arkindex-base-worker 0.5.2a1__tar.gz → 0.5.2a2__tar.gz

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arkindex-base-worker
-Version: 0.5.2a1
+Version: 0.5.2a2
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -23,6 +23,7 @@ Requires-Dist: humanize==4.15.0
 Requires-Dist: peewee~=3.17
 Requires-Dist: Pillow==11.3.0
 Requires-Dist: python-gnupg==0.5.6
+Requires-Dist: python-magic==0.4.27
 Requires-Dist: shapely==2.0.6
 Requires-Dist: teklia-toolbox==0.1.12
 Requires-Dist: zstandard==0.25.0

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/arkindex_base_worker.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arkindex-base-worker
-Version: 0.5.2a1
+Version: 0.5.2a2
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>
@@ -23,6 +23,7 @@ Requires-Dist: humanize==4.15.0
 Requires-Dist: peewee~=3.17
 Requires-Dist: Pillow==11.3.0
 Requires-Dist: python-gnupg==0.5.6
+Requires-Dist: python-magic==0.4.27
 Requires-Dist: shapely==2.0.6
 Requires-Dist: teklia-toolbox==0.1.12
 Requires-Dist: zstandard==0.25.0

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/arkindex_base_worker.egg-info/requires.txt

@@ -2,6 +2,7 @@ humanize==4.15.0
 peewee~=3.17
 Pillow==11.3.0
 python-gnupg==0.5.6
+python-magic==0.4.27
 shapely==2.0.6
 teklia-toolbox==0.1.12
 zstandard==0.25.0

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/arkindex_worker/image.py

@@ -38,7 +38,7 @@ if TYPE_CHECKING:
     from arkindex_worker.models import Element
 
 # See http://docs.python-requests.org/en/master/user/advanced/#timeouts
-DOWNLOAD_TIMEOUT = (30, 60)
+REQUEST_TIMEOUT = (30, 60)
 
 BoundingBox = namedtuple("BoundingBox", ["x", "y", "width", "height"])
 
@@ -346,7 +346,7 @@ def _retried_request(url, *args, method=requests.get, **kwargs):
         url,
         *args,
         headers={"User-Agent": IIIF_USER_AGENT},
-        timeout=DOWNLOAD_TIMEOUT,
+        timeout=REQUEST_TIMEOUT,
         verify=should_verify_cert(url),
         **kwargs,
     )

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/arkindex_worker/worker/__init__.py

@@ -424,12 +424,13 @@ class DatasetWorker(DatasetMixin, BaseWorker, TaskMixin):
         failed = 0
         for i, dataset_set in enumerate(dataset_sets, start=1):
             try:
-                assert dataset_set.dataset.state == DatasetState.Complete.value, (
-                    "When processing a set, its dataset state should be Complete."
-                )
-
-                logger.info(f"Retrieving data for {dataset_set} ({i}/{count})")
-                self.download_dataset_artifact(dataset_set.dataset)
+                if dataset_set.dataset.state == DatasetState.Complete.value:
+                    logger.info(f"Retrieving data for {dataset_set} ({i}/{count})")
+                    self.download_dataset_artifact(dataset_set.dataset)
+                else:
+                    logger.warning(
+                        f"The dataset {dataset_set.dataset} has its state set to `{dataset_set.dataset.state}`, its archive will not be downloaded"
+                    )
 
                 logger.info(f"Processing {dataset_set} ({i}/{count})")
                 self.process_set(dataset_set)
@@ -444,7 +445,7 @@ class DatasetWorker(DatasetMixin, BaseWorker, TaskMixin):
 
                 logger.warning(message, exc_info=e if self.args.verbose else None)
 
