datashare-python 0.6.2__py3-none-any.whl → 0.7.0__py3-none-any.whl

datashare_python/config.py

@@ -1,5 +1,5 @@
 from pathlib import Path
-from typing import ClassVar
+from typing import Annotated, Literal
 
 from icij_common.es import ESClient
 from icij_common.pydantic_utils import ICIJSettings
@@ -18,7 +18,6 @@ import datashare_python
 from .objects import BaseModel
 from .task_client import DatashareTaskClient
 from .types_ import TemporalClient
-from .utils import LogWithWorkerIDMixin
 
 _ALL_LOGGERS = [datashare_python.__name__]
 
@@ -76,11 +75,20 @@ class TemporalClientConfig(BaseModel):
         return self._client
 
 
-class WorkerConfig(ICIJSettings, LogWithWorkerIDMixin, BaseModel):
+LogLevel = Literal["CRITICAL", "ERROR", "WARNING", "INFO", "DEBUG"]
+
+
+class LoggingConfig(BaseModel):
+    log_in_json: bool = False
+    loggers: dict[str, LogLevel]
+
+
+class WorkerConfig(ICIJSettings, BaseModel):
     model_config = DS_WORKER_SETTINGS_CONFIG
 
-    loggers: ClassVar[list[str]] = Field(_ALL_LOGGERS, frozen=True)
-    log_level: str = Field(default="INFO")
+    logging: Annotated[LoggingConfig, Field(frozen=True)] = {
+        datashare_python.__name__: "INFO"
+    }
 
     datashare: DatashareClientConfig = DatashareClientConfig()
     elasticsearch: ESClientConfig = ESClientConfig()

datashare_python/conftest.py

@@ -12,6 +12,7 @@ from temporalio import workflow
 
 from datashare_python.config import (
     DatashareClientConfig,
+    LoggingConfig,
     TemporalClientConfig,
     WorkerConfig,
 )
@@ -19,10 +20,7 @@ from datashare_python.dependencies import (
     lifespan_es_client,
     lifespan_task_client,
     set_es_client,
-    set_event_loop,
-    set_loggers,
     set_task_client,
-    set_temporal_client,
     with_dependencies,
 )
 from datashare_python.objects import Document, TaskState
@@ -78,13 +76,7 @@ class MockedWorkflow:
 
 @pytest.fixture(scope="session")
 def test_deps() -> list[ContextManagerFactory]:
-    return [
-        set_loggers,
-        set_event_loop,
-        set_es_client,
-        set_temporal_client,
-        set_task_client,
-    ]
+    return [set_es_client, set_task_client]
 
 
 @pytest.fixture(scope="session")
@@ -99,8 +91,16 @@ def event_loop(
 
 @pytest.fixture(scope="session")
 def test_worker_config() -> WorkerConfig:
+    logging_config = LoggingConfig(
+        log_in_json=False,
+        loggers={
+            "datashare_python": "DEBUG",
+            "icij_common": "DEBUG",
+            "worker_template": "DEBUG",
+        },
+    )
     return WorkerConfig(
-        log_level="DEBUG",
+        logging=logging_config,
         datashare=DatashareClientConfig(url="http://localhost:8080"),
         temporal=TemporalClientConfig(host="localhost:7233"),
     )

datashare_python/dependencies.py

@@ -9,8 +9,9 @@ from typing import Any
 
 from icij_common.es import ESClient
 
-from .config import WorkerConfig
+from .config import LogLevel, WorkerConfig
 from .exceptions import DependencyInjectionError
+from .logging_ import setup_worker_loggers
 from .task_client import DatashareTaskClient
 from .types_ import ContextManagerFactory, TemporalClient
 
@@ -35,10 +36,13 @@ def lifespan_event_loop() -> AbstractEventLoop:
         raise DependencyInjectionError("event loop") from e
 
