arkindex-base-worker 0.4.0rc5__tar.gz → 0.4.0rc6__tar.gz

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arkindex-base-worker
-Version: 0.4.0rc5
+Version: 0.4.0rc6
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -41,19 +41,19 @@ Classifier: Programming Language :: Python :: 3.12
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: humanize==4.10.0
+Requires-Dist: humanize==4.11.0
 Requires-Dist: peewee~=3.17
-Requires-Dist: Pillow==10.4.0
-Requires-Dist: python-gnupg==0.5.2
+Requires-Dist: Pillow==11.0.0
+Requires-Dist: python-gnupg==0.5.3
 Requires-Dist: shapely==2.0.6
 Requires-Dist: teklia-toolbox==0.1.7b1
-Requires-Dist: zstandard==0.22.0
+Requires-Dist: zstandard==0.23.0
 Provides-Extra: docs
-Requires-Dist: black==24.4.2; extra == "docs"
-Requires-Dist: mkdocs-material==9.5.33; extra == "docs"
-Requires-Dist: mkdocstrings-python==1.11.1; extra == "docs"
+Requires-Dist: black==24.10.0; extra == "docs"
+Requires-Dist: mkdocs-material==9.5.48; extra == "docs"
+Requires-Dist: mkdocstrings-python==1.12.2; extra == "docs"
 Provides-Extra: tests
-Requires-Dist: pytest==8.3.2; extra == "tests"
+Requires-Dist: pytest==8.3.4; extra == "tests"
 Requires-Dist: pytest-mock==3.14.0; extra == "tests"
 Requires-Dist: pytest-responses==0.5.1; extra == "tests"
 

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/arkindex_base_worker.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: arkindex-base-worker
-Version: 0.4.0rc5
+Version: 0.4.0rc6
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -41,19 +41,19 @@ Classifier: Programming Language :: Python :: 3.12
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: humanize==4.10.0
+Requires-Dist: humanize==4.11.0
 Requires-Dist: peewee~=3.17
-Requires-Dist: Pillow==10.4.0
-Requires-Dist: python-gnupg==0.5.2
+Requires-Dist: Pillow==11.0.0
+Requires-Dist: python-gnupg==0.5.3
 Requires-Dist: shapely==2.0.6
 Requires-Dist: teklia-toolbox==0.1.7b1
-Requires-Dist: zstandard==0.22.0
+Requires-Dist: zstandard==0.23.0
 Provides-Extra: docs
-Requires-Dist: black==24.4.2; extra == "docs"
-Requires-Dist: mkdocs-material==9.5.33; extra == "docs"
-Requires-Dist: mkdocstrings-python==1.11.1; extra == "docs"
+Requires-Dist: black==24.10.0; extra == "docs"
+Requires-Dist: mkdocs-material==9.5.48; extra == "docs"
+Requires-Dist: mkdocstrings-python==1.12.2; extra == "docs"
 Provides-Extra: tests
-Requires-Dist: pytest==8.3.2; extra == "tests"
+Requires-Dist: pytest==8.3.4; extra == "tests"
 Requires-Dist: pytest-mock==3.14.0; extra == "tests"
 Requires-Dist: pytest-responses==0.5.1; extra == "tests"
 

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/arkindex_base_worker.egg-info/SOURCES.txt

@@ -49,6 +49,7 @@ tests/test_elements_worker/test_entity_create.py
 tests/test_elements_worker/test_entity_list_and_check.py
 tests/test_elements_worker/test_image.py
 tests/test_elements_worker/test_metadata.py
+tests/test_elements_worker/test_process.py
 tests/test_elements_worker/test_task.py
 tests/test_elements_worker/test_training.py
 tests/test_elements_worker/test_transcription_create.py

arkindex_base_worker-0.4.0rc6/arkindex_base_worker.egg-info/requires.txt

@@ -0,0 +1,17 @@
+humanize==4.11.0
+peewee~=3.17
+Pillow==11.0.0
+python-gnupg==0.5.3
+shapely==2.0.6
+teklia-toolbox==0.1.7b1
+zstandard==0.23.0
+
+[docs]
+black==24.10.0
+mkdocs-material==9.5.48
+mkdocstrings-python==1.12.2
+
+[tests]
+pytest==8.3.4
+pytest-mock==3.14.0
+pytest-responses==0.5.1

