cobweb-launcher 1.1.17__py3-none-any.whl → 1.1.19__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.
Potentially problematic release.
This version of cobweb-launcher might be problematic. Click here for more details.
- cobweb/crawlers/base_crawler.py +5 -5
- {cobweb_launcher-1.1.17.dist-info → cobweb_launcher-1.1.19.dist-info}/METADATA +1 -1
- {cobweb_launcher-1.1.17.dist-info → cobweb_launcher-1.1.19.dist-info}/RECORD +6 -6
- {cobweb_launcher-1.1.17.dist-info → cobweb_launcher-1.1.19.dist-info}/LICENSE +0 -0
- {cobweb_launcher-1.1.17.dist-info → cobweb_launcher-1.1.19.dist-info}/WHEEL +0 -0
- {cobweb_launcher-1.1.17.dist-info → cobweb_launcher-1.1.19.dist-info}/top_level.txt +0 -0
cobweb/crawlers/base_crawler.py
CHANGED
|
@@ -69,7 +69,7 @@ class Crawler(threading.Thread):
|
|
|
69
69
|
continue
|
|
70
70
|
|
|
71
71
|
elif seed.params.retry >= self.max_retries:
|
|
72
|
-
seed.params.
|
|
72
|
+
seed.params.seed_status = DealModel.fail
|
|
73
73
|
self.launcher_queue['done'].push(seed)
|
|
74
74
|
continue
|
|
75
75
|
|
|
@@ -83,13 +83,11 @@ class Crawler(threading.Thread):
|
|
|
83
83
|
|
|
84
84
|
for request_item in request_iterators:
|
|
85
85
|
|
|
86
|
-
if isinstance(request_item,
|
|
87
|
-
self.upload_queue.push(request_item)
|
|
88
|
-
|
|
89
|
-
elif isinstance(request_item, Request):
|
|
86
|
+
if isinstance(request_item, Request):
|
|
90
87
|
download_iterators = self.download(request_item)
|
|
91
88
|
if not isgenerator(download_iterators):
|
|
92
89
|
raise TypeError("download function isn't a generator")
|
|
90
|
+
|
|
93
91
|
for download_item in download_iterators:
|
|
94
92
|
if isinstance(download_item, Response):
|
|
95
93
|
logger.info(LogTemplate.download_info.format(
|
|
@@ -110,6 +108,8 @@ class Crawler(threading.Thread):
|
|
|
110
108
|
self.distribute(parse_item, seed)
|
|
111
109
|
else:
|
|
112
110
|
self.distribute(download_item, seed)
|
|
111
|
+
else:
|
|
112
|
+
self.distribute(request_item, seed)
|
|
113
113
|
except Exception as e:
|
|
114
114
|
logger.info(LogTemplate.download_exception.format(
|
|
115
115
|
detail=seed_detail_log_info,
|
|
@@ -10,7 +10,7 @@ cobweb/base/request.py,sha256=tEkgMVUfdQI-kZuzWuiit9P_q4Q9-_RZh9aXXpc0314,2352
|
|
|
10
10
|
cobweb/base/response.py,sha256=eB1DWMXFCpn3cJ3yzgCRU1WeZAdayGDohRgdjdMUFN4,406
|
|
11
11
|
cobweb/base/seed.py,sha256=Uz_VBRlAxNYQcFHk3tsZFMlU96yPOedHaWGTvk-zKd8,2908
|
|
12
12
|
cobweb/crawlers/__init__.py,sha256=Rr3DTjD-abMA1_FYcQJZYNvQvcWMuEVcsIU6duqHrw4,75
|
|
13
|
-
cobweb/crawlers/base_crawler.py,sha256=
|
|
13
|
+
cobweb/crawlers/base_crawler.py,sha256=6_v71I-opx6JhSKFWa7qF__ivzKwRwAgwGI8nwx3YMQ,5154
|
|
14
14
|
cobweb/crawlers/file_crawler.py,sha256=2Sjbdgxzqd41WykKUQE3QQlGai3T8k-pmHNmPlTchjQ,4454
|
|
15
15
|
cobweb/db/__init__.py,sha256=ut0iEyBLjcJL06WNG_5_d4hO5PJWvDrKWMkDOdmgh2M,30
|
|
16
16
|
cobweb/db/redis_db.py,sha256=NNI2QkRV1hEZI-z-COEncXt88z3pZN6wusKlcQzc8V4,4304
|
|
@@ -25,8 +25,8 @@ cobweb/pipelines/loghub_pipeline.py,sha256=cjPO6w6UJ0jNw2fVvdX0BCdlm58T7dmYXlxzX
|
|
|
25
25
|
cobweb/utils/__init__.py,sha256=JTE4sBfHnKHhD6w9Auk0MIT7O9BMOamCeryhlHNx3Zg,47
|
|
26
26
|
cobweb/utils/oss.py,sha256=qAl05ybL2Jp6KFjHDHVMfmeBHQmDKPtZleHjHAY7LZc,3277
|
|
27
27
|
cobweb/utils/tools.py,sha256=bVd3iRGBvwhohQAH7AXTTjbmQ54Z35K0O-fatEyhePU,1249
|
|
28
|
-
cobweb_launcher-1.1.
|
|
29
|
-
cobweb_launcher-1.1.
|
|
30
|
-
cobweb_launcher-1.1.
|
|
31
|
-
cobweb_launcher-1.1.
|
|
32
|
-
cobweb_launcher-1.1.
|
|
28
|
+
cobweb_launcher-1.1.19.dist-info/LICENSE,sha256=z1rxSIGOyzcSb3orZxFPxzx-0C1vTocmswqBNxpKfEk,1063
|
|
29
|
+
cobweb_launcher-1.1.19.dist-info/METADATA,sha256=e-o4DJ4zQ8qbHuR3lllCJOGcAdxACG_je1w_o5hcyTs,1246
|
|
30
|
+
cobweb_launcher-1.1.19.dist-info/WHEEL,sha256=ewwEueio1C2XeHTvT17n8dZUJgOvyCWCt0WVNLClP9o,92
|
|
31
|
+
cobweb_launcher-1.1.19.dist-info/top_level.txt,sha256=4GETBGNsKqiCUezmT-mJn7tjhcDlu7nLIV5gGgHBW4I,7
|
|
32
|
+
cobweb_launcher-1.1.19.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|
|
File without changes
|