apify 2.4.0b2__py3-none-any.whl → 2.4.0b4__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 apify might be problematic. Click here for more details.
- apify/_actor.py +38 -9
- apify/_crypto.py +38 -0
- apify/_platform_event_manager.py +3 -3
- apify/apify_storage_client/_key_value_store_client.py +17 -2
- {apify-2.4.0b2.dist-info → apify-2.4.0b4.dist-info}/METADATA +2 -2
- {apify-2.4.0b2.dist-info → apify-2.4.0b4.dist-info}/RECORD +8 -8
- {apify-2.4.0b2.dist-info → apify-2.4.0b4.dist-info}/WHEEL +0 -0
- {apify-2.4.0b2.dist-info → apify-2.4.0b4.dist-info}/licenses/LICENSE +0 -0
apify/_actor.py
CHANGED
|
@@ -3,6 +3,7 @@ from __future__ import annotations
|
|
|
3
3
|
import asyncio
|
|
4
4
|
import os
|
|
5
5
|
import sys
|
|
6
|
+
from contextlib import suppress
|
|
6
7
|
from datetime import timedelta
|
|
7
8
|
from typing import TYPE_CHECKING, Any, Callable, Literal, TypeVar, cast, overload
|
|
8
9
|
|
|
@@ -64,6 +65,7 @@ class _ActorType:
|
|
|
64
65
|
configuration: Configuration | None = None,
|
|
65
66
|
*,
|
|
66
67
|
configure_logging: bool = True,
|
|
68
|
+
exit_process: bool | None = None,
|
|
67
69
|
) -> None:
|
|
68
70
|
"""Create an Actor instance.
|
|
69
71
|
|
|
@@ -74,7 +76,10 @@ class _ActorType:
|
|
|
74
76
|
configuration: The Actor configuration to be used. If not passed, a new Configuration instance will
|
|
75
77
|
be created.
|
|
76
78
|
configure_logging: Should the default logging configuration be configured?
|
|
79
|
+
exit_process: Whether the Actor should call `sys.exit` when the context manager exits. The default is
|
|
80
|
+
True except for the IPython, Pytest and Scrapy environments.
|
|
77
81
|
"""
|
|
82
|
+
self._exit_process = self._get_default_exit_process() if exit_process is None else exit_process
|
|
78
83
|
self._is_exiting = False
|
|
79
84
|
|
|
80
85
|
self._configuration = configuration or Configuration.get_global_configuration()
|
|
@@ -141,9 +146,19 @@ class _ActorType:
|
|
|
141
146
|
|
|
142
147
|
return super().__repr__()
|
|
143
148
|
|
|
144
|
-
def __call__(
|
|
149
|
+
def __call__(
|
|
150
|
+
self,
|
|
151
|
+
configuration: Configuration | None = None,
|
|
152
|
+
*,
|
|
153
|
+
configure_logging: bool = True,
|
|
154
|
+
exit_process: bool | None = None,
|
|
155
|
+
) -> Self:
|
|
145
156
|
"""Make a new Actor instance with a non-default configuration."""
|
|
146
|
-
return self.__class__(
|
|
157
|
+
return self.__class__(
|
|
158
|
+
configuration=configuration,
|
|
159
|
+
configure_logging=configure_logging,
|
|
160
|
+
exit_process=exit_process,
|
|
161
|
+
)
|
|
147
162
|
|
|
148
163
|
@property
|
|
149
164
|
def apify_client(self) -> ApifyClientAsync:
|
|
@@ -281,13 +296,7 @@ class _ActorType:
|
|
|
281
296
|
await asyncio.wait_for(finalize(), cleanup_timeout.total_seconds())
|
|
282
297
|
self._is_initialized = False
|
|
283
298
|
|
|
284
|
-
if
|
|
285
|
-
self.log.debug(f'Not calling sys.exit({exit_code}) because Actor is running in IPython')
|
|
286
|
-
elif os.getenv('PYTEST_CURRENT_TEST', default=False): # noqa: PLW1508
|
|
287
|
-
self.log.debug(f'Not calling sys.exit({exit_code}) because Actor is running in an unit test')
|
|
288
|
-
elif os.getenv('SCRAPY_SETTINGS_MODULE'):
|
|
289
|
-
self.log.debug(f'Not calling sys.exit({exit_code}) because Actor is running with Scrapy')
|
|
290
|
-
else:
|
|
299
|
+
if self._exit_process:
|
|
291
300
|
sys.exit(exit_code)
|
|
292
301
|
|
|
293
302
|
async def fail(
|
|
@@ -1128,6 +1137,26 @@ class _ActorType:
|
|
|
1128
1137
|
|
|
1129
1138
|
return proxy_configuration
|
|
1130
1139
|
|
|
1140
|
+
def _get_default_exit_process(self) -> bool:
|
|
1141
|
+
"""Returns False for IPython, Pytest, and Scrapy environments, True otherwise."""
