unitlab 2.4.2__tar.gz → 2.4.4__tar.gz

{unitlab-2.4.2 → unitlab-2.4.4}/.claude/settings.local.json

@@ -41,7 +41,9 @@
       "Bash(ps:*)",
       "WebFetch(domain:127.0.0.1)",
       "Bash(DATASET_ID=\"b727a6fd-7786-4d40-b75b-56e7f52f378b\")",
-      "Bash(pipenv run:*)"
+      "Bash(pipenv run:*)",
+      "Bash(DS=\"b727a6fd-7786-4d40-b75b-56e7f52f378b\")",
+      "Bash(python3 -c \"import sys,json; d=json.load\\(sys.stdin\\); print\\(''''OK'''' if ''''file'''' in d else d\\)\")"
     ]
   }
 }

{unitlab-2.4.2 → unitlab-2.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: unitlab
-Version: 2.4.2
+Version: 2.4.4
 Summary: Python SDK for the Unitlab.ai data annotation platform
 Project-URL: Homepage, https://unitlab.ai
 Project-URL: Documentation, https://docs.unitlab.ai

{unitlab-2.4.2 → unitlab-2.4.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "unitlab"
-version = "2.4.2"
+version = "2.4.4"
 description = "Python SDK for the Unitlab.ai data annotation platform"
 readme = "README.md"
 license = "MIT"

{unitlab-2.4.2 → unitlab-2.4.4}/src/unitlab/cli.py

@@ -37,12 +37,6 @@ class DownloadType(str, Enum):
     files = "files"
 
 
-class DatasetSplit(str, Enum):
-    train = "train"
-    validation = "validation"
-    test = "test"
-
-
 @app.command(help="Configure the credentials")
 def configure(
     api_key: Annotated[
@@ -124,11 +118,14 @@ def dataset_download(
         typer.Option(help="Export type (e.g. COCO, YOLOv8, YOLOv5, UUEF)"),
     ] = "COCO",
     split_type: Annotated[
-        DatasetSplit, typer.Option(help="Dataset split (train, validation, test)")
-    ] = DatasetSplit.train,
+        str | None,
+        typer.Option(
+            help="Dataset split (e.g. train, validation). Auto-detected if omitted."
+        ),
+    ] = None,
 ):
     if download_type == DownloadType.annotation:
-        get_client(api_key).dataset_download(str(pk), export_type, split_type.value)
+        get_client(api_key).dataset_download(str(pk), export_type, split_type)
     else:
         typer.echo(get_client(api_key).dataset_download_files(str(pk)))
 

{unitlab-2.4.2 → unitlab-2.4.4}/src/unitlab/client.py

@@ -4,6 +4,7 @@ import asyncio
 import functools
 import logging
 import os
+import uuid
 from pathlib import Path
 from typing import Any
 from urllib.parse import urlparse
@@ -247,6 +248,9 @@ class UnitlabClient:
         num_batches = (num_files + batch_size - 1) // batch_size
         semaphore = asyncio.Semaphore(_UPLOAD_CONCURRENCY)
 
+        # Medical projects use session-based upload with finalize
+        session_id = str(uuid.uuid4()) if generic_type == "medical" else None
+
         async def post_file(
             client: httpx.AsyncClient, file: str, project_id: str
         ) -> int:
@@ -256,6 +260,8 @@ class UnitlabClient:
                     extra_data["sentences_per_chunk"] = str(sentences_per_chunk)
                 elif generic_type == "video":
                     extra_data["fps"] = str(fps)
+                if session_id:
+                    extra_data["session_id"] = session_id
 
                 try:
                     with open(file, "rb") as f:
@@ -288,14 +294,55 @@ class UnitlabClient:
                         for f in asyncio.as_completed(tasks):
                             pbar.update(await f)
 
+                    # Finalize medical session after all uploads
+                    if session_id:
+                        response = await client.post(
+                            f"/api/sdk/projects/{project_id}/"
+                            f"medical-upload-sessions/{session_id}/finalize/",
+                        )
+                        response.raise_for_status()
+                        logger.info("Medical session finalized")
+
         asyncio.run(main())
 
     def datasets(self, pretty: int = 0) -> Any:
         return self._get(f"/api/sdk/datasets/?pretty={pretty}")
 
+    def dataset_info(self, dataset_id: str) -> Any:
+        """Return metadata for a dataset.
+
+        Includes name, version, annotation_type, number_of_data,
+        download_formats and the list of available ``splits``.
+        """
+        return self._get(f"/api/sdk/datasets/{dataset_id}/")
+
     def dataset_download(
-        self, dataset_id: str, export_type: str, split_type: str
+        self,
+        dataset_id: str,
+        export_type: str,
+        split_type: str | None = None,
     ) -> str:
+        if split_type is None:
+            info = self.dataset_info(dataset_id)
+            splits = info.get("splits") or []
+            if not splits:
+                raise NetworkError(
+                    message=(
+                        "No splits available for this dataset. Nothing to download."
+                    ),
+                    detail=None,
+                )
+            if len(splits) > 1:
+                raise NetworkError(
+                    message=(
+                        f"Multiple splits available: {', '.join(splits)}. "
+                        f"Specify one with split_type=..."
+                    ),
+                    detail=None,
+                )
+            split_type = splits[0]
+            logger.info(f"Auto-selected split: {split_type}")
+
         response = self._post(
             f"/api/sdk/datasets/{dataset_id}/",
             data={