crawlee 1.0.5b21__py3-none-any.whl → 1.1.0__py3-none-any.whl

crawlee/request_loaders/_sitemap_request_loader.py

@@ -9,7 +9,7 @@ from typing import TYPE_CHECKING, Annotated, Any
 from pydantic import BaseModel, ConfigDict, Field
 from typing_extensions import override
 
-from crawlee import Request
+from crawlee import Request, RequestOptions
 from crawlee._utils.docs import docs_group
 from crawlee._utils.globs import Glob
 from crawlee._utils.recoverable_state import RecoverableState
@@ -18,9 +18,10 @@ from crawlee.request_loaders._request_loader import RequestLoader
 
 if TYPE_CHECKING:
     import re
-    from collections.abc import Sequence
+    from collections.abc import Callable, Sequence
     from types import TracebackType
 
+    from crawlee import RequestTransformAction
     from crawlee.http_clients import HttpClient
     from crawlee.proxy_configuration import ProxyInfo
     from crawlee.storage_clients.models import ProcessedRequest
@@ -112,6 +113,7 @@ class SitemapRequestLoader(RequestLoader):
         exclude: list[re.Pattern[Any] | Glob] | None = None,
         max_buffer_size: int = 200,
         persist_state_key: str | None = None,
+        transform_request_function: Callable[[RequestOptions], RequestOptions | RequestTransformAction] | None = None,
     ) -> None:
         """Initialize the sitemap request loader.
 
@@ -125,6 +127,9 @@ class SitemapRequestLoader(RequestLoader):
             persist_state_key: A key for persisting the loader's state in the KeyValueStore.
                 When provided, allows resuming from where it left off after interruption.
                 If None, no state persistence occurs.
+            transform_request_function: An optional function to transform requests
+                generated by the loader. It receives `RequestOptions` with `url` and should return either
+                modified `RequestOptions` or a `RequestTransformAction`.
         """
         self._http_client = http_client
         self._sitemap_urls = sitemap_urls
@@ -132,6 +137,7 @@ class SitemapRequestLoader(RequestLoader):
         self._exclude = exclude
         self._proxy_info = proxy_info
         self._max_buffer_size = max_buffer_size
+        self._transform_request_function = transform_request_function
 
         # Synchronization for queue operations
         self._queue_has_capacity = asyncio.Event()
@@ -313,8 +319,15 @@ class SitemapRequestLoader(RequestLoader):
 
             async with self._queue_lock:
                 url = state.url_queue.popleft()
-
-                request = Request.from_url(url)
+                request_option = RequestOptions(url=url)
+                if self._transform_request_function:
+                    transform_request_option = self._transform_request_function(request_option)
+                    if transform_request_option == 'skip':
+                        state.total_count -= 1
+                        continue
+                    if transform_request_option != 'unchanged':
+                        request_option = transform_request_option
+                request = Request.from_url(**request_option)
                 state.in_progress.add(request.url)
                 if len(state.url_queue) < self._max_buffer_size:
                     self._queue_has_capacity.set()

{crawlee-1.0.5b21.dist-info → crawlee-1.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: crawlee
-Version: 1.0.5b21
+Version: 1.1.0
 Summary: Crawlee for Python
 Project-URL: Apify Homepage, https://apify.com
 Project-URL: Changelog, https://crawlee.dev/python/docs/changelog

{crawlee-1.0.5b21.dist-info → crawlee-1.1.0.dist-info}/RECORD

@@ -137,7 +137,7 @@ crawlee/request_loaders/_request_list.py,sha256=SIalHBMuFanE5GLnFocI0QCppWUiJQjr
 crawlee/request_loaders/_request_loader.py,sha256=2Bg-AWWkIV1W-Dwjqo91dPY8nmc7H3teQy7d6OSgliQ,3620
 crawlee/request_loaders/_request_manager.py,sha256=qFizyJuV2meIb9iiPfuii7ciuERMrp4SldAufiH46dc,3000
 crawlee/request_loaders/_request_manager_tandem.py,sha256=lv-s94KPsoQAqx1KaXFch96ejhO147uOflF3UK5ORTk,4058
-crawlee/request_loaders/_sitemap_request_loader.py,sha256=s65D_N0mZxeIrGJEjqUYfu1uYj2AXSOkmErSnfAHv2A,15554
+crawlee/request_loaders/_sitemap_request_loader.py,sha256=W1_k_Szrtk0iE2LJBkHrrFeDtcKReXzr3DG32EnQaQE,16565
 crawlee/sessions/__init__.py,sha256=dJdelbL-6MK5sW4SMU4QrjFbb9kRZ9uRnN-VS3R5-8Y,190
 crawlee/sessions/_cookies.py,sha256=ihYbmpXfCzClzXDT7M2wefB_3KVzcMUdIzTZo6uGk6Y,9356
 crawlee/sessions/_models.py,sha256=JMRQgDUP30XUdZ32isncHowOsXvK9jC_m9QYegbBI1E,2916
@@ -199,8 +199,8 @@ crawlee/storages/_request_queue.py,sha256=bjBOGbpMaGUsqJPVB-JD2VShziPAYMI-GvWKKp
 crawlee/storages/_storage_instance_manager.py,sha256=72n0YlPwNpSQDJSPf4TxnI2GvIK6L-ZiTmHRbFcoVU0,8164
 crawlee/storages/_utils.py,sha256=Yz-5tEBYKYCFJemYT29--uGJqoJLApLDLgPcsnbifRw,439
 crawlee/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-crawlee-1.0.5b21.dist-info/METADATA,sha256=ERsOFxwDxPP0IVAwE8ZKLakRlDNcXKTYSi9ZKzFHCSQ,29533
-crawlee-1.0.5b21.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-crawlee-1.0.5b21.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
-crawlee-1.0.5b21.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
-crawlee-1.0.5b21.dist-info/RECORD,,
+crawlee-1.1.0.dist-info/METADATA,sha256=lSL2WVejm9D2tHBLJWJuobID9_WyglyirKeLsMzp1Z0,29530
+crawlee-1.1.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+crawlee-1.1.0.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
+crawlee-1.1.0.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
+crawlee-1.1.0.dist-info/RECORD,,