|
|
1142
|
+
if is_running_in_ipython():
|
|
1143
|
+
self.log.debug('Running in IPython, setting default `exit_process` to False.')
|
|
1144
|
+
return False
|
|
1145
|
+
|
|
1146
|
+
# Check if running in Pytest by detecting the relevant environment variable.
|
|
1147
|
+
if os.getenv('PYTEST_CURRENT_TEST'):
|
|
1148
|
+
self.log.debug('Running in Pytest, setting default `exit_process` to False.')
|
|
1149
|
+
return False
|
|
1150
|
+
|
|
1151
|
+
# Check if running in Scrapy by attempting to import it.
|
|
1152
|
+
with suppress(ImportError):
|
|
1153
|
+
import scrapy # noqa: F401
|
|
1154
|
+
|
|
1155
|
+
self.log.debug('Running in Scrapy, setting default `exit_process` to False.')
|
|
1156
|
+
return False
|
|
1157
|
+
|
|
1158
|
+
return True
|
|
1159
|
+
|
|
1131
1160
|
|
|
1132
1161
|
Actor = cast(_ActorType, Proxy(_ActorType))
|
|
1133
1162
|
"""The entry point of the SDK, through which all the Actor operations should be done."""
|
apify/_crypto.py
CHANGED
|
@@ -1,6 +1,9 @@
|
|
|
1
1
|
from __future__ import annotations
|
|
2
2
|
|
|
3
3
|
import base64
|
|
4
|
+
import hashlib
|
|
5
|
+
import hmac
|
|
6
|
+
import string
|
|
4
7
|
from typing import Any
|
|
5
8
|
|
|
6
9
|
from cryptography.exceptions import InvalidTag as InvalidTagException
|
|
@@ -153,3 +156,38 @@ def decrypt_input_secrets(private_key: rsa.RSAPrivateKey, input_data: Any) -> An
|
|
|
153
156
|
)
|
|
154
157
|
|
|
155
158
|
return input_data
|
|
159
|
+
|
|
160
|
+
|
|
161
|
+
CHARSET = string.digits + string.ascii_letters
|
|
162
|
+
|
|
163
|
+
|
|
164
|
+
def encode_base62(num: int) -> str:
|
|
165
|
+
"""Encode the given number to base62."""
|
|
166
|
+
if num == 0:
|
|
167
|
+
return CHARSET[0]
|
|
168
|
+
|
|
169
|
+
res = ''
|
|
170
|
+
while num > 0:
|
|
171
|
+
num, remainder = divmod(num, 62)
|
|
172
|
+
res = CHARSET[remainder] + res
|
|
173
|
+
return res
|
|
174
|
+
|
|
175
|
+
|
|
176
|
+
@ignore_docs
|
|
177
|
+
def create_hmac_signature(secret_key: str, message: str) -> str:
|
|
178
|
+
"""Generate an HMAC signature and encodes it using Base62. Base62 encoding reduces the signature length.
|
|
179
|
+
|
|
180
|
+
HMAC signature is truncated to 30 characters to make it shorter.
|
|
181
|
+
|
|
182
|
+
Args:
|
|
183
|
+
secret_key: Secret key used for signing signatures.
|
|
184
|
+
message: Message to be signed.
|
|
185
|
+
|
|
186
|
+
Returns:
|
|
187
|
+
Base62 encoded signature.
|
|
188
|
+
"""
|
|
189
|
+
signature = hmac.new(secret_key.encode('utf-8'), message.encode('utf-8'), hashlib.sha256).hexdigest()[:30]
|
|
190
|
+
|
|
191
|
+
decimal_signature = int(signature, 16)
|
|
192
|
+
|
|
193
|
+
return encode_base62(decimal_signature)
|
apify/_platform_event_manager.py
CHANGED
|
@@ -4,7 +4,7 @@ import asyncio
|
|
|
4
4
|
from datetime import datetime
|
|
5
5
|
from typing import TYPE_CHECKING, Annotated, Any, Literal, Union
|
|
6
6
|
|
|
7
|
-
import websockets.client
|
|
7
|
+
import websockets.asyncio.client
|
|
8
8
|
from pydantic import BaseModel, Discriminator, Field, TypeAdapter
|
|
9
9
|
from typing_extensions import Self, Unpack, override
|
|
10
10
|
|
|
@@ -143,7 +143,7 @@ class PlatformEventManager(EventManager):
|
|
|
143
143
|
but instead use it via the `Actor.on()` and `Actor.off()` methods.
