fix: respect <base> when enqueuing (#1590)

### Description

- This PR focuses on ensuring that `extract_links` and `enqueue_links`
respect the `<base>` tag on the page.

### Issues

- Closes: #1589

### Testing

- Update tests for enqueuing

Author: Mantisus

src/crawlee/crawlers/_abstract_http/_abstract_http_crawler.py

@@ -167,9 +167,15 @@ async def extract_links(
             kwargs.setdefault('strategy', 'same-hostname')
 
             links_iterator: Iterator[str] = iter(self._parser.find_links(parsed_content, selector=selector))
-            links_iterator = to_absolute_url_iterator(
-                context.request.loaded_url or context.request.url, links_iterator, logger=context.log
+
+            # Get base URL from <base> tag if present
+            extracted_base_urls = list(self._parser.find_links(parsed_content, 'base[href]'))
+            base_url: str = (
+                str(extracted_base_urls[0])
+                if extracted_base_urls
+                else context.request.loaded_url or context.request.url
             )
+            links_iterator = to_absolute_url_iterator(base_url, links_iterator, logger=context.log)
 
             if robots_txt_file:
                 skipped, links_iterator = partition(lambda url: robots_txt_file.is_allowed(url), links_iterator)

src/crawlee/crawlers/_playwright/_playwright_crawler.py

@@ -369,9 +369,12 @@ async def extract_links(
             links_iterator: Iterator[str] = iter(
                 [url for element in elements if (url := await element.get_attribute('href')) is not None]
             )
-            links_iterator = to_absolute_url_iterator(
-                context.request.loaded_url or context.request.url, links_iterator, logger=context.log
-            )
+
+            # Get base URL from <base> tag if present
+            extracted_base_url = await context.page.evaluate('document.baseURI')
+            base_url: str = extracted_base_url or context.request.loaded_url or context.request.url
+
+            links_iterator = to_absolute_url_iterator(base_url, links_iterator, logger=context.log)
 
             if robots_txt_file:
                 skipped, links_iterator = partition(lambda url: robots_txt_file.is_allowed(url), links_iterator)

tests/unit/crawlers/_beautifulsoup/test_beautifulsoup_crawler.py

@@ -58,6 +58,9 @@ async def request_handler(context: BeautifulSoupCrawlingContext) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -131,6 +134,9 @@ async def request_handler(context: BeautifulSoupCrawlingContext) -> None:
         str(server_url / 'sub_index'),
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
+        str(server_url / 'base_page'),
+        str(server_url / 'page_4'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
     # # all urls added to `enqueue_links` must have a custom header
@@ -164,6 +170,8 @@ async def request_handler(context: BeautifulSoupCrawlingContext) -> None:
     assert visited == {
         str(server_url / 'start_enqueue'),
         str(server_url / 'sub_index'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -221,6 +229,7 @@ async def skipped_hook(url: str, _reason: SkippedReason) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
     }
 
 

tests/unit/crawlers/_parsel/test_parsel_crawler.py

@@ -61,6 +61,9 @@ async def request_handler(context: ParselCrawlingContext) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -151,6 +154,9 @@ async def request_handler(context: ParselCrawlingContext) -> None:
         str(server_url / 'sub_index'),
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
+        str(server_url / 'page_4'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
     # # all urls added to `enqueue_links` must have a custom header
@@ -258,6 +264,8 @@ async def request_handler(context: ParselCrawlingContext) -> None:
     assert visited == {
         str(server_url / 'start_enqueue'),
         str(server_url / 'sub_index'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -315,6 +323,7 @@ async def skipped_hook(url: str, _reason: SkippedReason) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
     }
 
 

tests/unit/crawlers/_playwright/test_playwright_crawler.py

@@ -99,6 +99,9 @@ async def request_handler(context: PlaywrightCrawlingContext) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -668,6 +671,8 @@ async def request_handler(context: PlaywrightCrawlingContext) -> None:
     assert visited == {
         str(server_url / 'start_enqueue'),
         str(server_url / 'sub_index'),
+        str(server_url / 'base_page'),
+        str(server_url / 'base_subpath/page_5'),
     }
 
 
@@ -724,6 +729,7 @@ async def skipped_hook(url: str, _reason: SkippedReason) -> None:
         str(server_url / 'page_1'),
         str(server_url / 'page_2'),
         str(server_url / 'page_3'),
+        str(server_url / 'page_4'),
     }
 
 

tests/unit/server.py

@@ -15,6 +15,7 @@
 from yarl import URL
 
 from tests.unit.server_endpoints import (
+    BASE_INDEX,
     GENERIC_RESPONSE,
     HELLO_WORLD,
     INCAPSULA,
@@ -105,6 +106,7 @@ async def app(scope: dict[str, Any], receive: Receive, send: Send) -> None:
         'page_1': generic_response_endpoint,
         'page_2': generic_response_endpoint,
         'page_3': generic_response_endpoint,
+        'base_page': base_index_endpoint,
         'problematic_links': problematic_links_endpoint,
         'set_cookies': set_cookies,
         'set_complex_cookies': set_complex_cookies,
@@ -431,6 +433,16 @@ async def resource_loading_endpoint(_scope: dict[str, Any], _receive: Receive, s
     )
 
 
+async def base_index_endpoint(_scope: dict[str, Any], _receive: Receive, send: Send) -> None:
+    """Handle requests for the base index page."""
+    host = f'http://{get_headers_dict(_scope).get("host", "localhost")}'
+    content = BASE_INDEX.format(host=host).encode()
+    await send_html_response(
+        send,
+        content,
+    )
+
+
 class TestServer(Server):
     """A test HTTP server implementation based on Uvicorn Server."""
 

tests/unit/server_endpoints.py

@@ -24,6 +24,18 @@
 <body>
     <a href="/page_3">Link 3</a>
     <a href="/page_2">Link 4</a>
+    <a href="/base_page">Base Page</a>
+</body></html>"""
+
+BASE_INDEX = """\
+<html><head>
+    <base href="{host}/base_subpath/">
+    <base href="{host}/sub_index/">
+    <title>Hello</title>
+</head>
+<body>
+    <a href="page_5">Link 5</a>
+    <a href="/page_4">Link 6</a>
 </body></html>"""
 
 INCAPSULA = b"""\