datashare-python 0.9.4__tar.gz → 0.9.7__tar.gz

{datashare_python-0.9.4 → datashare_python-0.9.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: datashare-python
-Version: 0.9.4
+Version: 0.9.7
 Summary: Manage Python tasks and local resources in Datashare
 Project-URL: Homepage, https://icij.github.io/datashare-python/
 Project-URL: Documentation, https://icij.github.io/datashare-python/
@@ -14,8 +14,6 @@ Requires-Dist: hatchling~=1.27
 Requires-Dist: icij-common[elasticsearch]~=0.8.2
 Requires-Dist: langcodes~=3.5
 Requires-Dist: lru-dict~=1.4
-Requires-Dist: nest-asyncio~=1.6
-Requires-Dist: orjson~=3.11
 Requires-Dist: pydantic-extra-types[pycountry]>=2.11.1
 Requires-Dist: python-json-logger~=4.0
 Requires-Dist: pyyaml~=6.0

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/conftest.py

@@ -1,7 +1,5 @@
-import asyncio
 import shutil
-from asyncio import AbstractEventLoop
-from collections.abc import AsyncGenerator, Generator, Iterator, Sequence
+from collections.abc import AsyncGenerator, Generator, Sequence
 from pathlib import Path
 
 import aiohttp
@@ -10,7 +8,9 @@ import pytest
 from elasticsearch._async.helpers import async_streaming_bulk
 from icij_common.es import DOC_ROOT_ID, ES_DOCUMENT_TYPE, ID, ESClient
 from icij_common.test_utils import reset_env  # noqa: F401
+from pytest_asyncio import is_async_test
 from temporalio import workflow
+from temporalio.service import RPCError, RPCStatusCode
 
 from datashare_python.config import (
     DatashareClientConfig,
@@ -59,6 +59,13 @@ _INDEX_BODY = {
 }
 
 
+def pytest_collection_modifyitems(items: list) -> None:
+    pytest_asyncio_tests = (item for item in items if is_async_test(item))
+    session_scope_marker = pytest.mark.asyncio(loop_scope="session")
+    for async_test in pytest_asyncio_tests:
+        async_test.add_marker(session_scope_marker, append=False)
+
+
 @activity_defn(name="mocked-act")
 def mocked_act() -> None:
     pass
@@ -81,16 +88,6 @@ def test_deps() -> list[ContextManagerFactory]:
     return [set_es_client, set_task_client]
 
 
-@pytest.fixture(scope="session")
-def event_loop(
-    request: pytest.FixtureRequest,  # noqa: ARG001
-) -> Iterator[asyncio.AbstractEventLoop]:
-    """Create an instance of the default event loop for each test case."""
-    loop = asyncio.get_event_loop_policy().new_event_loop()
-    yield loop
-    loop.close()
-
-
 @pytest.fixture(scope="session")
 def test_worker_config() -> WorkerConfig:
     logging_config = LoggingConfig(
@@ -117,9 +114,7 @@ def test_worker_config_path(test_worker_config: WorkerConfig, tmpdir: Path) -> P
 
 @pytest.fixture(scope="session")
 async def worker_lifetime_deps(
-    event_loop: AbstractEventLoop,
-    test_deps: list[ContextManagerFactory],
-    test_worker_config: WorkerConfig,
+    test_deps: list[ContextManagerFactory], test_worker_config: WorkerConfig
 ) -> AsyncGenerator[None, None]:
     worker_id = "test-worker-id"
     ctx = "test application"
@@ -128,7 +123,6 @@ async def worker_lifetime_deps(
         ctx=ctx,
         worker_id=worker_id,
         worker_config=test_worker_config,
-        event_loop=event_loop,
     ):
         yield
 
@@ -174,11 +168,25 @@ async def test_task_client(
 @pytest.fixture(scope="session")
 async def test_temporal_client_session(
     test_worker_config: WorkerConfig,
-    event_loop: AbstractEventLoop,  # noqa: ARG001
 ) -> TemporalClient:  # noqa: ANN001
     return await test_worker_config.to_temporal_client()
 