-        # Cleanup the latest downloaded dataset artifact
+        # Cleanup the latest downloaded dataset artifact (if needed)
         self.cleanup_downloaded_artifact()
 
         message = f"Ran on {count} {pluralize('set', count)}: {count - failed} completed, {failed} failed"

arkindex_base_worker-0.5.2a2/arkindex_worker/worker/task.py

@@ -0,0 +1,100 @@
+"""
+BaseWorker methods for tasks.
+"""
+
+import uuid
+from collections.abc import Iterator
+from http.client import REQUEST_TIMEOUT
+from pathlib import Path
+
+import magic
+import requests
+
+from arkindex.compat import DownloadedFile
+from arkindex_worker import logger
+from arkindex_worker.models import Artifact
+from teklia_toolbox.requests import should_verify_cert
+
+
+class TaskMixin:
+    def list_artifacts(self, task_id: uuid.UUID) -> Iterator[Artifact]:
+        """
+        List artifacts associated to a task.
+
+        :param task_id: Task ID to find artifacts from.
+        :returns: An iterator of ``Artifact`` objects built from the ``ListArtifacts`` API endpoint.
+        """
+        assert task_id and isinstance(task_id, uuid.UUID), (
+            "task_id shouldn't be null and should be an UUID"
+        )
+
+        results = self.api_client.request("ListArtifacts", id=task_id)
+
+        return map(Artifact, results)
+
+    def download_artifact(
+        self, task_id: uuid.UUID, artifact: Artifact
+    ) -> DownloadedFile:
+        """
+        Download an artifact content.
+
+        :param task_id: Task ID the Artifact is from.
+        :param artifact: Artifact to download content from.
+        :returns: A temporary file containing the ``Artifact`` downloaded from the ``DownloadArtifact`` API endpoint.
+        """
+        assert task_id and isinstance(task_id, uuid.UUID), (
+            "task_id shouldn't be null and should be an UUID"
+        )
+        assert artifact and isinstance(artifact, Artifact), (
+            "artifact shouldn't be null and should be an Artifact"
+        )
+
+        return self.api_client.request(
+            "DownloadArtifact", id=task_id, path=artifact.path
+        )
+
+    def upload_artifact(self, path: Path) -> None:
+        """
+        Upload a single file as an Artifact of the current task.
+
+        :param path: Path of the single file to upload as an Artifact.
+        """
+        assert path and isinstance(path, Path) and path.exists(), (
+            "path shouldn't be null, should be a Path and should exist"
+        )
+
+        if self.is_read_only:
+            logger.warning("Cannot upload artifact as this worker is in read-only mode")
+            return
+
+        # Get path relative to task's data directory
+        relpath = str(path.relative_to(self.work_dir))
+
+        # Get file size
+        size = path.stat().st_size
+
+        # Detect content type
+        try:
+            content_type = magic.from_file(path, mime=True)
+        except Exception as e:
+            logger.warning(f"Failed to get a mime type for {path}: {e}")
+            content_type = "application/octet-stream"
+
+        # Create artifact on API to get an S3 url
+        artifact = self.api_client.request(
+            "CreateArtifact",
+            id=self.task_id,
+            body={"path": relpath, "content_type": content_type, "size": size},
+        )
+
+        # Upload the file content to S3
+        s3_put_url = artifact["s3_put_url"]
+        with path.open("rb") as content:
+            resp = requests.put(
+                s3_put_url,
+                data=content,
+                headers={"Content-Type": content_type},
+                timeout=REQUEST_TIMEOUT,
+                verify=should_verify_cert(s3_put_url),
+            )
+            resp.raise_for_status()

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "arkindex-base-worker"
-version = "0.5.2a1"
+version = "0.5.2a2"
 description = "Base Worker to easily build Arkindex ML workflows"
 license-files = ["LICENSE"]
 dependencies = [
@@ -12,6 +12,7 @@ dependencies = [
     "peewee~=3.17",
     "Pillow==11.3.0",
     "python-gnupg==0.5.6",
+    "python-magic==0.4.27",
     "shapely==2.0.6",
     "teklia-toolbox==0.1.12",
     "zstandard==0.25.0",

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/tests/test_dataset_worker.py

