arkindex-base-worker 0.3.7rc10__tar.gz → 0.4.0a1__tar.gz

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arkindex-base-worker
-Version: 0.3.7rc10
+Version: 0.4.0a1
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -37,28 +37,22 @@ Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
-Classifier: Topic :: Text Processing :: Linguistic
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: peewee==3.17.0
-Requires-Dist: Pillow==10.2.0
-Requires-Dist: pymdown-extensions==10.7
+Requires-Dist: peewee~=3.17
+Requires-Dist: Pillow==10.3.0
 Requires-Dist: python-gnupg==0.5.2
 Requires-Dist: shapely==2.0.3
-Requires-Dist: teklia-toolbox==0.1.4
+Requires-Dist: teklia-toolbox==0.1.5
 Requires-Dist: zstandard==0.22.0
 Provides-Extra: docs
-Requires-Dist: black==24.2.0; extra == "docs"
-Requires-Dist: doc8==1.1.1; extra == "docs"
-Requires-Dist: mkdocs==1.5.3; extra == "docs"
-Requires-Dist: mkdocs-material==9.5.10; extra == "docs"
-Requires-Dist: mkdocstrings==0.24.0; extra == "docs"
-Requires-Dist: mkdocstrings-python==1.8.0; extra == "docs"
-Requires-Dist: recommonmark==0.7.1; extra == "docs"
+Requires-Dist: black==24.4.0; extra == "docs"
+Requires-Dist: mkdocs-material==9.5.17; extra == "docs"
+Requires-Dist: mkdocstrings-python==1.9.2; extra == "docs"
 Provides-Extra: tests
-Requires-Dist: pytest==8.0.1; extra == "tests"
-Requires-Dist: pytest-mock==3.12.0; extra == "tests"
+Requires-Dist: pytest==8.1.1; extra == "tests"
+Requires-Dist: pytest-mock==3.14.0; extra == "tests"
 Requires-Dist: pytest-responses==0.5.1; extra == "tests"
 
 # Arkindex base Worker
@@ -73,7 +67,7 @@ The [documentation](https://workers.arkindex.org/) is made with [Material for Mk
 
 ## Create a new worker using our template
 
-```
+```shell
 pip install --user cookiecutter
 cookiecutter git@gitlab.teklia.com:workers/base-worker.git
 ```

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/README.md

@@ -10,7 +10,7 @@ The [documentation](https://workers.arkindex.org/) is made with [Material for Mk
 
 ## Create a new worker using our template
 
-```
+```shell
 pip install --user cookiecutter
 cookiecutter git@gitlab.teklia.com:workers/base-worker.git
 ```

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_base_worker.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arkindex-base-worker
-Version: 0.3.7rc10
+Version: 0.4.0a1
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -37,28 +37,22 @@ Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
-Classifier: Topic :: Text Processing :: Linguistic
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: peewee==3.17.0
-Requires-Dist: Pillow==10.2.0
-Requires-Dist: pymdown-extensions==10.7
+Requires-Dist: peewee~=3.17
+Requires-Dist: Pillow==10.3.0
 Requires-Dist: python-gnupg==0.5.2
 Requires-Dist: shapely==2.0.3
-Requires-Dist: teklia-toolbox==0.1.4
+Requires-Dist: teklia-toolbox==0.1.5
 Requires-Dist: zstandard==0.22.0
 Provides-Extra: docs
-Requires-Dist: black==24.2.0; extra == "docs"
-Requires-Dist: doc8==1.1.1; extra == "docs"
-Requires-Dist: mkdocs==1.5.3; extra == "docs"
-Requires-Dist: mkdocs-material==9.5.10; extra == "docs"
-Requires-Dist: mkdocstrings==0.24.0; extra == "docs"
-Requires-Dist: mkdocstrings-python==1.8.0; extra == "docs"
-Requires-Dist: recommonmark==0.7.1; extra == "docs"
+Requires-Dist: black==24.4.0; extra == "docs"
+Requires-Dist: mkdocs-material==9.5.17; extra == "docs"
+Requires-Dist: mkdocstrings-python==1.9.2; extra == "docs"
 Provides-Extra: tests
-Requires-Dist: pytest==8.0.1; extra == "tests"
-Requires-Dist: pytest-mock==3.12.0; extra == "tests"
+Requires-Dist: pytest==8.1.1; extra == "tests"
+Requires-Dist: pytest-mock==3.14.0; extra == "tests"
 Requires-Dist: pytest-responses==0.5.1; extra == "tests"
 
 # Arkindex base Worker
@@ -73,7 +67,7 @@ The [documentation](https://workers.arkindex.org/) is made with [Material for Mk
 
 ## Create a new worker using our template
 
-```
+```shell
 pip install --user cookiecutter
 cookiecutter git@gitlab.teklia.com:workers/base-worker.git
 ```

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_base_worker.egg-info/SOURCES.txt

@@ -14,9 +14,11 @@ arkindex_worker/utils.py
 arkindex_worker/worker/__init__.py