 
+@pytest.fixture
+async def test_temporal_client(
+    test_temporal_client_session: TemporalClient,
+) -> TemporalClient:  # noqa: ANN001
+    client = test_temporal_client_session
+    async for wf in client.list_workflows():
+        try:
+            await client.get_workflow_handle(wf.id).terminate()
+        except RPCError as e:
+            if e.status != RPCStatusCode.NOT_FOUND:
+                raise
+
+    return client
+
+
 @pytest.fixture
 async def populate_es(
     test_es_client: ESClient,

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/discovery.py

@@ -88,7 +88,9 @@ def discover(
         worker_config_cls = discover_worker_config_cls()
     else:
         worker_config_cls = WorkerConfig
-    discovered += f"- worker config class: {worker_config_cls}"
+    if discovered:
+        discovered += "\n"
+    discovered += f"- worker config class: {worker_config_cls.__name__}"
     logger.info("discovered:\n%s", discovered)
     return wfs, acts, deps, worker_config_cls
 

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/interceptors.py

@@ -6,7 +6,7 @@ from collections.abc import Callable, Generator, Mapping
 from contextlib import contextmanager
 from contextvars import ContextVar
 from copy import deepcopy
-from functools import partial
+from functools import partial, wraps
 from inspect import signature
 from types import UnionType
 from typing import (
@@ -52,7 +52,12 @@ from temporalio.workflow import (
 )
 
 from .objects import BaseModel
-from .types_ import ProgressRateHandler, Weight
+from .types_ import (
+    AsyncProgressRateHandler,
+    ProgressRateHandler,
+    SyncProgressRateHandler,
+    Weight,
+)
 from .utils import (
     PROGRESS_HANDLER_ARG,
     PYDANTIC_DATA_CONVERTER,
@@ -62,6 +67,11 @@ from .utils import (
 
 _TRACEPARENT = "traceparent"
 _DEFAULT_PAYLOAD_CONVERTER = DataConverter.default.payload_converter
+_PROGRESS_TYPES = {
+    ProgressRateHandler,
+    AsyncProgressRateHandler,
+    SyncProgressRateHandler,
+}
 
 
 class TraceContext(BaseModel):
@@ -292,11 +302,13 @@ def _get_progress_handler(act_fn: Callable) -> ProgressRateHandler:
 
 
 def _is_progress(t: type) -> bool:
-    if t is ProgressRateHandler:
+    if any(t is p_cls for p_cls in _PROGRESS_TYPES):
         return True
     return bool(
         isinstance(t, UnionType)
-        and any(sub_t is ProgressRateHandler for sub_t in get_args(t))
+        and any(
+            any(sub_t is p_cls for p_cls in _PROGRESS_TYPES) for sub_t in get_args(t)
+        )
     )
 
 
@@ -315,16 +327,21 @@ class _ProgressInboundInterceptor(ActivityInboundInterceptor):
         # https://github.com/temporalio/sdk-python/blob/631ebaf0e20fb214b16589b45627b358048a5d77/temporalio/worker/_activity.py#L600
         # we have to force it here again
         progress_handler = _get_progress_handler(input.fn)
+        new_args = []
+        act_definition = _Definition.must_from_callable(input.fn)
         if input.args:
             data_converter = PYDANTIC_DATA_CONVERTER
-            arg_types = _Definition.must_from_callable(input.fn).arg_types
+            arg_types = act_definition.arg_types
             arg_types = _without_progress(arg_types)
             arg_types = arg_types[: len(input.args)]
             encoded = await data_converter.encode(input.args)
             new_args = await data_converter.decode(encoded, type_hints=arg_types)
-            new_args.append(progress_handler)
-        else:
-            new_args = [progress_handler]
+        injected_progress = (
+            progress_handler
+            if act_definition.is_async
+            else _sync_progress(progress_handler)
+        )
+        new_args.append(injected_progress)
         new_input = dataclasses.replace(input, args=new_args)
         await progress_handler(0.0)
         res = await super().execute_activity(new_input)
@@ -374,3 +391,15 @@ class _HeartbeatInboundInterceptor(ActivityInboundInterceptor):
             if heartbeat_task:
                 heartbeat_task.cancel()
                 await asyncio.wait([heartbeat_task])
+
+
+def _sync_progress(
+    progress_handler: AsyncProgressRateHandler,
+) -> SyncProgressRateHandler:
+    @wraps(progress_handler)
+    def p(progress: float, event_loop: asyncio.AbstractEventLoop) -> None:
+        asyncio.run_coroutine_threadsafe(
+            progress_handler(progress), event_loop
+        ).result()
+
+    return p

