro(
            self.assertIgnored(Request("http://site.local/admin/main"), middleware)
        )
        d3 = deferred_from_coro(
            self.assertIgnored(Request("http://site.local/static/"), middleware)
        )
        await maybe_deferred_to_future(
            DeferredList([d1, d2, d3], fireOnOneErrback=True)
        )

    @pytest.mark.only_asyncio
    @coroutine_test
    async def test_robotstxt_multiple_reqs_asyncio(self) -> None:
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        await asyncio.gather(
            self.assertNotIgnored(Request("http://site.local/allowed"), middleware),
            self.assertIgnored(Request("http://site.local/admin/main"), middleware),
            self.assertIgnored(Request("http://site.local/static/"), middleware),
        )

    @coroutine_test
    async def test_robotstxt_multiple_reqs_cancelled(self) -> None:
        """Cancelling one request doesn't affect others."""
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        release: Deferred[None] = Deferred()

        async def return_response(request: Request) -> Response:
            await maybe_deferred_to_future(release)
            return TextResponse(request.url, body=b"User-Agent: *\nDisallow: /admin/")

        self.crawler.engine.download_async.side_effect = return_response
        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        d1 = deferred_from_coro(
            self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        )
        d2 = deferred_from_coro(
            middleware.process_request(Request("http://site.local/admin/1"))
        )
        d3 = deferred_from_coro(
            self.assertIgnored(Request("http://site.local/admin/2"), middleware)
        )
        await sleep(0)
        d2.cancel()
        with pytest.raises((CancelledError, asyncio.CancelledError)):
            await maybe_deferred_to_future(d2)
        release.callback(None)
        await maybe_deferred_to_future(DeferredList([d1, d3], fireOnOneErrback=True))

    @coroutine_test
    async def test_robotstxt_ready_parser(self):
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)

    @coroutine_test
    async def test_robotstxt_meta(self):
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        meta = {"dont_obey_robotstxt": True}
        await self.assertNotIgnored(
            Request("http://site.local/allowed", meta=meta), middleware
        )
        await self.assertNotIgnored(
            Request("http://site.local/admin/main", meta=meta), middleware
        )
        await self.assertNotIgnored(
            Request("http://site.local/static/", meta=meta), middleware
        )

    def _get_garbage_crawler(self) -> mock.MagicMock:
        crawler = self.crawler
        crawler.settings.set("ROBOTSTXT_OBEY", True)
        response = Response(
            "http://site.local/robots.txt", body=b"GIF89a\xd3\x00\xfe\x00\xa2"
        )

        async def return_response(request: Request) -> Response:
            deferred: Deferred[Response] = Deferred()
            call_later(0, deferred.callback, response)
            return await maybe_deferred_to_future(deferred)

        crawler.engine.download_async.side_effect = return_response
        return crawler

    @coroutine_test
    async def test_robotstxt_garbage(self):
        # garbage response should be discarded, equal 'allow all'
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_garbage_crawler()
        )
        await self.assertNotIgnored(Request("http://site.local"), middleware)
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        await self.assertNotIgnored(Request("http://site.local/admin/main"), middleware)
        await self.assertNotIgnored(Request("http://site.local/static/"), middleware)

    def _get_emptybody_crawler(self) -> mock.MagicMock:
        crawler = self.crawler
        crawler.settings.set("ROBOTSTXT_OBEY", True)
        response = Response("http://site.local/robots.txt")

        async def return_response(request: Request) -> Response:
            deferred: Deferred[Response] = Deferred()
            call_later(0, deferred.callback, response)
            return await maybe_deferred_to_future(deferred)

        crawler.engine.download_async.side_effect = return_response
        return crawler

    @coroutine_test
    async def test_robotstxt_empty_response(self):
        # empty response should equal 'allow all'
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_emptybody_crawler()
        )
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        await self.assertNotIgnored(Request("http://site.local/admin/main"), middleware)
        await self.assertNotIgnored(Request("http://site.local/static/"), middleware)

    @coroutine_test
    async def test_robotstxt_error(self, caplog: pytest.LogCaptureFixture) -> None:
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        err = CannotResolveHostError("Robotstxt address not found")

        async def return_failure(request: Request) -> Response:
            deferred: Deferred[Response] = Deferred()
            call_later(0, deferred.errback, failure.Failure(err))
            return await maybe_deferred_to_future(deferred)

        self.crawler.engine.download_async.side_effect = return_failure

        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        await middleware.process_request(Request("http://site.local"))
        assert "Robotstxt address not found" in caplog.text

