re-common 10.0.30__py3-none-any.whl → 10.0.32__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.
- re_common/v2/baselibrary/tools/hdfs_base_processor.py +11 -2
- re_common/v2/baselibrary/tools/resume_tracker.py +1 -1
- {re_common-10.0.30.dist-info → re_common-10.0.32.dist-info}/METADATA +1 -1
- {re_common-10.0.30.dist-info → re_common-10.0.32.dist-info}/RECORD +7 -7
- {re_common-10.0.30.dist-info → re_common-10.0.32.dist-info}/LICENSE +0 -0
- {re_common-10.0.30.dist-info → re_common-10.0.32.dist-info}/WHEEL +0 -0
- {re_common-10.0.30.dist-info → re_common-10.0.32.dist-info}/top_level.txt +0 -0
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
import abc
|
|
2
2
|
import asyncio
|
|
3
|
+
import sys
|
|
3
4
|
from concurrent.futures import ProcessPoolExecutor
|
|
4
5
|
import gzip
|
|
5
6
|
|
|
@@ -17,6 +18,7 @@ from re_common.v2.baselibrary.tools.resume_tracker import ResumeTracker
|
|
|
17
18
|
|
|
18
19
|
|
|
19
20
|
_pool = None
|
|
21
|
+
_loop = None
|
|
20
22
|
|
|
21
23
|
|
|
22
24
|
class HDFSBaseProcessor(abc.ABC):
|
|
@@ -169,8 +171,15 @@ class HDFSBaseProcessor(abc.ABC):
|
|
|
169
171
|
def _process_file_wrapper(self, args):
|
|
170
172
|
"""为多进程执行准备的同步包装函数"""
|
|
171
173
|
hdfs_file_path, process_func = args
|
|
172
|
-
|
|
173
|
-
|
|
174
|
+
if sys.platform == "win32":
|
|
175
|
+
loop = asyncio.get_event_loop()
|
|
176
|
+
return loop.run_until_complete(self._retry_process_file(hdfs_file_path, process_func))
|
|
177
|
+
else:
|
|
178
|
+
global _loop
|
|
179
|
+
if _loop is None:
|
|
180
|
+
_loop = asyncio.new_event_loop()
|
|
181
|
+
asyncio.set_event_loop(_loop)
|
|
182
|
+
return _loop.run_until_complete(self._retry_process_file(hdfs_file_path, process_func))
|
|
174
183
|
|
|
175
184
|
async def _run_multi_process(self, gz_files, process_func):
|
|
176
185
|
"""多进程并发运行文件处理任务"""
|
|
@@ -11,7 +11,7 @@ class ResumeTracker:
|
|
|
11
11
|
self,
|
|
12
12
|
db_path: Union[str, Path] = "processed.db",
|
|
13
13
|
timeout: float = 10.0,
|
|
14
|
-
isolation_level: Literal["DEFERRED", "EXCLUSIVE", "IMMEDIATE"]
|
|
14
|
+
isolation_level: Union[Literal["DEFERRED", "EXCLUSIVE", "IMMEDIATE"], None] = "DEFERRED",
|
|
15
15
|
):
|
|
16
16
|
self.db_path = Path(db_path)
|
|
17
17
|
self.timeout = timeout
|
|
@@ -179,12 +179,12 @@ re_common/v2/baselibrary/tools/ac_ahocorasick.py,sha256=c63y5RtKVLD37nyPCnBqfNyg
|
|
|
179
179
|
re_common/v2/baselibrary/tools/concurrency.py,sha256=ctKBoeEbq1SGmhPp7oVR_QSXTKVAGLMByAHQKI6dAqU,981
|
|
180
180
|
