datashare-python 0.11.0__py3-none-any.whl → 0.11.2__py3-none-any.whl

datashare_python/objects.py

@@ -304,11 +304,16 @@ class Document(DatashareModel):
             )
         resource_name = self.metadata.get(TIKA_METADATA_RESOURCENAME)
         if resource_name is None:
-            msg = (
-                f"doc {self} can't turn ES document into a {ProcessedFile.__name__} "
-                f"without tika metadata resources name in metadata: {self.metadata}"
-            )
-            raise ValueError(msg)
+            # Backward compat for project before resourcename was mandatory
+            if self.is_root_document:
+                resource_name = self.path.name
+            else:
+                msg = (
+                    f"doc {self} can't turn ES document into a"
+                    f" {ProcessedFile.__name__} without tika metadata resources name"
+                    f" in metadata: {self.metadata}"
+                )
+                raise ValueError(msg)
         resource_name = cast(str, resource_name)
         if self.is_root_document:
             path = self.path
@@ -361,7 +366,17 @@ class ManifestEntryStatus(StrEnum):
     PARTIAL = "partial"
 
 
+class User(DatashareModel):
+    id: str
+    name: str | None = None
+    email: str | None = None
+    provider: str | None = None
+    details: dict[str, Any] = Field(default_factory=dict)
+
+
 class TaskArgs(DatashareModel, ABC):
+    user: User | None = None
+
     def as_manifest_task_input(self) -> dict[str, Any]:
         # This is a base implementation, if the input is too large to be dumped,
         # override this and pop large keys

datashare_python/utils.py

@@ -512,13 +512,21 @@ def _write_artifact_bytes(path: Path, artifact: bytes | BytesIO | Path) -> None:
             with path.open("wb") as f:
                 f.write(artifact.read())
         case Path():
-            path.unlink(missing_ok=True)
+            _remove_previous_artifacts(path)
             shutil.move(artifact, path)
         case _:
             msg = f"unsupported artifact type: {artifact.__class__.__name__}"
             raise ValueError(msg)
 
 
+def _remove_previous_artifacts(path: Path) -> None:
+    if isinstance(path, Path) and path.exists():
+        with contextlib.suppress(FileNotFoundError):
+            path.rmdir()
+    else:
+        path.unlink(missing_ok=True)
+
+
 def debuggable_name(
     path: str, component_size_limit: int = 10, *, deterministic: bool = False
 ) -> str:

{datashare_python-0.11.0.dist-info → datashare_python-0.11.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: datashare-python
-Version: 0.11.0
+Version: 0.11.2
 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-0.11.0.dist-info → datashare_python-0.11.2.dist-info}/RECORD

@@ -10,19 +10,19 @@ datashare_python/exceptions.py,sha256=bVHEAXxDPKfxeeMC0hJXEsrJkgsKO2ESAhxWU96GA4
 datashare_python/interceptors.py,sha256=-XC-y_MvGyHlZQ6jItfMRPAmKwQkQzzncpAGxceKlwM,16485
 datashare_python/logging_.py,sha256=96KNXhP-JR5ybu9dld4VLtF7v7PfiphOCR1Nio-mmUs,4483
 datashare_python/mimetypes_.py,sha256=mjQgaqnSFZEcl3LoLDGBaBBjTTzwYfSNZND9Fmkk2zs,49682
-datashare_python/objects.py,sha256=_FQsZs2kdD2aqCdlov55lZq0-_W4H18NaH-6Y1N6eRg,18346
+datashare_python/objects.py,sha256=RuDEC4dZPndMYwFKhmMlToY8jjwjupG7mGGCG5tzc6g,18788
 datashare_python/task_client.py,sha256=oTmP8bvZW0UyhLNMi1AV3XIAx7hrdbxNRss2Mw2azEc,8435
 datashare_python/template.py,sha256=N1uM8TRXh5tQPV7BC1oJj-Pdm2Gs_SkjaXv50RnNYso,4682
 datashare_python/types_.py,sha256=cPcCkIk6ecILrKetPnAsrb0njHYuWJU3eZicUXy2lF8,1278
-datashare_python/utils.py,sha256=8OzlY9Y1QfXhhzERck4-Iizdmqt87SHNT0r19C0U7BA,28371
-datashare_python/worker-template.tar.gz,sha256=e8DonliLDGrC9Neh2icooEqpQDrEw4XOtHFbCDBli34,235943
+datashare_python/utils.py,sha256=k85QP34hSHBRFo8YwGA7rYPO5cqCt33i3ntl2oSfHkE,28603
+datashare_python/worker-template.tar.gz,sha256=3Lu6zLEXlndprxXPGMN9L19XrplmxUki3v15HymMmls,236583
 datashare_python/worker.py,sha256=ad36_yJOgwtj8KBheWndGkqaP_2aXTPSrYaHrbCXTvg,8389
 datashare_python/cli/__init__.py,sha256=9BPWtssDgsVfWMsZ1TtZCla0EC_kai4RHttr8oNLYOE,1401
 datashare_python/cli/project.py,sha256=w32Gy9AOL5B00uDT4in7YUCt2g68FnNbvwg2M3a8G6o,946
 datashare_python/cli/task.py,sha256=GsyR5o0fN4Pd8tJKAUgivAvCBKT3L4Ay5j_BCW1XBHk,5828
 datashare_python/cli/utils.py,sha256=p69CQb0zfixuyBkiZprhdMCc_NuYwXyAn6vC9H1UzAw,911
 datashare_python/cli/worker.py,sha256=x0p3a1_NsZPbmV7-WSVMRyW3_iadYcgyZKbDvWgymQc,5790
-datashare_python-0.11.0.dist-info/METADATA,sha256=Dl-HsxpC1OdtC_Ls9Ree246a_RwNYUbkSp7XnGAbNsw,1012
-datashare_python-0.11.0.dist-info/WHEEL,sha256=zOwg4jB6zX2kU910N-cMawjivD6tO8NEWvE12je1bVk,87
-datashare_python-0.11.0.dist-info/entry_points.txt,sha256=ILE7auxabHWiu3GC-AunWnzjhOI_SbZp7D4GqZHlLw4,68
-datashare_python-0.11.0.dist-info/RECORD,,
+datashare_python-0.11.2.dist-info/METADATA,sha256=STmFFjil2tcX6BgZOyEBdlraEtIs69qEWV6FGHGhGnY,1012
+datashare_python-0.11.2.dist-info/WHEEL,sha256=zOwg4jB6zX2kU910N-cMawjivD6tO8NEWvE12je1bVk,87
+datashare_python-0.11.2.dist-info/entry_points.txt,sha256=ILE7auxabHWiu3GC-AunWnzjhOI_SbZp7D4GqZHlLw4,68
+datashare_python-0.11.2.dist-info/RECORD,,