|
|
144
144
|
"""
|
|
145
145
|
|
|
146
|
-
_platform_events_websocket: websockets.client.
|
|
146
|
+
_platform_events_websocket: websockets.asyncio.client.ClientConnection | None = None
|
|
147
147
|
_process_platform_messages_task: asyncio.Task | None = None
|
|
148
148
|
_send_system_info_interval_task: asyncio.Task | None = None
|
|
149
149
|
_connected_to_platform_websocket: asyncio.Future = asyncio.Future()
|
|
@@ -196,7 +196,7 @@ class PlatformEventManager(EventManager):
|
|
|
196
196
|
|
|
197
197
|
async def _process_platform_messages(self, ws_url: str) -> None:
|
|
198
198
|
try:
|
|
199
|
-
async with websockets.client.connect(ws_url) as websocket:
|
|
199
|
+
async with websockets.asyncio.client.connect(ws_url) as websocket:
|
|
200
200
|
self._platform_events_websocket = websocket
|
|
201
201
|
self._connected_to_platform_websocket.set_result(True)
|
|
202
202
|
|
|
@@ -4,10 +4,13 @@ from contextlib import asynccontextmanager
|
|
|
4
4
|
from typing import TYPE_CHECKING, Any
|
|
5
5
|
|
|
6
6
|
from typing_extensions import override
|
|
7
|
+
from yarl import URL
|
|
7
8
|
|
|
8
9
|
from crawlee.storage_clients._base import KeyValueStoreClient as BaseKeyValueStoreClient
|
|
9
10
|
from crawlee.storage_clients.models import KeyValueStoreListKeysPage, KeyValueStoreMetadata, KeyValueStoreRecord
|
|
10
11
|
|
|
12
|
+
from apify._crypto import create_hmac_signature
|
|
13
|
+
|
|
11
14
|
if TYPE_CHECKING:
|
|
12
15
|
from collections.abc import AsyncIterator
|
|
13
16
|
from contextlib import AbstractAsyncContextManager
|
|
@@ -89,6 +92,18 @@ class KeyValueStoreClient(BaseKeyValueStoreClient):
|
|
|
89
92
|
Args:
|
|
90
93
|
key: The key for which the URL should be generated.
|
|
91
94
|
"""
|
|
92
|
-
|
|
95
|
+
if self._client.resource_id is None:
|
|
96
|
+
raise ValueError('resource_id cannot be None when generating a public URL')
|
|
97
|
+
|
|
98
|
+
public_url = (
|
|
99
|
+
URL(self._api_public_base_url) / 'v2' / 'key-value-stores' / self._client.resource_id / 'records' / key
|
|
100
|
+
)
|
|
101
|
+
|
|
102
|
+
key_value_store = await self.get()
|
|
103
|
+
|
|
104
|
+
if key_value_store is not None and isinstance(key_value_store.model_extra, dict):
|
|
105
|
+
url_signing_secret_key = key_value_store.model_extra.get('urlSigningSecretKey')
|
|
106
|
+
if url_signing_secret_key:
|
|
107
|
+
public_url = public_url.with_query(signature=create_hmac_signature(url_signing_secret_key, key))
|
|
93
108
|
|
|
94
|
-
return
|
|
109
|
+
return str(public_url)
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: apify
|
|
3
|
-
Version: 2.4.