 
-def set_loggers(worker_config: WorkerConfig, worker_id: str) -> None:
-    worker_config.setup_loggers(worker_id=worker_id)
+def set_loggers(
+    worker_config: WorkerConfig, worker_id: str, loggers: dict[str, LogLevel]
+) -> None:
+    setup_worker_loggers(
+        loggers=loggers, worker_id=worker_id, in_json=worker_config.logging.log_in_json
+    )
     logger.info("worker loggers ready to log 💬")
-    logger.info("app config: %s", worker_config.model_dump_json(indent=2))
 
 
 def set_worker_config(worker_config: WorkerConfig) -> None:

datashare_python/discovery.py

@@ -4,7 +4,7 @@ from collections.abc import Callable, Iterable
 from importlib.metadata import entry_points
 
 from .config import WorkerConfig
-from .dependencies import set_worker_config
+from .dependencies import set_loggers, set_worker_config
 from .types_ import ContextManagerFactory
 from .utils import ActivityWithProgress
 
@@ -29,6 +29,8 @@ _Discovery = tuple[
     type[WorkerConfig],
 ]
 
+_MANDATORY_DEPS = [set_worker_config, set_loggers]
+
 
 def discover(
     wf_names: list[str] | None, *, act_names: list[str] | None, deps_name: str | None
@@ -68,8 +70,9 @@ def discover(
     deps = []
     if deps_name is not None:
         deps = discover_dependencies(deps_name)
-    if set_worker_config not in deps:
-        deps.append(set_worker_config)
+    for mandatory in _MANDATORY_DEPS:
+        if mandatory not in deps:
+            deps.append(mandatory)
     if deps:
         n_deps = len(deps)
         discovered += "\n"

datashare_python/logging_.py

@@ -0,0 +1,87 @@
+import logging
+import sys
+from copy import copy
+
+from icij_common.logging_utils import (
+    DATE_FMT,
+    STREAM_HANDLER_FMT,
+    STREAM_HANDLER_FMT_WITH_WORKER_ID,
+)
+from pythonjsonlogger.core import RESERVED_ATTRS, BaseJsonFormatter
+from pythonjsonlogger.orjson import OrjsonFormatter
+from temporalio import activity, workflow
+
+from .config import LogLevel
+
+_ACT_LOGGER_ATTRS = [
+    "activity_type",
+    "activity_id",
+    "activity_run_id",
+]
+
+_WF_LOGGED_ATTRS = [
+    "workflow_type",
+    "workflow_id",
+    "workflow_run_id",
+]
+_LOGGED_ATTRIBUTES = (
+    copy(RESERVED_ATTRS) + _WF_LOGGED_ATTRS + _ACT_LOGGER_ATTRS + ["worker_id"]
+)
+
+
+def setup_worker_loggers(
+    loggers: dict[str, LogLevel], *, worker_id: str | None, in_json: bool
+) -> None:
+    worker_filter = WorkerFilter(worker_id)
+    for logger_name, level_str in loggers.items():
+        level = getattr(logging, level_str)
+        logger = logging.getLogger(logger_name)
+        logger.setLevel(level)
+        logger.handlers = []
+        for handler in _get_worker_handlers(level, worker_id, in_json=in_json):
+            logger.addHandler(handler)
+        logger.addFilter(worker_filter)
+
+
+def _get_worker_handlers(
+    level: int, worker_id: str | None, *, in_json: bool
+) -> list[logging.Handler]:
+    stream_handler = logging.StreamHandler(sys.stderr)
+    if in_json:
+        fmt = _json_formatter(datefmt=DATE_FMT, worker_id=worker_id)
+    else:
+        if worker_id is not None:
+            fmt = STREAM_HANDLER_FMT_WITH_WORKER_ID
+        else:
+            fmt = STREAM_HANDLER_FMT
+        fmt = logging.Formatter(fmt, DATE_FMT)
+    stream_handler.setFormatter(fmt)
+    stream_handler.setLevel(level)
+    return [stream_handler]
+
+
+class WorkerFilter(logging.Filter):
+    def __init__(self, worker_id: str) -> None:
+        super().__init__()
+        self._worker_id = worker_id
+
+    def filter(self, record: logging.LogRecord) -> bool:
+        record.worker_id = self._worker_id
+        if workflow.in_workflow():
+            wf_info = workflow.info()
+            for attr in _WF_LOGGED_ATTRS:
+                setattr(record, attr, getattr(wf_info, attr))
+        if activity.in_activity():
+            act_info = activity.info()
+            for attr in _ACT_LOGGER_ATTRS:
+                setattr(record, attr, getattr(act_info, attr))
+        return True
+
+
+def _json_formatter(datefmt: str, worker_id: str) -> BaseJsonFormatter:
+    fmt = OrjsonFormatter(  # let's keep logging as fast as possible
+        _LOGGED_ATTRIBUTES,
+        extra={"worker_id": worker_id},
+        datefmt=datefmt,
+    )
+    return fmt

