cobweb-launcher 1.2.66__py3-none-any.whl → 1.2.68__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.
- cobweb/crawlers/crawler.py +2 -19
- {cobweb_launcher-1.2.66.dist-info → cobweb_launcher-1.2.68.dist-info}/METADATA +1 -1
- {cobweb_launcher-1.2.66.dist-info → cobweb_launcher-1.2.68.dist-info}/RECORD +6 -6
- {cobweb_launcher-1.2.66.dist-info → cobweb_launcher-1.2.68.dist-info}/LICENSE +0 -0
- {cobweb_launcher-1.2.66.dist-info → cobweb_launcher-1.2.68.dist-info}/WHEEL +0 -0
- {cobweb_launcher-1.2.66.dist-info → cobweb_launcher-1.2.68.dist-info}/top_level.txt +0 -0
cobweb/crawlers/crawler.py
CHANGED
@@ -1,7 +1,6 @@
|
|
1
1
|
import json
|
2
|
-
import os
|
3
|
-
import threading
|
4
2
|
import time
|
3
|
+
import threading
|
5
4
|
import traceback
|
6
5
|
from inspect import isgenerator
|
7
6
|
from typing import Union, Callable, Mapping
|
@@ -108,7 +107,7 @@ class Crawler(threading.Thread):
|
|
108
107
|
elif seed.params.retry > self.max_retries:
|
109
108
|
seed.params.seed_status = DealModel.fail
|
110
109
|
if self._record_failed:
|
111
|
-
self.parse(Response(seed, "failed"))
|
110
|
+
self.parse(Response(seed, "failed", max_retries=True))
|
112
111
|
else:
|
113
112
|
self._delete_seed(seed)
|
114
113
|
continue
|
@@ -202,22 +201,6 @@ class Crawler(threading.Thread):
|
|
202
201
|
))
|
203
202
|
seed.params.retry += 1
|
204
203
|
self._set_seed(seed)
|
205
|
-
# time.sleep(self.time_sleep * seed.params.retry)
|
206
|
-
# except Exception as e:
|
207
|
-
# logger.info(LogTemplate.download_exception.format(
|
208
|
-
# detail=seed_detail_log_info,
|
209
|
-
# retry=seed.params.retry,
|
210
|
-
# priority=seed.params.priority,
|
211
|
-
# seed_version=seed.params.seed_version,
|
212
|
-
# identifier=seed.identifier or "",
|
213
|
-
# exception=''.join(traceback.format_exception(type(e), e, e.__traceback__))
|
214
|
-
# ))
|
215
|
-
# seed.params.retry += 1
|
216
|
-
# # self._todo.push(seed)
|
217
|
-
# self._set_seed(seed)
|
218
|
-
# # time.sleep(self.time_sleep * seed.params.retry)
|
219
|
-
finally:
|
220
|
-
time.sleep(0.1)
|
221
204
|
logger.info("spider thread close")
|
222
205
|
|
223
206
|
def run(self):
|
@@ -13,7 +13,7 @@ cobweb/base/response.py,sha256=eB1DWMXFCpn3cJ3yzgCRU1WeZAdayGDohRgdjdMUFN4,406
|
|
13
13
|
cobweb/base/seed.py,sha256=A-F1urjbE5hYNWTCwq3sUV4nrxlK_RGMoCmjBmIwYsI,3158
|
14
14
|
cobweb/crawlers/__init__.py,sha256=msvkB9mTpsgyj8JfNMsmwAcpy5kWk_2NrO1Adw2Hkw0,29
|
15
15
|
cobweb/crawlers/base_crawler.py,sha256=ee_WSDnPQpPTk6wlFuY2UEx5L3hcsAZFcr6i3GLSry8,5751
|
16
|
-
cobweb/crawlers/crawler.py,sha256=
|
16
|
+
cobweb/crawlers/crawler.py,sha256=_V2jJtrf13Wwg9PwOPtOCPRv0aBIq420JHutEz0Ci0E,8496
|
17
17
|
cobweb/crawlers/file_crawler.py,sha256=2Sjbdgxzqd41WykKUQE3QQlGai3T8k-pmHNmPlTchjQ,4454
|
18
18
|
cobweb/db/__init__.py,sha256=uZwSkd105EAwYo95oZQXAfofUKHVIAZZIPpNMy-hm2Q,56
|
19
19
|
cobweb/db/api_db.py,sha256=bDc5dJQxq4z04h70KUTHd0OqUOEY7Cm3wcNJZtTvJIM,3015
|
@@ -38,8 +38,8 @@ cobweb/utils/bloom.py,sha256=vng-YbKgh9HbtpAWYf_nkUSbfVTOj40aqUUejRYlsCU,1752
|
|
38
38
|
cobweb/utils/dotting.py,sha256=mVICaa26R-dQ4JGmPK-kkR6QjX38QiRewXZnGb2DCIc,1784
|
39
39
|
cobweb/utils/oss.py,sha256=6x_ugXanh1R-6ZylQzUDQh4OeFZHujhWFCOxbzy53JY,3984
|
40
40
|
cobweb/utils/tools.py,sha256=5JEaaAwYoV9Sdla2UBIJn6faUBuXmxUMagm9ck6FVqs,1253
|
41
|
-
cobweb_launcher-1.2.
|
42
|
-
cobweb_launcher-1.2.
|
43
|
-
cobweb_launcher-1.2.
|
44
|
-
cobweb_launcher-1.2.
|
45
|
-
cobweb_launcher-1.2.
|
41
|
+
cobweb_launcher-1.2.68.dist-info/LICENSE,sha256=z1rxSIGOyzcSb3orZxFPxzx-0C1vTocmswqBNxpKfEk,1063
|
42
|
+
cobweb_launcher-1.2.68.dist-info/METADATA,sha256=Rr1Pm1Xd3RDv-ldwwocn1S5OMAIZ4E7AfgBv1tWa5Ck,6510
|
43
|
+
cobweb_launcher-1.2.68.dist-info/WHEEL,sha256=ewwEueio1C2XeHTvT17n8dZUJgOvyCWCt0WVNLClP9o,92
|
44
|
+
cobweb_launcher-1.2.68.dist-info/top_level.txt,sha256=4GETBGNsKqiCUezmT-mJn7tjhcDlu7nLIV5gGgHBW4I,7
|
45
|
+
cobweb_launcher-1.2.68.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|