 arkindex_worker/worker/base.py
 arkindex_worker/worker/classification.py
+arkindex_worker/worker/corpus.py
 arkindex_worker/worker/dataset.py
 arkindex_worker/worker/element.py
 arkindex_worker/worker/entity.py
+arkindex_worker/worker/image.py
 arkindex_worker/worker/metadata.py
 arkindex_worker/worker/task.py
 arkindex_worker/worker/training.py
@@ -35,9 +37,11 @@ tests/test_utils.py
 tests/test_elements_worker/__init__.py
 tests/test_elements_worker/test_classifications.py
 tests/test_elements_worker/test_cli.py
+tests/test_elements_worker/test_corpus.py
 tests/test_elements_worker/test_dataset.py
 tests/test_elements_worker/test_elements.py
 tests/test_elements_worker/test_entities.py
+tests/test_elements_worker/test_image.py
 tests/test_elements_worker/test_metadata.py
 tests/test_elements_worker/test_task.py
 tests/test_elements_worker/test_training.py

arkindex_base_worker-0.4.0a1/arkindex_base_worker.egg-info/requires.txt

@@ -0,0 +1,16 @@
+peewee~=3.17
+Pillow==10.3.0
+python-gnupg==0.5.2
+shapely==2.0.3
+teklia-toolbox==0.1.5
+zstandard==0.22.0
+
+[docs]
+black==24.4.0
+mkdocs-material==9.5.17
+mkdocstrings-python==1.9.2
+
+[tests]
+pytest==8.1.1
+pytest-mock==3.14.0
+pytest-responses==0.5.1

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_worker/models.py

@@ -261,6 +261,12 @@ class Transcription(ArkindexModel):
     """
 
 
+class Image(ArkindexModel):
+    """
+    Describes an Arkindex image.
+    """
+
+
 class Dataset(ArkindexModel):
     """
     Describes an Arkindex dataset.

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_worker/utils.py

@@ -5,7 +5,6 @@ import tarfile
 import tempfile
 from pathlib import Path
 
-import zstandard
 import zstandard as zstd
 
 logger = logging.getLogger(__name__)
@@ -38,7 +37,7 @@ def decompress_zst_archive(compressed_archive: Path) -> tuple[int, Path]:
     :param compressed_archive: Path to the target ZST-compressed archive
     :return: File descriptor and path to the uncompressed tar archive
     """
-    dctx = zstandard.ZstdDecompressor()
+    dctx = zstd.ZstdDecompressor()
     archive_fd, archive_path = tempfile.mkstemp(suffix=".tar")
     archive_path = Path(archive_path)
 
@@ -50,7 +49,7 @@ def decompress_zst_archive(compressed_archive: Path) -> tuple[int, Path]:
         ):
             dctx.copy_stream(compressed, decompressed)
         logger.debug(f"Successfully uncompressed archive {compressed_archive}")
-    except zstandard.ZstdError as e:
+    except zstd.ZstdError as e:
         raise Exception(f"Couldn't uncompressed archive: {e}") from e
 
     return archive_fd, archive_path
@@ -129,7 +128,7 @@ def zstd_compress(
                 archive_hasher.update(compressed_chunk)
                 archive_file.write(compressed_chunk)
         logger.debug(f"Successfully compressed {source}")
-    except zstandard.ZstdError as e:
+    except zstd.ZstdError as e:
         raise Exception(f"Couldn't compress archive: {e}") from e
     return file_d, destination, archive_hasher.hexdigest()
 

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_worker/worker/__init__.py

@@ -19,9 +19,11 @@ from arkindex_worker.cache import CachedElement
 from arkindex_worker.models import Dataset, Element, Set
 from arkindex_worker.worker.base import BaseWorker
 from arkindex_worker.worker.classification import ClassificationMixin
+from arkindex_worker.worker.corpus import CorpusMixin
 from arkindex_worker.worker.dataset import DatasetMixin, DatasetState
 from arkindex_worker.worker.element import ElementMixin
 from arkindex_worker.worker.entity import EntityMixin
+from arkindex_worker.worker.image import ImageMixin
 from arkindex_worker.worker.metadata import MetaDataMixin, MetaType  # noqa: F401
 from arkindex_worker.worker.task import TaskMixin
 from arkindex_worker.worker.transcription import TranscriptionMixin
@@ -57,11 +59,13 @@ class ActivityState(Enum):
 class ElementsWorker(
     BaseWorker,
     ClassificationMixin,
+    CorpusMixin,
     ElementMixin,
     TranscriptionMixin,
     WorkerVersionMixin,
     EntityMixin,
     MetaDataMixin,
+    ImageMixin,
 ):
     """
     Base class for ML workers that operate on Arkindex elements.