@@ -435,34 +435,6 @@ def test_run_no_sets(mocker, caplog, mock_dataset_worker):
     ]
 
 
-def test_run_initial_dataset_state_error(
-    mocker, responses, caplog, mock_dataset_worker, default_dataset
-):
-    default_dataset.state = DatasetState.Building.value
-    mocker.patch(
-        "arkindex_worker.worker.DatasetWorker.list_sets",
-        return_value=[Set(name="train", dataset=default_dataset)],
-    )
-
-    with pytest.raises(SystemExit):
-        mock_dataset_worker.run()
-
-    assert len(responses.calls) == len(BASE_API_CALLS) * 2
-    assert [
-        (call.request.method, call.request.url) for call in responses.calls
-    ] == BASE_API_CALLS * 2
-
-    assert [(level, message) for _, level, message in caplog.record_tuples] == [
-        (logging.INFO, "Loaded Worker Fake worker @ 123412 from API"),
-        (logging.INFO, "Modern configuration is not available"),
-        (
-            logging.WARNING,
-            "Failed running worker on Set (train) from Dataset (dataset_id): AssertionError('When processing a set, its dataset state should be Complete.')",
-        ),
-        (logging.ERROR, "Ran on 1 set: 0 completed, 1 failed"),
-    ]
-
-
 def test_run_download_dataset_artifact_api_error(
     mocker,
     tmp_path,
@@ -570,16 +542,18 @@ def test_run_no_downloaded_dataset_artifact_error(
     ]
 
 
+@pytest.mark.parametrize("dataset_state", DatasetState)
 def test_run(
     mocker,
     tmp_path,
     responses,
     caplog,
+    dataset_state,
     mock_dataset_worker,
     default_dataset,
     default_artifact,
 ):
-    default_dataset.state = DatasetState.Complete.value
+    default_dataset.state = dataset_state.value
     mocker.patch(
         "arkindex_worker.worker.DatasetWorker.list_sets",
         return_value=[Set(name="train", dataset=default_dataset)],
@@ -590,55 +564,68 @@ def test_run(
     )
     mock_process = mocker.patch("arkindex_worker.worker.DatasetWorker.process_set")
 
-    archive_path = (
-        FIXTURES_DIR
-        / "extract_parent_archives"
-        / "first_parent"
-        / "arkindex_data.tar.zst"
-    )
-    responses.add(
-        responses.GET,
-        f"http://testserver/api/v1/task/{default_dataset.task_id}/artifacts/",
-        status=200,
-        json=[default_artifact],
-    )
-    responses.add(
-        responses.GET,
-        f"http://testserver/api/v1/task/{default_dataset.task_id}/artifact/dataset_id.tar.zst",
-        status=200,
-        body=archive_path.read_bytes(),
-        content_type="application/zstd",
-    )
+    if dataset_state == DatasetState.Complete:
+        archive_path = (
+            FIXTURES_DIR
+            / "extract_parent_archives"
+            / "first_parent"
+            / "arkindex_data.tar.zst"
+        )
+        responses.add(
+            responses.GET,
+            f"http://testserver/api/v1/task/{default_dataset.task_id}/artifacts/",
+            status=200,
+            json=[default_artifact],
+        )
+        responses.add(
+            responses.GET,
+            f"http://testserver/api/v1/task/{default_dataset.task_id}/artifact/dataset_id.tar.zst",
+            status=200,
+            body=archive_path.read_bytes(),
+            content_type="application/zstd",
+        )
 
     mock_dataset_worker.run()
 