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/arkindex_worker/worker/__init__.py

@@ -27,7 +27,7 @@ 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.process import ActivityState, ProcessMode
+from arkindex_worker.worker.process import ActivityState, ProcessMixin, ProcessMode
 from arkindex_worker.worker.task import TaskMixin
 from arkindex_worker.worker.transcription import TranscriptionMixin
 from arkindex_worker.worker.version import WorkerVersionMixin
@@ -44,6 +44,7 @@ class ElementsWorker(
     EntityMixin,
     MetaDataMixin,
     ImageMixin,
+    ProcessMixin,
 ):
     """
     Base class for ML workers that operate on Arkindex elements.
@@ -119,6 +120,9 @@ class ElementsWorker(
             return list(
                 chain.from_iterable(map(self.list_set_elements, self.list_sets()))
             )
+        elif self.process_mode == ProcessMode.Export:
+            # For export mode processes, use list_process_elements and return element IDs
+            return {item["id"] for item in self.list_process_elements()}
 
         invalid_element_ids = list(filter(invalid_element_id, out))
         assert (

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/arkindex_worker/worker/process.py

@@ -1,5 +1,11 @@
+from collections.abc import Iterator
 from enum import Enum
 
+from arkindex_worker.cache import unsupported_cache
+
+# Increases the number of elements returned per page by the API
+PROCESS_ELEMENTS_PAGE_SIZE = 500
+
 
 class ActivityState(Enum):
     """
@@ -66,3 +72,21 @@ class ProcessMode(Enum):
     """
     Export processes.
     """
+
+
+class ProcessMixin:
+    @unsupported_cache
+    def list_process_elements(self, with_image: bool = False) -> Iterator[dict]:
+        """
+        List the elements of a process.
+
+        :param with_image: whether or not to include zone and image information in the elements response.
+        :returns: the process' elements.
+        """
+        return self.api_client.paginate(
+            "ListProcessElements",
+            id=self.process_information["id"],
+            with_image=with_image,
+            allow_missing_data=True,
+            page_size=PROCESS_ELEMENTS_PAGE_SIZE,
+        )

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/pyproject.toml

@@ -4,17 +4,17 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "arkindex-base-worker"
-version = "0.4.0rc5"
+version = "0.4.0rc6"
 description = "Base Worker to easily build Arkindex ML workflows"
 license = { file = "LICENSE" }
 dependencies = [
-    "humanize==4.10.0",
+    "humanize==4.11.0",
     "peewee~=3.17",
-    "Pillow==10.4.0",
-    "python-gnupg==0.5.2",
+    "Pillow==11.0.0",
+    "python-gnupg==0.5.3",
     "shapely==2.0.6",
     "teklia-toolbox==0.1.7b1",
-    "zstandard==0.22.0",
+    "zstandard==0.23.0",
 ]
 authors = [
     { name = "Teklia", email = "contact@teklia.com" },
@@ -44,12 +44,12 @@ Authors = "https://teklia.com"
 
 [project.optional-dependencies]
 docs = [
-    "black==24.4.2",
-    "mkdocs-material==9.5.33",
-    "mkdocstrings-python==1.11.1",
+    "black==24.10.0",
+    "mkdocs-material==9.5.48",
+    "mkdocstrings-python==1.12.2",
 ]
 tests = [
-    "pytest==8.3.2",
+    "pytest==8.3.4",
     "pytest-mock==3.14.0",
     "pytest-responses==0.5.1",
 ]

