arkindex-base-worker 0.5.1rc1__py3-none-any.whl → 0.5.1rc2__py3-none-any.whl

{arkindex_base_worker-0.5.1rc1.dist-info → arkindex_base_worker-0.5.1rc2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: arkindex-base-worker
-Version: 0.5.1rc1
+Version: 0.5.1rc2
 Summary: Base Worker to easily build Arkindex ML workflows
 Author-email: Teklia <contact@teklia.com>
 Maintainer-email: Teklia <contact@teklia.com>

{arkindex_base_worker-0.5.1rc1.dist-info → arkindex_base_worker-0.5.1rc2.dist-info}/RECORD

@@ -1,9 +1,9 @@
-arkindex_base_worker-0.5.1rc1.dist-info/licenses/LICENSE,sha256=NVshRi1efwVezMfW7xXYLrdDr2Li1AfwfGOd5WuH1kQ,1063
+arkindex_base_worker-0.5.1rc2.dist-info/licenses/LICENSE,sha256=NVshRi1efwVezMfW7xXYLrdDr2Li1AfwfGOd5WuH1kQ,1063
 arkindex_worker/__init__.py,sha256=Sdt5KXn8EgURb2MurYVrUWaHbH3iFA1XLRo0Lc5AJ44,250
 arkindex_worker/cache.py,sha256=XpEXMSnbhYCvrJquwA9XXqZo-ajMLpaCxKG5wH3Gp6Y,10959
 arkindex_worker/image.py,sha256=sGE8to5iykXv25bpkftOEWzlh5NzBZSKy4lSRoHYHPU,20929
 arkindex_worker/models.py,sha256=7GnKqpWPOSxyR_eKlDNVBe_r3TcE4ofK-1GzaonJEdM,10132
-arkindex_worker/utils.py,sha256=MbbJT8oh8DMHHR-vidFeXdUH0TSXGWm7ZDGWzrRXoEY,9933
+arkindex_worker/utils.py,sha256=yq_LmRlqfWOzB09Aiz2XYx4xPZnoEXR3As48h2HxOVc,10974
 arkindex_worker/worker/__init__.py,sha256=SzD0s1_m6gMV02EUF-NeciqZdVPA4dpXI84tSj-g494,17869
 arkindex_worker/worker/base.py,sha256=-ASOgLzemHOAz7EnPZ9hT6wON_9GJ7uadxx7oZLgUD4,22149
 arkindex_worker/worker/classification.py,sha256=qvykymkgd4nGywHCxL8obo4egstoGsmWNS4Ztc1qNWQ,11024
@@ -29,7 +29,7 @@ tests/test_element.py,sha256=hlj5VSF4plwC7uz9R4LGOOXZJQcHZiYCIDZT5V6EIB8,14334
 tests/test_image.py,sha256=yAM5mMfpQcIurT1KLHmu0AhSX2Qm3YvCu7afyZ3XUdU,28314
 tests/test_merge.py,sha256=REpZ13jkq_qm_4L5URQgFy5lxvPZtXxQEiWfYLMdmF0,7956
 tests/test_modern_config.py,sha256=Bm-a4LYQXgLZWQX7AmVyfJW0LNoLy1wj2d2GjzDkcBk,2683
-tests/test_utils.py,sha256=nYL1s2ViZoLoMiNpLGDaWwxf8dJ1D8aT522AO-PVaEQ,3607
+tests/test_utils.py,sha256=tgzNqyJMpddpeFWEjgsew_yDzmqnCA9HDaA5IpevAcM,5353
 tests/test_elements_worker/__init__.py,sha256=2t3NciCIOun_N-Wv63FWGsTm5W9N3mbwAWVuFORlMg8,308
 tests/test_elements_worker/test_classification.py,sha256=nya7veSPR_O9G41Enodp2-o6AifMBcaSTWJP2vXSSJ4,30133
 tests/test_elements_worker/test_cli.py,sha256=a23i1pUDbXi23MUtbWwGEcLLrmc_YlrbDgOG3h66wLM,2620
@@ -55,7 +55,7 @@ worker-demo/tests/conftest.py,sha256=XzNMNeg6pmABUAH8jN6eZTlZSFGLYjS3-DTXjiRN6Yc
 worker-demo/tests/test_worker.py,sha256=3DLd4NRK4bfyatG5P_PK4k9P9tJHx9XQq5_ryFEEFVg,304
 worker-demo/worker_demo/__init__.py,sha256=2BPomV8ZMNf3YXJgloatKeHQCE6QOkwmsHGkO6MkQuM,125
 worker-demo/worker_demo/worker.py,sha256=Rt-DjWa5iBP08k58NDZMfeyPuFbtNcbX6nc5jFX7GNo,440
-arkindex_base_worker-0.5.1rc1.dist-info/METADATA,sha256=Tj9sPHkLgV6wIDNTYut7GB-SmeEoH2_M6-__QNqo8iU,3138
-arkindex_base_worker-0.5.1rc1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-arkindex_base_worker-0.5.1rc1.dist-info/top_level.txt,sha256=-vNjP2VfROx0j83mdi9aIqRZ88eoJjxeWz-R_gPgyXU,49
-arkindex_base_worker-0.5.1rc1.dist-info/RECORD,,
+arkindex_base_worker-0.5.1rc2.dist-info/METADATA,sha256=Hur6AEbZLwqTw-hbggG-IOgDib14f0p6ujRGx6nD_jE,3138
+arkindex_base_worker-0.5.1rc2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+arkindex_base_worker-0.5.1rc2.dist-info/top_level.txt,sha256=-vNjP2VfROx0j83mdi9aIqRZ88eoJjxeWz-R_gPgyXU,49
+arkindex_base_worker-0.5.1rc2.dist-info/RECORD,,