     assert mock_process.call_count == 1
 
-    assert len(responses.calls) == len(BASE_API_CALLS) * 2 + 2
+    # We only download the dataset archive when it is Complete
+    extra_calls = []
+    if dataset_state == DatasetState.Complete:
+        extra_calls = [
+            (
+                "GET",
+                f"http://testserver/api/v1/task/{default_dataset.task_id}/artifacts/",
+            ),
+            (
+                "GET",
+                f"http://testserver/api/v1/task/{default_dataset.task_id}/artifact/dataset_id.tar.zst",
+            ),
+        ]
+
+    assert len(responses.calls) == len(BASE_API_CALLS) * 2 + len(extra_calls)
     assert [
         (call.request.method, call.request.url) for call in responses.calls
-    ] == BASE_API_CALLS * 2 + [
-        (
-            "GET",
-            f"http://testserver/api/v1/task/{default_dataset.task_id}/artifacts/",
-        ),
-        (
-            "GET",
-            f"http://testserver/api/v1/task/{default_dataset.task_id}/artifact/dataset_id.tar.zst",
-        ),
-    ]
+    ] == BASE_API_CALLS * 2 + extra_calls
 
-    assert [(level, message) for _, level, message in caplog.record_tuples] == [
+    logs = [
         (logging.INFO, "Loaded Worker Fake worker @ 123412 from API"),
         (logging.INFO, "Modern configuration is not available"),
         (
-            logging.INFO,
-            "Retrieving data for Set (train) from Dataset (dataset_id) (1/1)",
+            logging.WARNING,
+            f"The dataset Dataset (dataset_id) has its state set to `{dataset_state.value}`, its archive will not be downloaded",
         ),
-        (logging.INFO, "Downloading artifact for Dataset (dataset_id)"),
         (logging.INFO, "Processing Set (train) from Dataset (dataset_id) (1/1)"),
         (logging.INFO, "Ran on 1 set: 1 completed, 0 failed"),
     ]
+    if dataset_state == DatasetState.Complete:
+        logs[2] = (
+            logging.INFO,
+            "Retrieving data for Set (train) from Dataset (dataset_id) (1/1)",
+        )
+        logs.insert(3, (logging.INFO, "Downloading artifact for Dataset (dataset_id)"))
+
+    assert [(level, message) for _, level, message in caplog.record_tuples] == logs
 
 
 def test_run_read_only(

{arkindex_base_worker-0.5.2a1 → arkindex_base_worker-0.5.2a2}/tests/test_elements_worker/test_task.py