datashare_python/objects.py

@@ -1,11 +1,12 @@
 import logging
+import os
 from collections.abc import Awaitable, Callable
 from dataclasses import dataclass
 from datetime import UTC, datetime
 from enum import StrEnum, unique
 from io import BytesIO
 from pathlib import Path
-from typing import Any, Literal, Self, TypeVar, cast
+from typing import Annotated, Any, Literal, Self, TypeVar, cast
 
 from temporalio import workflow
 
@@ -30,8 +31,8 @@ from icij_common.pydantic_utils import (
     merge_configs,
     no_enum_values_config,
 )
+from pydantic import AfterValidator, Field
 from pydantic import BaseModel as _BaseModel
-from pydantic import Field
 from pydantic.main import IncEx
 
 logger = logging.getLogger(__name__)
@@ -162,9 +163,17 @@ class DocumentLocation(StrEnum):
     WORKDIR = "workdir"
 
 
+def _is_relative(value: Path) -> Path:
+    if value.is_absolute():
+        raise ValueError(
+            f"FilesystemDocument path should always be relative, found {value}"
+        )
+    return value
+
+
 class FilesystemDocument(DatashareModel):
     id: str
-    path: Path
+    path: Annotated[Path, AfterValidator(_is_relative)]
     index: str
     location: DocumentLocation
     resource_name: str
@@ -174,11 +183,11 @@ class FilesystemDocument(DatashareModel):
     ) -> Path:
         from datashare_python.utils import artifacts_dir  # noqa: PLC0415
 
-        project = self.index
         match self.location:
             case DocumentLocation.ORIGINAL:
                 return original_root / self.path
             case DocumentLocation.ARTIFACTS:
+                project = self.index
                 return artifacts_root / artifacts_dir(self.id, project=project) / "raw"
             case DocumentLocation.WORKDIR:
                 return workdir / self.path
@@ -236,6 +245,10 @@ class Document(DatashareModel):
                 raise ValueError(msg)
             path = artifacts_dir(doc_id=self.id, project=self.index) / "raw"
             location = DocumentLocation.ARTIFACTS
+        # The filesystem dod is alway relative to the base location, let's make sure
+        # we store a relative path otherwise joining with the location will fail
+        if path.parts and path.parts[0] == os.path.sep:
+            path = Path(*path.parts[1:])
         return FilesystemDocument(
             id=self.id,
             path=path,

datashare_python/worker.py

@@ -8,6 +8,7 @@ from asyncio import AbstractEventLoop
 from collections.abc import AsyncGenerator, Callable
 from concurrent.futures import ThreadPoolExecutor
 from contextlib import asynccontextmanager
+from copy import copy
 from typing import Any
 