@@ -88,7 +92,7 @@ class ElementsWorker(
         )
         self.parser.add_argument(
             "--element",
-            type=uuid.UUID,
+            type=str,
             nargs="+",
             help="One or more Arkindex element ID",
         )
@@ -107,11 +111,23 @@ class ElementsWorker(
         the cache database when enabled.
 
         :return: An iterable of [CachedElement][arkindex_worker.cache.CachedElement] when cache support is enabled,
-           and a list of strings representing element IDs otherwise.
+           or a list of strings representing element IDs otherwise.
         """
         assert not (
             self.args.elements_list and self.args.element
         ), "elements-list and element CLI args shouldn't be both set"
+
+        def invalid_element_id(value: str) -> bool:
+            """
+            Return whether the ID of an element is a valid UUID or not
+            """
+            try:
+                uuid.UUID(value)
+            except Exception:
+                return True
+
+            return False
+
         out = []
 
         # Load from the cache when available
@@ -131,6 +147,11 @@ class ElementsWorker(
         elif self.args.element:
             out += self.args.element
 
+        invalid_element_ids = list(filter(invalid_element_id, out))
+        assert (
+            not invalid_element_ids
+        ), f"These element IDs are invalid: {', '.join(invalid_element_ids)}"
+
         return out
 
     @property

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_worker/worker/base.py

@@ -13,14 +13,8 @@ from tempfile import mkdtemp
 import gnupg
 import yaml
 from apistar.exceptions import ErrorResponse
-from tenacity import (
-    before_sleep_log,
-    retry,
-    retry_if_exception,
-    stop_after_attempt,
-    wait_exponential,
-)
 
+from arkindex import options_from_env
 from arkindex_worker import logger
 from arkindex_worker.cache import (
     check_version,
@@ -30,7 +24,7 @@ from arkindex_worker.cache import (
     merge_parents_cache,
 )
 from arkindex_worker.utils import close_delete_file, extract_tar_zst_archive
-from teklia_toolbox.requests import _get_arkindex_client, _is_500_error
+from teklia_toolbox.requests import get_arkindex_client
 
 
 class ExtrasDirNotFoundError(Exception):
@@ -185,7 +179,7 @@ class BaseWorker:
         Create an ArkindexClient to make API requests towards Arkindex instances.
         """
         # Build Arkindex API client from environment variables
-        self.api_client = _get_arkindex_client()
+        self.api_client = get_arkindex_client(**options_from_env())
         logger.debug(f"Setup Arkindex API client on {self.api_client.document.url}")
 
     def configure_for_developers(self):
@@ -477,23 +471,9 @@ class BaseWorker:
             # Clean up
             shutil.rmtree(base_extracted_path)
 
-    @retry(
-        retry=retry_if_exception(_is_500_error),
-        wait=wait_exponential(multiplier=2, min=3),
-        reraise=True,
-        stop=stop_after_attempt(5),
-        before_sleep=before_sleep_log(logger, logging.INFO),
-    )
     def request(self, *args, **kwargs):
         """
         Wrapper around the ``ArkindexClient.request`` method.
-
-        The API call will be retried up to 5 times in case of HTTP 5xx errors,
-        with an exponential sleep time of 3, 4, 8 and 16 seconds between calls.
-        If the 5th call still causes an HTTP 5xx error, the exception is re-raised
-        and the caller should catch it.
-
-        Log messages are displayed when an HTTP 5xx error occurs, before waiting for the next call.
         """
         return self.api_client.request(*args, **kwargs)
 

arkindex_base_worker-0.4.0a1/arkindex_worker/worker/corpus.py

@@ -0,0 +1,69 @@
+"""
+BaseWorker methods for corpora.
+"""
+
+from enum import Enum
+from operator import itemgetter
+from tempfile import _TemporaryFileWrapper
+
+from arkindex_worker import logger
+
+
+class CorpusExportState(Enum):
+    """
+    State of a corpus export.
+    """
+
+    Created = "created"
+    """
+    The corpus export is created, awaiting its processing.