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/logging_.py

@@ -1,10 +1,10 @@
+import json
 import logging
 import numbers
 import sys
 from copy import copy
 from typing import Any
 
-import orjson
 from icij_common.logging_utils import DATE_FMT, STREAM_HANDLER_FMT
 from pythonjsonlogger.core import BaseJsonFormatter
 from pythonjsonlogger.orjson import OrjsonFormatter
@@ -67,8 +67,6 @@ class WorkerFilter(logging.Filter):
         self.worker_id = worker_id
 
     def filter(self, record: logging.LogRecord) -> bool:
-        if not hasattr(record, "message") and record.msg is not None:
-            setattr(record, "msg", record.getMessage())  # noqa: B010
         if self.worker_id is not None:
             record.worker_id = self.worker_id
         if workflow.in_workflow():
@@ -134,7 +132,7 @@ def _encode_value(value: Any) -> str:
         return "true" if value else "false"
     if isinstance(value, numbers.Number):
         return str(value)
-    return orjson.dumps(value).decode()
+    return json.dumps(value).decode()
 
 
 def _json_formatter(datefmt: str) -> BaseJsonFormatter:

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/types_.py

@@ -1,3 +1,4 @@
+import asyncio
 from collections.abc import Coroutine
 from contextlib import AbstractAsyncContextManager, AbstractContextManager
 from dataclasses import dataclass
@@ -8,19 +9,34 @@ from temporalio.client import Client
 TemporalClient = Client
 
 
-class ProgressRateHandler(Protocol):
+class AsyncProgressRateHandler(Protocol):
     async def __call__(self, progress_rate: float) -> None:
         pass
 
 
+class SyncProgressRateHandler(Protocol):
+    def __call__(
+        self, progress_rate: float, event_loop: asyncio.AbstractEventLoop
+    ) -> None:
+        pass
+
+
+ProgressRateHandler = SyncProgressRateHandler | AsyncProgressRateHandler
+
+
 @dataclass
 class Weight:
     value: float
 
 
-class RawProgressHandler(Protocol):
-    async def __call__(self, iteration: int) -> None:
-        pass
+class RawAsyncProgressHandler(Protocol):
+    async def __call__(self, iteration: int) -> None: ...
+
+
+class RawSyncProgressHandler(Protocol):
+    async def __call__(
+        self, iteration: int, event_loop: asyncio.AbstractEventLoop
+    ) -> None: ...
 
 
 FactoryReturnType = (

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/utils.py

@@ -15,7 +15,6 @@ from pathlib import Path
 from typing import Any, ParamSpec, TypeVar
 from uuid import uuid4
 
-import nest_asyncio
 import temporalio
 from pydantic import ValidationError
 from temporalio import activity, workflow
@@ -31,9 +30,15 @@ from temporalio.converter import (
 )
 from temporalio.exceptions import ApplicationError
 
+from datashare_python.types_ import (
+    AsyncProgressRateHandler,
+    RawSyncProgressHandler,
+    SyncProgressRateHandler,
+)
+
 from .constants import METADATA_JSON
 from .objects import DocArtifact, DocumentLocation, FilesystemDocument
-from .types_ import ProgressRateHandler, RawProgressHandler
+from .types_ import RawAsyncProgressHandler
 