re_common/v2/baselibrary/tools/dict_tools.py,sha256=eSMwPTLp3oSjuviC_wlXg0I-dnkkmZfUfCRLX5djWV8,1365
|
|
181
181
|
re_common/v2/baselibrary/tools/dolphinscheduler.py,sha256=1m7UGYDiuvJUCI6ik6CGM2fO8U5XteJzn55VRbwB9ts,7978
|
|
182
|
-
re_common/v2/baselibrary/tools/hdfs_base_processor.py,sha256=
|
|
182
|
+
re_common/v2/baselibrary/tools/hdfs_base_processor.py,sha256=PFTyhth9RBoMTxsOzUxpyEfQegRi_Sd-C0lvIEOCxCg,8746
|
|
183
183
|
re_common/v2/baselibrary/tools/hdfs_bulk_processor.py,sha256=8FjuZbcBXC_27zEQonLKvb0fMGwcIH9MPRTOmdMAuLU,2396
|
|
184
184
|
re_common/v2/baselibrary/tools/hdfs_data_processer.py,sha256=g0DaNjXM1hIUblFQ6YBwnwEBKIXn48X8Y9Eiok4dVlQ,14824
|
|
185
185
|
re_common/v2/baselibrary/tools/hdfs_line_processor.py,sha256=h1J_mOPoNvsjw7zYMsD7rr0Q6bXvVzo9tRJZVAbei1s,2732
|
|
186
186
|
re_common/v2/baselibrary/tools/list_tools.py,sha256=1NxGVM4EytSXh4IGAEfZQnvq0Ev-UOF-PGZBg2EQbOg,2132
|
|
187
|
-
re_common/v2/baselibrary/tools/resume_tracker.py,sha256=
|
|
187
|
+
re_common/v2/baselibrary/tools/resume_tracker.py,sha256=rVvkd3-eF5lX0VesIHc3YjVRE6bsgAd1VT0G8Uq5DVM,3252
|
|
188
188
|
re_common/v2/baselibrary/tools/search_hash_tools.py,sha256=2ENLtZE8opRsfkwRtTNMzITmpTsjO7wZ1ZkfkqpOH9U,1937
|
|
189
189
|
re_common/v2/baselibrary/tools/text_matcher.py,sha256=cPMoFxaA0-ce3tLRxVSs8_3pTYS1oVIHDnNy_AlPU-4,10756
|
|
190
190
|
re_common/v2/baselibrary/tools/unionfind_tools.py,sha256=VYHZZPXwBYljsm7TjV1B6iCgDn3O3btzNf9hMvQySVU,2965
|
|
@@ -236,8 +236,8 @@ re_common/vip/title/transform/TransformRegulationTitleToZt.py,sha256=LKRdIsWKues
|
|
|
236
236
|
re_common/vip/title/transform/TransformStandardTitleToZt.py,sha256=-fCKAbSBzXVyQDCE61CalvR9E_QzQMA08QOO_NePFNI,5563
|
|
237
237
|
re_common/vip/title/transform/TransformThesisTitleToZt.py,sha256=QS-uV0cQrpUFAcKucuJQ9Ue2VRQH-inmfn_X3IplfRo,5488
|
|
238
238
|
re_common/vip/title/transform/__init__.py,sha256=m83-CWyRq_VHPYHaALEQlmXrkTdrZ3e4B_kCfBYE-uc,239
|
|
239
|
-
re_common-10.0.
|
|
240
|
-
re_common-10.0.
|
|
241
|
-
re_common-10.0.
|
|
242
|
-
re_common-10.0.
|
|
243
|
-
re_common-10.0.
|
|
239
|
+
re_common-10.0.32.dist-info/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
|
|
240
|
+
re_common-10.0.32.dist-info/METADATA,sha256=dhCvbocMDtJ_7KmIPeQFMypp6LOSS-HQ6YovmjRwvI8,764
|
|
241
|
+
re_common-10.0.32.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
|
242
|
+
re_common-10.0.32.dist-info/top_level.txt,sha256=_H9H23zoLIalm1AIY_KYTVh_H0ZnmjxQIxsvXtLv45o,10
|
|
243
|
+
re_common-10.0.32.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|
|
File without changes
|