@@ -1,6 +1,9 @@
+import tempfile
 import uuid
+from pathlib import Path
 
 import pytest
+from requests import HTTPError
 
 from arkindex.exceptions import ErrorResponse
 from arkindex_worker.models import Artifact
@@ -196,3 +199,112 @@ def test_download_artifact(
     ] == BASE_API_CALLS + [
         ("GET", f"http://testserver/api/v1/task/{TASK_ID}/artifact/dataset_id.tar.zst"),
     ]
+
+
+@pytest.mark.parametrize(
+    ("payload", "error"),
+    [
+        # Path
+        (
+            {"path": None},
+            "path shouldn't be null, should be a Path and should exist",
+        ),
+        (
+            {"path": "not path type"},
+            "path shouldn't be null, should be a Path and should exist",
+        ),
+        (
+            {"path": Path("i_do_no_exist.oops")},
+            "path shouldn't be null, should be a Path and should exist",
+        ),
+    ],
+)
+def test_upload_artifact_wrong_param_path(mock_dataset_worker, payload, error):
+    with pytest.raises(AssertionError, match=error):
+        mock_dataset_worker.upload_artifact(**payload)
+
+
+@pytest.fixture
+def tmp_file(mock_dataset_worker):
+    with tempfile.NamedTemporaryFile(
+        mode="w", suffix=".txt", dir=mock_dataset_worker.work_dir
+    ) as file:
+        file.write("Some content...")
+        file.seek(0)
+
+        yield Path(file.name)
+
+
+def test_upload_artifact_api_error(responses, mock_dataset_worker, tmp_file):
+    responses.add(
+        responses.POST,
+        "http://testserver/api/v1/task/my_task/artifacts/",
+        status=418,
+    )
+
+    with pytest.raises(ErrorResponse):
+        mock_dataset_worker.upload_artifact(path=tmp_file)
+
+    assert len(responses.calls) == len(BASE_API_CALLS) + 1
+    assert [
+        (call.request.method, call.request.url) for call in responses.calls
+    ] == BASE_API_CALLS + [("POST", "http://testserver/api/v1/task/my_task/artifacts/")]
+
+
+def test_upload_artifact_s3_upload_error(
+    responses,
+    mock_dataset_worker,
+    tmp_file,
+):
+    responses.add(
+        responses.POST,
+        "http://testserver/api/v1/task/my_task/artifacts/",
+        json={
+            "id": "11111111-1111-1111-1111-111111111111",
+            "path": tmp_file.name,
+            "size": 15,
+            "content_type": "text/plain",
+            "s3_put_url": "http://example.com/oops.txt",
+        },
+    )
+    responses.add(responses.PUT, "http://example.com/oops.txt", status=500)
+
+    with pytest.raises(HTTPError):
+        mock_dataset_worker.upload_artifact(path=tmp_file)
+
+    assert len(responses.calls) == len(BASE_API_CALLS) + 2
+    assert [
+        (call.request.method, call.request.url) for call in responses.calls
+    ] == BASE_API_CALLS + [
+        ("POST", "http://testserver/api/v1/task/my_task/artifacts/"),
+        ("PUT", "http://example.com/oops.txt"),
+    ]
+
+
+def test_upload_artifact(
+    responses,
+    mock_dataset_worker,
+    tmp_file,
+):
+    responses.add(
+        responses.POST,
+        "http://testserver/api/v1/task/my_task/artifacts/",
+        json={
+            "id": "11111111-1111-1111-1111-111111111111",
+            "path": tmp_file.name,
+            "size": 15,
+            "content_type": "text/plain",
+            "s3_put_url": "http://example.com/test.txt",
+        },
+    )
+    responses.add(responses.PUT, "http://example.com/test.txt")
+
+    mock_dataset_worker.upload_artifact(path=tmp_file)
+
+    assert len(responses.calls) == len(BASE_API_CALLS) + 2
+    assert [
+        (call.request.method, call.request.url) for call in responses.calls
+    ] == BASE_API_CALLS + [
+        ("POST", "http://testserver/api/v1/task/my_task/artifacts/"),
+        ("PUT", "http://example.com/test.txt"),
+    ]

arkindex_base_worker-0.5.2a1/arkindex_worker/worker/task.py

@@ -1,47 +0,0 @@
-"""
-BaseWorker methods for tasks.
-"""
-
-import uuid
-from collections.abc import Iterator
-
-from arkindex.compat import DownloadedFile
-from arkindex_worker.models import Artifact
-
-
-class TaskMixin:
-    def list_artifacts(self, task_id: uuid.UUID) -> Iterator[Artifact]:
-        """
-        List artifacts associated to a task.
-
-        :param task_id: Task ID to find artifacts from.
-        :returns: An iterator of ``Artifact`` objects built from the ``ListArtifacts`` API endpoint.
-        """
-        assert task_id and isinstance(task_id, uuid.UUID), (
-            "task_id shouldn't be null and should be an UUID"
-        )
-
-        results = self.api_client.request("ListArtifacts", id=task_id)
-
-        return map(Artifact, results)
-
-    def download_artifact(
-        self, task_id: uuid.UUID, artifact: Artifact
-    ) -> DownloadedFile:
-        """
-        Download an artifact content.
-
-        :param task_id: Task ID the Artifact is from.
-        :param artifact: Artifact to download content from.
-        :returns: A temporary file containing the ``Artifact`` downloaded from the ``DownloadArtifact`` API endpoint.
-        """
-        assert task_id and isinstance(task_id, uuid.UUID), (
-            "task_id shouldn't be null and should be an UUID"
-        )
-        assert artifact and isinstance(artifact, Artifact), (
-            "artifact shouldn't be null and should be an Artifact"
-        )
-
-        return self.api_client.request(
-            "DownloadArtifact", id=task_id, path=artifact.path
-        )