crawlee 1.1.1b1__py3-none-any.whl → 1.1.1b2__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- crawlee/crawlers/_basic/_basic_crawler.py +2 -2
- {crawlee-1.1.1b1.dist-info → crawlee-1.1.1b2.dist-info}/METADATA +1 -1
- {crawlee-1.1.1b1.dist-info → crawlee-1.1.1b2.dist-info}/RECORD +6 -6
- {crawlee-1.1.1b1.dist-info → crawlee-1.1.1b2.dist-info}/WHEEL +0 -0
- {crawlee-1.1.1b1.dist-info → crawlee-1.1.1b2.dist-info}/entry_points.txt +0 -0
- {crawlee-1.1.1b1.dist-info → crawlee-1.1.1b2.dist-info}/licenses/LICENSE +0 -0
|
@@ -1043,8 +1043,8 @@ class BasicCrawler(Generic[TCrawlingContext, TStatisticsState]):
|
|
|
1043
1043
|
return target_url.hostname == origin_url.hostname
|
|
1044
1044
|
|
|
1045
1045
|
if strategy == 'same-domain':
|
|
1046
|
-
origin_domain = self._tld_extractor.extract_str(origin_url.hostname).
|
|
1047
|
-
target_domain = self._tld_extractor.extract_str(target_url.hostname).
|
|
1046
|
+
origin_domain = self._tld_extractor.extract_str(origin_url.hostname).top_domain_under_public_suffix
|
|
1047
|
+
target_domain = self._tld_extractor.extract_str(target_url.hostname).top_domain_under_public_suffix
|
|
1048
1048
|
return origin_domain == target_domain
|
|
1049
1049
|
|
|
1050
1050
|
if strategy == 'same-origin':
|
|
@@ -65,7 +65,7 @@ crawlee/crawlers/_adaptive_playwright/_rendering_type_predictor.py,sha256=TM4mkb
|
|
|
65
65
|
crawlee/crawlers/_adaptive_playwright/_result_comparator.py,sha256=NAfw5VKzTnkvARtLr_zrZj6UGeMp05Voc6Oi8oPxU3w,1747
|
|
66
66
|
crawlee/crawlers/_adaptive_playwright/_utils.py,sha256=EUYVz5i2YkLpL_gbVRp9BAD5u6w1xJ_AFzc_qB9bdDQ,1102
|
|
67
67
|
crawlee/crawlers/_basic/__init__.py,sha256=LPln8SiBBXSMqrApiFUfpqz3hvqxN5HUa1cHQXMVKgU,280
|
|
68
|
-
crawlee/crawlers/_basic/_basic_crawler.py,sha256
|
|
68
|
+
crawlee/crawlers/_basic/_basic_crawler.py,sha256=w8JSPF1zw2QG_1Phek5dENL_d5BC2N4dSZ6oR0MoUSY,73052
|
|
69
69
|
crawlee/crawlers/_basic/_basic_crawling_context.py,sha256=fjxm2RQXMDkDlWu38dQ3xn5rrGUOhJXkXiqkgbFJFk4,155
|
|
70
70
|
crawlee/crawlers/_basic/_context_pipeline.py,sha256=vM8EEvnCoguERjRV3oyrxUq2Ln2F9DzY7P5dAEiuMHo,5869
|
|
71
71
|
crawlee/crawlers/_basic/_logging_utils.py,sha256=jp5mEwSq5a_BgzUhNPJ9WrIDcoIeYGbeHstcRqCcP0s,3093
|
|
@@ -199,8 +199,8 @@ crawlee/storages/_request_queue.py,sha256=bjBOGbpMaGUsqJPVB-JD2VShziPAYMI-GvWKKp
|
|
|
199
199
|
crawlee/storages/_storage_instance_manager.py,sha256=72n0YlPwNpSQDJSPf4TxnI2GvIK6L-ZiTmHRbFcoVU0,8164
|
|
200
200
|
crawlee/storages/_utils.py,sha256=Yz-5tEBYKYCFJemYT29--uGJqoJLApLDLgPcsnbifRw,439
|
|
201
201
|
crawlee/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
202
|
-
crawlee-1.1.
|
|
203
|
-
crawlee-1.1.
|
|
204
|
-
crawlee-1.1.
|
|
205
|
-
crawlee-1.1.
|
|
206
|
-
crawlee-1.1.
|
|
202
|
+
crawlee-1.1.1b2.dist-info/METADATA,sha256=LMwCP8_OXNKMiOwWfP1tF-6gy2OTkjJdrjtfKHRfdLs,29532
|
|
203
|
+
crawlee-1.1.1b2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
|
|
204
|
+
crawlee-1.1.1b2.dist-info/entry_points.txt,sha256=1p65X3dA-cYvzjtlxLL6Kn1wpY-3uEDVqJLp53uNPeo,45
|
|
205
|
+
crawlee-1.1.1b2.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
|
|
206
|
+
crawlee-1.1.1b2.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|
|
File without changes
|