arkindex_base_worker-0.4.0rc6/tests/test_elements_worker/test_process.py

@@ -0,0 +1,89 @@
+import pytest
+
+from tests import PROCESS_ID
+
+
+@pytest.mark.parametrize(
+    ("with_image", "elements"),
+    [
+        (
+            False,
+            [
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 1",
+                    "confidence": 1,
+                    "image_id": None,
+                    "image_width": None,
+                    "image_height": None,
+                    "image_url": None,
+                    "polygon": None,
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaa0",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 2",
+                    "confidence": 1,
+                    "image_id": None,
+                    "image_width": None,
+                    "image_height": None,
+                    "image_url": None,
+                    "polygon": None,
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+            ],
+        ),
+        (
+            True,
+            [
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 1",
+                    "confidence": 1,
+                    "image_id": "aaa2aaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "image_width": 76,
+                    "image_height": 138,
+                    "image_url": "http://somewhere.com/iiif/image.jpeg",
+                    "polygon": [[0, 0], [0, 40], [20, 40], [20, 0]],
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaa0",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 2",
+                    "confidence": 1,
+                    "image_id": "aaa2aaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "image_width": 138,
+                    "image_height": 76,
+                    "image_url": "http://somewhere.com/iiif/image.jpeg",
+                    "polygon": [[0, 0], [0, 40], [20, 40], [20, 0]],
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+            ],
+        ),
+    ],
+)
+def test_list_process_elements_with_image(
+    responses, mock_elements_worker, with_image, elements
+):
+    responses.add(
+        responses.GET,
+        f"http://testserver/api/v1/process/{PROCESS_ID}/elements/?page_size=500&with_count=true&with_image={with_image}",
+        status=200,
+        json={
+            "count": 2,
+            "next": None,
+            "results": elements,
+        },
+    )
+    assert (
+        list(mock_elements_worker.list_process_elements(with_image=with_image))
+        == elements
+    )

{arkindex_base_worker-0.4.0rc5 → arkindex_base_worker-0.4.0rc6}/tests/test_elements_worker/test_worker.py

@@ -16,6 +16,7 @@ from arkindex_worker.models import Element
 from arkindex_worker.worker import ActivityState, ElementsWorker
 from arkindex_worker.worker.dataset import DatasetState
 from arkindex_worker.worker.process import ProcessMode
+from tests import PROCESS_ID
 
 from . import BASE_API_CALLS
 
@@ -523,6 +524,51 @@ def test_get_elements_both_args_error(mocker, mock_elements_worker, tmp_path):
         worker.get_elements()
 
 
+def test_get_elements_export_process(mock_elements_worker, responses):
+    responses.add(
+        responses.GET,
+        f"http://testserver/api/v1/process/{PROCESS_ID}/elements/?page_size=500&with_count=true&with_image=False",
+        status=200,
+        json={
+            "count": 2,
+            "next": None,
+            "results": [
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 1",
+                    "confidence": 1,
+                    "image_id": None,
+                    "image_width": None,
+                    "image_height": None,
+                    "image_url": None,
+                    "polygon": None,
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+                {
+                    "id": "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaa0",
+                    "type_id": "baaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+                    "name": "element 2",
+                    "confidence": 1,
+                    "image_id": None,
+                    "image_width": None,
+                    "image_height": None,
+                    "image_url": None,
+                    "polygon": None,
+                    "rotation_angle": 0,
+                    "mirrored": False,
+                },
+            ],
+        },
+    )
+    mock_elements_worker.process_information["mode"] = "export"
+    assert set(mock_elements_worker.get_elements()) == {
+        "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa",
+        "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaa0",
+    }
+
+
 @pytest.mark.usefixtures("_mock_worker_run_api")
 def test_activities_disabled(responses, monkeypatch):
     """Test worker process elements without updating activities when they are disabled for the process"""

arkindex_base_worker-0.4.0rc5/arkindex_base_worker.egg-info/requires.txt

@@ -1,17 +0,0 @@
-humanize==4.10.0
-peewee~=3.17
-Pillow==10.4.0
-python-gnupg==0.5.2
-shapely==2.0.6
-teklia-toolbox==0.1.7b1
-zstandard==0.22.0
-
-[docs]
-black==24.4.2
-mkdocs-material==9.5.33
-mkdocstrings-python==1.11.1
-
-[tests]
-pytest==8.3.2
-pytest-mock==3.14.0
-pytest-responses==0.5.1