datashare-python 0.12.9__tar.gz → 0.12.11__tar.gz

{datashare_python-0.12.9 → datashare_python-0.12.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: datashare-python
-Version: 0.12.9
+Version: 0.12.11
 Summary: Manage Python tasks and local resources in Datashare
 Project-URL: Homepage, https://icij.github.io/datashare-python/
 Project-URL: Documentation, https://icij.github.io/datashare-python/

{datashare_python-0.12.9 → datashare_python-0.12.11}/datashare_python/conftest.py

@@ -48,6 +48,7 @@ _DEFAULT_BROKER_URL = (
 _DEFAULT_AUTH = aiohttp.BasicAuth(login="guest", password="guest", encoding="utf-8")
 
 TEST_PROJECT = "test-project"
+DS_URL = "https://datashare.org"
 
 _INDEX_BODY = {
     "mappings": {

{datashare_python-0.12.9 → datashare_python-0.12.11}/datashare_python/exceptions.py

@@ -1,5 +1,7 @@
 from abc import ABC
 
+from .objects import DocArtifact
+
 
 class DatashareError(Exception, ABC): ...
 
@@ -16,3 +18,9 @@ class DependencyInjectionError(DatashareError, RuntimeError):
     def __init__(self, name: str):
         msg = f"{name} was not injected"
         super().__init__(msg)
+
+
+class ArtifactNotFound(DatashareError):
+    def __init__(self, doc_id: str, artifact_type: type[DocArtifact]):
+        msg = f"could not find {artifact_type.__name__} for {doc_id}"
+        super().__init__(msg)

{datashare_python-0.12.9 → datashare_python-0.12.11}/datashare_python/objects.py

@@ -858,10 +858,12 @@ class WorkerFile(
         )
 
 
-def _file_discriminator(v: Any) -> str:
+def _file_discriminator(v: Any) -> str | None:
     parent = v.get("parent") if isinstance(v, dict) else getattr(v, "parent", None)
     if parent is None:
-        raise ValueError(f"{v} has no parent")
+        # No match the parent is neither a DatashareFile of WorkerFile it could be a
+        # Document or we're just deserializing something else
+        return None
     if isinstance(parent, dict):
         is_document = parent.get("type") == "Document"
     else:

{datashare_python-0.12.9 → datashare_python-0.12.11}/datashare_python/utils.py

@@ -54,6 +54,7 @@ from datashare_python.types_ import (
 )
 
 from .constants import MANIFEST_JSON, METADATA_JSON
+from .exceptions import ArtifactNotFound
 from .mimetypes_ import types_map
 from .objects import (
     BaseModel,
@@ -401,6 +402,37 @@ def artifacts_dir(doc_id: str, *, project: str) -> Path:
     return Path(project, safe_dir(doc_id), doc_id)
 
 
+async def async_read_artifact(
+    paths: WorkerPaths, doc_id: str, artifact_type: type[DocArtifact], *, project: str
+) -> bytes:
+    # Reading the manifest can be long, we load it only if the artifact exists then
+    # we check that the entry is in the manifest
+    artifact_path = (
+        paths.artifacts
+        / artifacts_dir(doc_id, project=project)
+        / artifact_type.filename
+    )
+    if not artifact_path.exists():
+        raise ArtifactNotFound(doc_id, artifact_type)
+    m_path = paths.artifacts / _manifest_path(doc_id, project=project)
+
+    try:
+        async with async_open(m_path, "r") as f:
+            manifest_json = await f.read()
+    except OSError as e:
+        raise ArtifactNotFound(doc_id, artifact_type) from e
+
+    manifest = json.loads(manifest_json)
+    if artifact_type.type.value not in manifest:
+        raise ArtifactNotFound(doc_id, artifact_type)
+
+    try:
+        async with async_open(artifact_path, "rb") as f:
+            return await f.read()
+    except OSError as e:
+        raise ArtifactNotFound(doc_id, artifact_type) from e
+
+
 def _metadata_path(doc_id: str, *, project: str) -> Path:
     metadata_path = artifacts_dir(doc_id, project=project) / METADATA_JSON
     return metadata_path

{datashare_python-0.12.9 → datashare_python-0.12.11}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "datashare-python"
-version = "0.12.9"
+version = "0.12.11"
 description = "Manage Python tasks and local resources in Datashare"
 authors = [
   { name = "Clément Doumouro", email = "cdoumouro@icij.org" },