|
|
3
|
+
Version: 2.4.0b4
|
|
4
4
|
Summary: Apify SDK for Python
|
|
5
5
|
Project-URL: Homepage, https://docs.apify.com/sdk/python/
|
|
6
6
|
Project-URL: Apify homepage, https://apify.com
|
|
@@ -231,7 +231,7 @@ Requires-Dist: httpx>=0.27.0
|
|
|
231
231
|
Requires-Dist: lazy-object-proxy>=1.10.0
|
|
232
232
|
Requires-Dist: more-itertools>=10.2.0
|
|
233
233
|
Requires-Dist: typing-extensions>=4.1.0
|
|
234
|
-
Requires-Dist: websockets
|
|
234
|
+
Requires-Dist: websockets>=14.0
|
|
235
235
|
Provides-Extra: scrapy
|
|
236
236
|
Requires-Dist: scrapy>=2.11.0; extra == 'scrapy'
|
|
237
237
|
Description-Content-Type: text/markdown
|
|
@@ -1,11 +1,11 @@
|
|
|
1
1
|
apify/__init__.py,sha256=HpgKg2FZWJuSPfDygzJ62psylhw4NN4tKFnoYUIhcd4,838
|
|
2
|
-
apify/_actor.py,sha256=
|
|
2
|
+
apify/_actor.py,sha256=PQqFDpAqSbh_aP3EjD8yLGYOmLZMo1qLzqrbVT2KjWE,49697
|
|
3
3
|
apify/_charging.py,sha256=m7hJIQde4M7vS4g_4hsNRP5xHNXjYQ8MyqOEGeNb7VY,12267
|
|
4
4
|
apify/_configuration.py,sha256=yidcWHsu-IJ2mmLmXStKq_HHcdfQxZq7koYjlZfRnQ8,11128
|
|
5
5
|
apify/_consts.py,sha256=_Xq4hOfOA1iZ3n1P967YWdyncKivpbX6RTlp_qanUoE,330
|
|
6
|
-
apify/_crypto.py,sha256=
|
|
6
|
+
apify/_crypto.py,sha256=8BgeQC0ZhYP5KdmLxxLQAW87Gq-Z4HlREbYGXr46w0U,6607
|
|
7
7
|
apify/_models.py,sha256=-Y0rljBJWxMMCp8iDCTG4UV3bEvNZzp-kx2SYbPfeIY,7919
|
|
8
|
-
apify/_platform_event_manager.py,sha256=
|
|
8
|
+
apify/_platform_event_manager.py,sha256=k1e5ruSJdcMKr6j-_XIF-gfhrgyMzdSenYW0QoJROu8,7916
|
|
9
9
|
apify/_proxy_configuration.py,sha256=c-O6_PZ9pUD-i4J0RFEKTtfyJPP2rTRJJA1TH8NVsV8,13189
|
|
10
10
|
apify/_utils.py,sha256=92byxeXTpDFwhBq7ZS-obeXKtKWvVzCZMV0Drg3EjhQ,1634
|
|
11
11
|
apify/log.py,sha256=j-E4t-WeA93bc1NCQRG8sTntehQCiiN8ia-MdQe3_Ts,1291
|
|
@@ -14,7 +14,7 @@ apify/apify_storage_client/__init__.py,sha256=-UbR68bFsDR6ln8OFs4t50eqcnY36hujO-
|
|
|
14
14
|
apify/apify_storage_client/_apify_storage_client.py,sha256=qeWYsEQGeyyhJzS9TZTQFNqdSl8JzHz_4_HDKGY4I_Y,2736
|
|
15
15
|
apify/apify_storage_client/_dataset_client.py,sha256=9RxxhrJMic5QRJn2Vl4J-FnSlEigIpYW5Z_2B1dcRzM,5597
|
|
16
16
|
apify/apify_storage_client/_dataset_collection_client.py,sha256=gf5skMTkfpGhEscRy5bgo13vznxGZrSd7w9Ivh3Usyc,1516
|
|
17
|
-
apify/apify_storage_client/_key_value_store_client.py,sha256=
|
|
17
|
+
apify/apify_storage_client/_key_value_store_client.py,sha256=OCFUAW0o-8KQvUpL8zmlZrpU3yRmDKdsO2529H2v40I,4002
|
|
18
18
|
apify/apify_storage_client/_key_value_store_collection_client.py,sha256=zjsbRW4zjme6dIzxxlHyCW3voBA5489MUhdjl5YMaro,1596
|
|
19
19
|
apify/apify_storage_client/_request_queue_client.py,sha256=cNMhXz85s1ZtjLpVqkduYl1y6o9QyNdcIGoy6ccD-h0,5178
|
|
20
20
|
apify/apify_storage_client/_request_queue_collection_client.py,sha256=MTLM2cG0txAe3cSjkGbXyq2Ek0R7wlsMbGGULmQGD3I,1603
|
|
@@ -36,7 +36,7 @@ apify/scrapy/pipelines/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSu
|
|
|
36
36
|
apify/storages/__init__.py,sha256=FW-z6ubuPnHGM-Wp15T8mR5q6lnpDGrCW-IkgZd5L30,177
|
|
37
37
|
apify/storages/_request_list.py,sha256=7WpcdWvT3QxEBthynBpTVCSNDLXq6UbpQQmfUVyJ1jE,5849
|
|
38
38
|
apify/storages/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
39
|
-
apify-2.4.
|
|
40
|
-
apify-2.4.
|
|
41
|
-
apify-2.4.
|
|
42
|
-
apify-2.4.
|
|
39
|
+
apify-2.4.0b4.dist-info/METADATA,sha256=euAPCSZDsD14y2c1WnbKu9qMyXYHLQkeHZJ2Pzpnxvk,21558
|
|
40
|
+
apify-2.4.0b4.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
|
|
41
|
+
apify-2.4.0b4.dist-info/licenses/LICENSE,sha256=AsFjHssKjj4LGd2ZCqXn6FBzMqcWdjQre1byPPSypVw,11355
|
|
42
|
+
apify-2.4.0b4.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|