+    """
+
+    Running = "running"
+    """
+    The corpus export is being built.
+    """
+
+    Failed = "failed"
+    """
+    The corpus export failed.
+    """
+
+    Done = "done"
+    """
+    The corpus export ended in success.
+    """
+
+
+class CorpusMixin:
+    def download_latest_export(self) -> _TemporaryFileWrapper:
+        """
+        Download the latest export in `done` state of the current corpus.
+
+        :returns: The downloaded export stored in a temporary file.
+        """
+        # List all exports on the corpus
+        exports = self.api_client.paginate("ListExports", id=self.corpus_id)
+
+        # Find the latest that is in "done" state
+        exports: list[dict] = sorted(
+            list(
+                filter(
+                    lambda export: export["state"] == CorpusExportState.Done.value,
+                    exports,
+                )
+            ),
+            key=itemgetter("updated"),
+            reverse=True,
+        )
+        assert (
+            len(exports) > 0
+        ), f'No available exports found for the corpus ({self.corpus_id}) with state "{CorpusExportState.Done.value.capitalize()}".'
+
+        # Download latest export
+        export_id: str = exports[0]["id"]
+        logger.info(f"Downloading export ({export_id})...")
+        export: _TemporaryFileWrapper = self.request("DownloadExport", id=export_id)
+        logger.info(f"Downloaded export ({export_id}) @ `{export.name}`")
+
+        return export

arkindex_base_worker-0.4.0a1/arkindex_worker/worker/image.py

@@ -0,0 +1,21 @@
+"""
+ElementsWorker methods for images.
+"""
+
+from arkindex_worker.models import Image
+
+
+class ImageMixin:
+    def create_iiif_url(self, url: str) -> Image:
+        """
+        Create an image from an existing IIIF image by URL.
+        The URL should be of the image's identifier, not of its Image Information request (`/info.json`).
+
+        :param url: URL of the image.
+        :returns: The created image.
+        """
+        assert url and isinstance(
+            url, str
+        ), "url shouldn't be null and should be of type str"
+
+        return Image(self.api_client.request("CreateIIIFURL", body={"url": url}))

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/arkindex_worker/worker/training.py

@@ -83,6 +83,9 @@ class TrainingMixin:
 
     @property
     def is_finetuning(self) -> bool:
+        """
+        Whether or not this worker is fine-tuning an existing model version.
+        """
         return bool(self.model_version_id)
 
     @skip_if_read_only
@@ -280,8 +283,17 @@ class TrainingMixin:
                 },
             )
         except ErrorResponse as e:
+            # Temporary fix while waiting for `ValidateModelVersion` refactoring as it can
+            # return errors even when the model version is properly validated
+            if e.status_code in [403, 500]:
+                logger.warning(
+                    f'An error occurred while validating model version {self.model_version["id"]}, please check its status.'
+                )
+                return
+
             if e.status_code != 409:
                 raise e
+
             logger.warning(
                 f"An available model version exists with hash {hash}, using it instead of the pending version."
             )

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/pyproject.toml

@@ -4,16 +4,15 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "arkindex-base-worker"
-version = "0.3.7rc10"
+version = "0.4.0a1"
 description = "Base Worker to easily build Arkindex ML workflows"
 license = { file = "LICENSE" }
 dependencies = [
-    "peewee==3.17.0",
-    "Pillow==10.2.0",
-    "pymdown-extensions==10.7",
+    "peewee~=3.17",
+    "Pillow==10.3.0",
     "python-gnupg==0.5.2",
     "shapely==2.0.3",
-    "teklia-toolbox==0.1.4",
+    "teklia-toolbox==0.1.5",
     "zstandard==0.22.0",
 ]
 authors = [
@@ -32,8 +31,6 @@ classifiers = [
     "Programming Language :: Python :: 3 :: Only",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
-    # Topics
-    "Topic :: Text Processing :: Linguistic",
 ]
 
 [project.urls]
@@ -45,17 +42,13 @@ Authors = "https://teklia.com"
 
 [project.optional-dependencies]
 docs = [
-    "black==24.2.0",
-    "doc8==1.1.1",
-    "mkdocs==1.5.3",
-    "mkdocs-material==9.5.10",
-    "mkdocstrings==0.24.0",
-    "mkdocstrings-python==1.8.0",
-    "recommonmark==0.7.1",
+    "black==24.4.0",
+    "mkdocs-material==9.5.17",