arkindex_worker/utils.py

@@ -4,6 +4,7 @@ import logging
 import os
 import tarfile
 import tempfile
+import zipfile
 from collections.abc import Callable, Generator
 from itertools import islice
 from pathlib import Path
@@ -225,6 +226,32 @@ def create_tar_zst_archive(
     return zst_fd, zst_archive, zst_hash, tar_hash
 
 
+def create_zip_archive(source: Path, destination: Path | None = None) -> Path:
+    """Helper to create a ZIP archive from a source folder.
+
+    :param source: Path to the folder whose content should be archived.
+    :param destination: Path to the created archive, defaults to None. If unspecified, a temporary file will be created.
+    :return: The file descriptor of the created tempfile (if one was created), path to the archive.
+    """
+    # Parse destination and create a tmpfile if none was specified
+    file_d, destination = (
+        tempfile.mkstemp(prefix="teklia-", suffix=".zip")
+        if destination is None
+        else (None, destination)
+    )
+    destination = Path(destination)
+    logger.debug(f"Compressing file to {destination}")
+
+    with zipfile.ZipFile(
+        destination, mode="w", compression=zipfile.ZIP_BZIP2
+    ) as archive:
+        for p in source.rglob("*"):
+            relpath = p.relative_to(source)
+            archive.write(p, arcname=relpath)
+
+        return archive, destination
+
+
 DEFAULT_BATCH_SIZE = 50
 """Batch size used for bulk publication to Arkindex"""
 

tests/test_utils.py

@@ -1,4 +1,5 @@
 import logging
+import zipfile
 
 import pytest
 
@@ -7,6 +8,7 @@ from arkindex_worker.utils import (
     DEFAULT_BATCH_SIZE,
     batch_publication,
     close_delete_file,
+    create_zip_archive,
     extract_tar_zst_archive,
     parse_source_id,
 )
@@ -118,3 +120,43 @@ def test_batch_publication_decorator_alongside_unsupported_cache(caplog):
             "This API helper `custom_publication_in_batches_without_cache` did not update the cache database",
         ),
     ]
+
+
+def test_zip_archive():
+    # Create zip archive from fixtures
+    _, archive = create_zip_archive(FIXTURES_DIR / "extract_parent_archives/expected")
+
+    # Check the files in the archive
+    with zipfile.ZipFile(archive, mode="r") as f:
+        assert sorted(f.namelist()) == [
+            "test/",
+            "test/images/",
+            "test/images/f2649ce7-333e-44d2-ae73-387f18aad1f6.png",
+            "test/labels/",
+            "test/labels/f2649ce7-333e-44d2-ae73-387f18aad1f6.png",
+            "test/labels_json/",
+            "test/labels_json/f2649ce7-333e-44d2-ae73-387f18aad1f6.json",
+            "train/",
+            "train/images/",
+            "train/images/98115546-df07-448c-a2f0-34aa24789b77.png",
+            "train/images/ebeaa451-9287-4df7-9c40-07eb25cadb78.png",
+            "train/labels/",
+            "train/labels/98115546-df07-448c-a2f0-34aa24789b77.png",
+            "train/labels/ebeaa451-9287-4df7-9c40-07eb25cadb78.png",
+            "train/labels_json/",
+            "train/labels_json/98115546-df07-448c-a2f0-34aa24789b77.json",
+            "train/labels_json/ebeaa451-9287-4df7-9c40-07eb25cadb78.json",
+            "val/",
+            "val/images/",
+            "val/images/2987176d-4338-40f2-90d9-6d2cb4fd4a00.png",
+            "val/images/e3f91312-9201-45b7-9c32-e04a97ff1334.png",
+            "val/labels/",
+            "val/labels/2987176d-4338-40f2-90d9-6d2cb4fd4a00.png",
+            "val/labels/e3f91312-9201-45b7-9c32-e04a97ff1334.png",
+            "val/labels_json/",
+            "val/labels_json/2987176d-4338-40f2-90d9-6d2cb4fd4a00.json",
+            "val/labels_json/e3f91312-9201-45b7-9c32-e04a97ff1334.json",
+        ]
+
+    # Cleanup
+    archive.unlink()