 from temporalio.worker import PollerBehaviorSimpleMaximum, Worker
@@ -142,12 +143,27 @@ async def worker_context(
     task_queue: str,
     dependencies: list[ContextManagerFactory] | None = None,
 ) -> AsyncGenerator[DatashareWorker, None]:
+    discovered = []
+    if activities is not None:
+        discovered.extend(activities)
+    if workflows is not None:
+        discovered.extend(workflows)
+    if dependencies is not None:
+        discovered.extend(dependencies)
+    discovered.append(worker_config)
+    loggers = copy(worker_config.logging.loggers)
+    discovered_loggers = {_get_object_package(o).__name__ for o in discovered}
+    for logger in discovered_loggers:
+        if logger not in loggers:
+            # Log in info by default
+            loggers[logger] = "INFO"
     deps_cm = (
         with_dependencies(
             dependencies,
             worker_config=worker_config,
             worker_id=worker_id,
             event_loop=event_loop,
+            loggers=loggers,
         )
         if dependencies
         else _do_nothing_cm()
@@ -181,3 +197,9 @@ def _get_class_from_method(method: Callable) -> type:
     class_name = method.__qualname__.rsplit(".", 1)[0]
     module = sys.modules[method.__module__]
     return getattr(module, class_name)
+
+
+def _get_object_package(obj: Any) -> Any:
+    mod = inspect.getmodule(obj)
+    base, _, _ = mod.__name__.partition(".")
+    return sys.modules[base]

{datashare_python-0.6.2.dist-info → datashare_python-0.7.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: datashare-python
-Version: 0.6.2
+Version: 0.7.0
 Summary: Manage Pythoœn tasks and local resources in Datashare
 Project-URL: Homepage, https://icij.github.io/datashare-python/
 Project-URL: Documentation, https://icij.github.io/datashare-python/
@@ -8,13 +8,14 @@ Project-URL: Repository, https://github.com/ICIJ/datashare-python
 Project-URL: Issues, https://github.com/ICIJ/datashare-python/issues
 Author-email: Clément Doumouro <cdoumouro@icij.org>, Clément Doumouro <clement.doumouro@gmail.com>, Lion Summerbell <lsummerbell@icij.org>
 Requires-Python: <4,>=3.11
-Requires-Dist: aiohttp~=3.11.9
-Requires-Dist: alive-progress~=3.2.0
-Requires-Dist: hatchling~=1.27.0
+Requires-Dist: aiohttp~=3.11
+Requires-Dist: alive-progress~=3.2
+Requires-Dist: hatchling~=1.27
 Requires-Dist: icij-common[elasticsearch]~=0.8.2
-Requires-Dist: nest-asyncio~=1.6.0
-Requires-Dist: python-json-logger~=4.0.0
+Requires-Dist: nest-asyncio~=1.6
+Requires-Dist: orjson~=3.11
+Requires-Dist: python-json-logger~=4.0
 Requires-Dist: pyyaml~=6.0
-Requires-Dist: temporalio~=1.23.0
+Requires-Dist: temporalio~=1.23
 Requires-Dist: tomlkit~=0.14.0
 Requires-Dist: typer~=0.15.4