+    "mkdocstrings-python==1.9.2",
 ]
 tests = [
-    "pytest==8.0.1",
-    "pytest-mock==3.12.0",
+    "pytest==8.1.1",
+    "pytest-mock==3.14.0",
     "pytest-responses==0.5.1",
 ]
 
@@ -114,7 +107,3 @@ known-third-party = [
     "yaml",
     "zstandard",
 ]
-
-[tool.doc8]
-ignore-path=["*.egg-info", "public", ".git"]
-extensions=[".md"]

arkindex_base_worker-0.4.0a1/tests/__init__.py

@@ -0,0 +1,8 @@
+from pathlib import Path
+
+BASE_DIR = Path(__file__).resolve().parent
+FIXTURES_DIR = BASE_DIR / "data"
+SAMPLES_DIR = BASE_DIR / "samples"
+
+CORPUS_ID = "11111111-1111-1111-1111-111111111111"
+PROCESS_ID = "cafecafe-cafe-cafe-cafe-cafecafecafe"

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/tests/conftest.py

@@ -26,11 +26,7 @@ from arkindex_worker.models import Artifact, Dataset, Set
 from arkindex_worker.worker import BaseWorker, DatasetWorker, ElementsWorker
 from arkindex_worker.worker.dataset import DatasetState
 from arkindex_worker.worker.transcription import TextOrientation
-
-FIXTURES_DIR = Path(__file__).resolve().parent / "data"
-SAMPLES_DIR = Path(__file__).resolve().parent / "samples"
-
-PROCESS_ID = "cafecafe-cafe-cafe-cafe-cafecafecafe"
+from tests import CORPUS_ID, FIXTURES_DIR, PROCESS_ID, SAMPLES_DIR
 
 __yaml_cache = {}
 
@@ -93,7 +89,7 @@ def _setup_api(responses, monkeypatch, _cache_yaml):
 
     # Fallback to prod environment
     if schema_url is None:
-        schema_url = "https://arkindex.teklia.com/api/v1/openapi/?format=openapi-json"
+        schema_url = "https://arkindex.teklia.com/api/v1/openapi/?format=json"
         monkeypatch.setenv("ARKINDEX_API_SCHEMA_URL", schema_url)
 
     # Allow accessing remote API schemas
@@ -159,7 +155,7 @@ def _mock_worker_run_api(responses):
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
             "state": "running",
             "mode": "workers",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
             "use_cache": False,
             "activity_state": "ready",
             "model_id": None,
@@ -226,7 +222,7 @@ def _mock_worker_run_no_revision_api(responses):
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
             "state": "running",
             "mode": "workers",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
             "use_cache": False,
             "activity_state": "ready",
             "model_id": None,

{arkindex-base-worker-0.3.7rc10 → arkindex_base_worker-0.4.0a1}/tests/test_base_worker.py

@@ -11,7 +11,7 @@ from arkindex.mock import MockApiClient
 from arkindex_worker import logger
 from arkindex_worker.worker import BaseWorker, ElementsWorker
 from arkindex_worker.worker.base import ExtrasDirNotFoundError
-from tests.conftest import FIXTURES_DIR
+from tests import CORPUS_ID, FIXTURES_DIR
 
 
 def test_init_default_local_share():
@@ -178,7 +178,7 @@ def test_configure_worker_run(mocker, responses, caplog):
         "model_version": None,
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -270,7 +270,7 @@ def test_configure_user_configuration_defaults(mocker, responses):
         "model_version": None,
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -319,7 +319,7 @@ def test_configure_user_config_debug(mocker, responses, debug):
         },
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -367,7 +367,7 @@ def test_configure_worker_run_missing_conf(mocker, responses):
         "configuration": {"id": "bbbbbbbbb-bbbb-bbbb-bbbb-bbbbbbbbbbbb", "name": "BBB"},
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -409,7 +409,7 @@ def test_configure_worker_run_no_worker_run_conf(mocker, responses):
         "configuration": None,
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -458,7 +458,7 @@ def test_configure_load_model_configuration(mocker, responses):
         },
         "process": {
             "id": "aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeffff",
-            "corpus": "11111111-1111-1111-1111-111111111111",
+            "corpus": CORPUS_ID,
         },
         "summary": "Worker Fake worker @ 123412",
     }
@@ -669,8 +669,7 @@ def test_find_parents_file_paths(responses, mock_base_worker_with_cache, tmp_pat
     ):
         (tmp_path / parent_id).mkdir()
         file_path = tmp_path / parent_id / filename
-        with file_path.open("w", encoding="utf-8") as f:
-            f.write(content)
+        file_path.write_text(content)
 
     # Configure worker with a specific data directory
     mock_base_worker_with_cache.task_data_dir = tmp_path