 DependencyLabel = str | None
 DependencySetup = Callable[..., None]
@@ -76,10 +81,13 @@ class ProgressSignal:
 
 
 class ActivityWithProgress:
-    def __init__(self, temporal_client: Client, event_loop: asyncio.AbstractEventLoop):
+    def __init__(
+        self,
+        temporal_client: Client,
+        event_loop: asyncio.AbstractEventLoop | None = None,
+    ):
         self._temporal_client = temporal_client
-        nest_asyncio.apply()
-        self._event_loop = event_loop
+        self._event_loop = event_loop or asyncio.get_event_loop()
 
 
 class WorkflowWithProgress:
@@ -276,9 +284,9 @@ def fatal_error_from_exception(exc: Exception) -> ApplicationError:
     return ApplicationError(str(exc), details, type=exc_type, non_retryable=True)
 
 
-def to_raw_progress(
-    progress: ProgressRateHandler, max_progress: int
-) -> RawProgressHandler:
+def to_raw_async_progress(
+    progress: AsyncProgressRateHandler, max_progress: int
+) -> RawAsyncProgressHandler:
     if not max_progress > 0:
         raise ValueError("max_progress must be > 0")
 
@@ -288,9 +296,21 @@ def to_raw_progress(
     return raw
 
 
-def to_scaled_progress(
-    progress: ProgressRateHandler, *, start: float = 0.0, end: float = 1.0
-) -> ProgressRateHandler:
+def to_raw_sync_progress(
+    progress: SyncProgressRateHandler, max_progress: int
+) -> RawSyncProgressHandler:
+    if not max_progress > 0:
+        raise ValueError("max_progress must be > 0")
+
+    def raw(iteration: int, event_loop: asyncio.AbstractEventLoop) -> None:
+        progress(iteration / max_progress, event_loop)
+
+    return raw
+
+
+def to_scaled_async_progress(
+    progress: AsyncProgressRateHandler, *, start: float = 0.0, end: float = 1.0
+) -> AsyncProgressRateHandler:
     if not 0 <= start < end:
         raise ValueError("start must be [0, end[")
     if not start < end <= 1.0:

{datashare_python-0.9.4 → datashare_python-0.9.7}/datashare_python/worker.py

@@ -1,3 +1,4 @@
+import asyncio
 import inspect
 import logging
 import os
@@ -157,7 +158,7 @@ async def worker_context(
     workflows: list[type] | None = None,
     worker_config: WorkerConfig,
     client: TemporalClient,
-    event_loop: AbstractEventLoop,
+    event_loop: AbstractEventLoop | None = None,
     task_queue: str,
     dependencies: list[ContextManagerFactory] | None = None,
     sandboxed: bool = True,
@@ -169,6 +170,8 @@ async def worker_context(
         discovered.extend(workflows)
     if dependencies is not None:
         discovered.extend(dependencies)
+    if event_loop is None:
+        event_loop = asyncio.get_event_loop()
     discovered.append(worker_config)
     loggers = copy(worker_config.logging.loggers)
     discovered_loggers = {_get_object_package(o).__name__ for o in discovered}

{datashare_python-0.9.4 → datashare_python-0.9.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "datashare-python"
-version = "0.9.4"
+version = "0.9.7"
 description = "Manage Python tasks and local resources in Datashare"
 authors = [
   { name = "Clément Doumouro", email = "cdoumouro@icij.org" },
@@ -15,8 +15,6 @@ dependencies = [
   "hatchling~=1.27",
   "icij-common[elasticsearch]~=0.8.2",
   "langcodes~=3.5",
-  "nest-asyncio~=1.6",
-  "orjson~=3.11",
   "python-json-logger~=4.0",
   "pyyaml~=6.0",
   "pydantic-extra-types[pycountry]>=2.11.1",
@@ -76,6 +74,7 @@ dev = [
 
 [tool.pytest.ini_options]
 asyncio_mode = "auto"
+asyncio_debug = true
 asyncio_default_fixture_loop_scope = "session"
 markers = [
   "integration",