    @coroutine_test
    async def test_robotstxt_immediate_error(self):
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        err = CannotResolveHostError("Robotstxt address not found")

        async def immediate_failure(request):
            raise err

        self.crawler.engine.download_async.side_effect = immediate_failure

        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        await self.assertNotIgnored(Request("http://site.local"), middleware)

    @coroutine_test
    async def test_robotstxt_request_download_error(
        self, caplog: pytest.LogCaptureFixture
    ) -> None:
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        request = Request("http://site.local/robots.txt")
        await middleware.process_request(request)
        err = CannotResolveHostError("Robotstxt address not found")
        await middleware.process_exception(request, err)
        assert "Robotstxt address not found" in caplog.text
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)

    @pytest.mark.parametrize(
        "follow_up_url",
        ["http://site.local/robots.txt", "https://site.local/robots.txt"],
    )
    @pytest.mark.parametrize("fails", [False, True])
    @coroutine_test
    async def test_robotstxt_request_retried_or_redirected(
        self, follow_up_url: str, fails: bool
    ) -> None:
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        request = Request("http://site.local/robots.txt")
        await middleware.process_request(request)
        follow_up = request.replace(url=follow_up_url)
        await self.assertNotIgnored(follow_up, middleware)
        blocked = Request("http://site.local/blocked")
        if fails:
            err = CannotResolveHostError("Robotstxt address not found")
            await middleware.process_exception(follow_up, err)
            await self.assertNotIgnored(blocked, middleware)
        else:
            body = b"User-agent: *\nDisallow: /blocked\n"
            response = TextResponse(follow_up_url, body=body)
            await middleware.process_response(follow_up, response)
            await self.assertIgnored(blocked, middleware)
        assert not self.crawler.engine.download_async.called

    @coroutine_test
    async def test_robotstxt_process_exception_ignores_unrelated_request(
        self,
    ) -> None:
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        await self.assertNotIgnored(Request("http://site.local/allowed"), middleware)
        err = CannotResolveHostError("Unrelated request failed")
        with mock.patch(
            "scrapy.downloadermiddlewares.robotstxt.logger"
        ) as mw_module_logger:
            await middleware.process_exception(
                Request("http://site.local/allowed"), err
            )
            assert not mw_module_logger.error.called

    @coroutine_test
    async def test_ignore_robotstxt_request(self):
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)

        async def ignore_request(request: Request) -> Response:
            deferred: Deferred[Response] = Deferred()
            call_later(0, deferred.errback, failure.Failure(IgnoreRequest()))
            return await maybe_deferred_to_future(deferred)

        self.crawler.engine.download_async.side_effect = ignore_request

        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        with mock.patch(
            "scrapy.downloadermiddlewares.robotstxt.logger"
        ) as mw_module_logger:
            await self.assertNotIgnored(
                Request("http://site.local/allowed"), middleware
            )
            assert not mw_module_logger.error.called

    def test_robotstxt_user_agent_setting(self):
        crawler = self._get_successful_crawler()
        crawler.settings.set("ROBOTSTXT_USER_AGENT", "Examplebot")
        crawler.settings.set("USER_AGENT", "Mozilla/5.0 (X11; Linux x86_64)")
        middleware = build_from_crawler(RobotsTxtMiddleware, crawler)
        rp = mock.MagicMock(return_value=True)
        middleware.process_request_2(rp, Request("http://site.local/allowed"))
        rp.allowed.assert_called_once_with("http://site.local/allowed", "Examplebot")

    @coroutine_test
    async def test_robotstxt_user_agent_none(self):
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_successful_crawler()
        )
        headers = {"User-Agent": None}
        await self.assertNotIgnored(
            Request("http://site.local/allowed", headers=headers), middleware
        )
        await self.assertIgnored(
            Request("http://site.local/admin/main", headers=headers), middleware
        )

    @coroutine_test
    async def test_robotstxt_local_file(self):
        middleware = build_from_crawler(
            RobotsTxtMiddleware, self._get_emptybody_crawler()
        )
        middleware.process_request_2 = mock.MagicMock()  # type: ignore[method-assign]

        await middleware.process_request(Request("data:text/plain,Hello World data"))
        assert not middleware.process_request_2.called

        await middleware.process_request(
            Request("file:///tests/sample_data/test_site/nothinghere.html")
        )
        assert not middleware.process_request_2.called

        await middleware.process_request(Request("http://site.local/allowed"))
        assert middleware.process_request_2.called

    @coroutine_test
    async def test_robotstxt_download_handler(self) -> None:
        middleware = RobotsTxtMiddleware(self._get_successful_crawler())
        await self.assertNotIgnored(
            Request("http://site.local/allowed", meta={"download_handler": "named"}),
            middleware,
        )
        robotsreq = self.crawler.engine.download_async.call_args_list[0][0][0]
        assert robotsreq.meta["download_handler"] == "named"
        assert robotsreq.meta["is_robotstxt_request"] is True