{datashare_python-0.6.2.dist-info → datashare_python-0.7.0.dist-info}/RECORD

@@ -1,25 +1,26 @@
 datashare_python/.gitignore,sha256=e-SRgnvGGdsjRrqgKsTzALz6Obx8IYiOjr0yaAxT6v8,22
 datashare_python/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datashare_python/__main__.py,sha256=g-fvS46zl9umKmGrSpl-OG-8PSuZgjqvTCqjpsZtSps,101
-datashare_python/config.py,sha256=Mn43zAqUH8bmVpPQ4DK42ffg5LmRtC7ZHHgBlIINlD4,3814
-datashare_python/conftest.py,sha256=zUEB9d3CcYVVfsuEXkvF1G5JCEchTknB6pFUiM8FbXA,8512
+datashare_python/config.py,sha256=Q4iu3ZGaQB7npaHJWclxPIfgzZTf_-8VxxhXrB9nlpE,3928
+datashare_python/conftest.py,sha256=_VESUGax8wt0OlmdJmDFyTJlGO14bt1ccXsjHor0XVk,8592
 datashare_python/constants.py,sha256=a8-ceZKBVMXydcoNQ35fSjFjxeJ7dt-N6eAvqtPpf9g,320
-datashare_python/dependencies.py,sha256=4UsVFKRjd2Q0ghg_fUU24P26tFYhg_SnAENj2mKErrY,4060
-datashare_python/discovery.py,sha256=UsfIb_pL56BQ5i5xvfuhOO0bDefjA_oToxrCeJQQbcU,6925
+datashare_python/dependencies.py,sha256=KJuAp6Dmv8DQuFnGjbWiHu7StzZj97eBPDyZ_RfCQRc,4141
+datashare_python/discovery.py,sha256=BPB_Ak6d1-vcf9vAQA63IRb2U8h83_mIIi8MbKbFzQ0,7020
 datashare_python/exceptions.py,sha256=bVHEAXxDPKfxeeMC0hJXEsrJkgsKO2ESAhxWU96GA4M,496
-datashare_python/objects.py,sha256=mz025_6DogMMXSk5PqEctp0slvY4EKkuMDm80Kpm5_A,7020
+datashare_python/logging_.py,sha256=bhOI5ynVrTe2bU3jUNbaWXfe4zQEG_lLQxJ3Esrz9LQ,2589
+datashare_python/objects.py,sha256=pE0DGNNkl1etxz5ed7T-EaGo1o9TONjH2Lg9u1qdAWU,7571
 datashare_python/task_client.py,sha256=oTmP8bvZW0UyhLNMi1AV3XIAx7hrdbxNRss2Mw2azEc,8435
 datashare_python/template.py,sha256=RxKTYLXoS_EQ8Jc41JkBXppPdbCFqDWfP3BmC0gvB5o,4024
 datashare_python/types_.py,sha256=9Hk1XqpdXbM1TnEzwvJ5G9ABbaCZW9KgBTtiPBVn_7k,649
 datashare_python/utils.py,sha256=ZGZKO9Q4_aLVVilZUCkmHQ21M_37hVOCr7G-qZPOflU,17234
-datashare_python/worker-template.tar.gz,sha256=f_9jyCcZDTgQvzl7K3vD5Ved9MaHXUV9Ox_uZWzub10,274849
-datashare_python/worker.py,sha256=UdSCWZw8qhkhzok89oU7J35VTDZwxRAqef-5Z8yt95A,6333
+datashare_python/worker-template.tar.gz,sha256=Krs3td9sVkwXP9KPfIdTUgYWUNoU0G5YLGnn_cNxeTY,280218
+datashare_python/worker.py,sha256=M_I8AL3KZbKeA7-ObaoMUxz3pHBNslJD0XChOcmyafk,7076
 datashare_python/cli/__init__.py,sha256=9BPWtssDgsVfWMsZ1TtZCla0EC_kai4RHttr8oNLYOE,1401
 datashare_python/cli/project.py,sha256=w32Gy9AOL5B00uDT4in7YUCt2g68FnNbvwg2M3a8G6o,946
 datashare_python/cli/task.py,sha256=8mvKGS21bZ14BgZ0Uo-dfameljkaI2ZBha80ywCy-E8,5822
 datashare_python/cli/utils.py,sha256=p69CQb0zfixuyBkiZprhdMCc_NuYwXyAn6vC9H1UzAw,911
 datashare_python/cli/worker.py,sha256=I4KTpFIpXFowioFn72Rm6LBCYlY-Dhp4NBIPvtRgUXE,5283
-datashare_python-0.6.2.dist-info/METADATA,sha256=-tF3-ZSn0AA_xOJe-25-cJUnE-am9-yqDBRZqeNWbFs,907
-datashare_python-0.6.2.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
-datashare_python-0.6.2.dist-info/entry_points.txt,sha256=ILE7auxabHWiu3GC-AunWnzjhOI_SbZp7D4GqZHlLw4,68
-datashare_python-0.6.2.dist-info/RECORD,,
+datashare_python-0.7.0.dist-info/METADATA,sha256=GRZEpX-eTktdvsnSyrvyKzG9kgjqyW4tFs_FDVaEQWU,923
+datashare_python-0.7.0.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
+datashare_python-0.7.0.dist-info/entry_points.txt,sha256=ILE7auxabHWiu3GC-AunWnzjhOI_SbZp7D4GqZHlLw4,68
+datashare_python-0.7.0.dist-info/RECORD,,