    @pytest.mark.parametrize("ignored_first", [False, True])
    @coroutine_test
    async def test_robotstxt_per_scheme(self, ignored_first: bool) -> None:
        allowed_url = "http://site.local/blocked"
        ignored_url = "https://site.local/blocked"
        self.crawler.settings.set("ROBOTSTXT_OBEY", True)
        robots = {
            "http://site.local/robots.txt": b"User-agent: *\nDisallow:\n",
            "https://site.local/robots.txt": b"User-agent: *\nDisallow: /blocked\n",
        }

        async def return_response(request: Request) -> Response:
            return TextResponse(request.url, body=robots[request.url])

        self.crawler.engine.download_async.side_effect = return_response
        middleware = build_from_crawler(RobotsTxtMiddleware, self.crawler)
        if ignored_first:
            await self.assertIgnored(Request(ignored_url), middleware)
            await self.assertNotIgnored(Request(allowed_url), middleware)
        else:
            await self.assertNotIgnored(Request(allowed_url), middleware)
            await self.assertIgnored(Request(ignored_url), middleware)
        calls = self.crawler.engine.download_async.call_args_list
        assert sorted(call[0][0].url for call in calls) == sorted(robots)

    async def assertNotIgnored(
        self, request: Request, middleware: RobotsTxtMiddleware
    ) -> None:
        try:
            await middleware.process_request(request)
        except IgnoreRequest:
            pytest.fail("IgnoreRequest was raised unexpectedly")

    async def assertIgnored(
        self, request: Request, middleware: RobotsTxtMiddleware
    ) -> None:
        with pytest.raises(IgnoreRequest):
            await middleware.process_request(request)

    def assertRobotsTxtRequested(self, base_url: str) -> None:
        calls = self.crawler.engine.download_async.call_args_list
        request = calls[0][0][0]
        assert request.url == f"{base_url}/robots.txt"
        assert request.callback == NO_CALLBACK


class _IgnoreSpider(Spider):
    name = "test"

    def parse(self, response: Response) -> None:
        pass


class _FollowSpider(Spider):
    name = "test"

    def parse(self, response: Response) -> Iterable[Request]:
        yield response.follow("/deny/a")


class TestRobotsTxtDeniedCloseReason:
    @staticmethod
    async def _finish_reason(
        spider_cls: type[Spider], settings: dict[str, Any] | None = None, **kwargs: Any
    ) -> Any:
        crawler = get_crawler(
            spider_cls, settings_dict={"ROBOTSTXT_OBEY": True, **(settings or {})}
        )
        await crawler.crawl_async(**kwargs)
        assert crawler.stats
        return crawler.stats.get_value("finish_reason")

    @coroutine_test
    async def test_all_denied(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _IgnoreSpider,
            start_urls=[mockserver.url("/deny/a"), mockserver.url("/deny/b")],
        )
        assert reason == "robotstxt_denied"

    @coroutine_test
    async def test_all_denied_low_concurrency(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _IgnoreSpider,
            settings={"CONCURRENT_REQUESTS": 1},
            start_urls=[mockserver.url("/deny/a"), mockserver.url("/deny/b")],
        )
        assert reason == "robotstxt_denied"

    @coroutine_test
    async def test_all_denied_after_redirect(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _IgnoreSpider, start_urls=[mockserver.url("/redirect-to?goto=/deny/a")]
        )
        assert reason == "robotstxt_denied"

    @coroutine_test
    async def test_denied_and_download_failure(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _IgnoreSpider,
            settings={"RETRY_ENABLED": False},
            start_urls=[mockserver.url("/deny/a"), mockserver.url("/drop")],
        )
        assert reason == "robotstxt_denied"

    @coroutine_test
    async def test_some_denied(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _IgnoreSpider,
            start_urls=[mockserver.url("/deny/a"), mockserver.url("/text")],
        )
        assert reason == "finished"

    @coroutine_test
    async def test_denied_follow_up_request(self, mockserver: MockServer) -> None:
        reason = await self._finish_reason(
            _FollowSpider, start_urls=[mockserver.url("/text")]
        )
        assert reason == "finished"


@pytest.mark.skipif(not rerp_available(), reason="Rerp parser is not installed")
class TestRobotsTxtMiddlewareWithRerp(TestRobotsTxtMiddleware):
    def setup_method(self):
        super().setup_method()
        self.crawler.settings.set(
            "ROBOTSTXT_PARSER", "scrapy.robotstxt.RerpRobotParser"
        )