v4.2.12 Release

Author: khustup2

python/deeplake/__init__.py

@@ -16,7 +16,7 @@ def progress_bar(iterable, *args, **kwargs):
 from deeplake.ingestion import from_coco
 
 
-__version__ = "4.2.8"
+__version__ = "4.2.12"
 
 __all__ = [
     "__version__",
@@ -141,6 +141,7 @@ def progress_bar(iterable, *args, **kwargs):
     "core",
     "create",
     "create_async",
+    "_create_global_cache",
     "delete",
     "disconnect",
     "exists",

python/deeplake/__init__.pyi

@@ -129,6 +129,7 @@ __all__ = [
     "core",
     "create",
     "create_async",
+    "_create_global_cache",
     "delete",
     "disconnect",
     "exists",
@@ -2460,6 +2461,29 @@ class Dataset(DatasetView):
         When the dataset was created. The value is auto-generated at creation time.
         """
 
+    auto_commit_enabled: bool
+    """
+    This property controls whether the dataset will perform time-based auto-commits.
+
+    <!-- test-context
+    ```python
+    import deeplake
+    ds = deeplake.create("mem://auto_commit_ds")
+    ds.auto_commit_enabled = False
+    ds.add_column("column_name", deeplake.types.Text(deeplake.types.BM25))
+    a = ['a']*10_000
+    ds.append({"column_name":a})
+    ds.commit()
+    ```
+    -->
+
+    Examples:
+        ```python
+        ds = deeplake.open("mem://auto_commit_ds")
+        ds.auto_commit_enabled = True
+        ```
+    """
+
     indexing_mode: IndexingMode
     """
     The indexing mode of the dataset. This property can be set to change the indexing mode of the dataset for the current session,
@@ -2468,7 +2492,7 @@ class Dataset(DatasetView):
     <!-- test-context
     ```python
     import deeplake
-    ds = deeplake.create("mem://ds_id")
+    ds = deeplake.create("mem://indexing_mode_ds")
     ds.indexing_mode = deeplake.IndexingMode.Off
     ds.add_column("column_name", deeplake.types.Text(deeplake.types.BM25))
     a = ['a']*10_000
@@ -2479,7 +2503,7 @@ class Dataset(DatasetView):
 
     Examples:
         ```python
-        ds = deeplake.open("mem://ds_id")
+        ds = deeplake.open("mem://indexing_mode_ds")
         ds.indexing_mode = deeplake.IndexingMode.Automatic
         ds.commit()
         ```
@@ -3562,6 +3586,15 @@ def create_async(
         RuntimeError: if a dataset already exists at the given URL (will be raised when the future is awaited)
     """
 
+def _create_global_cache(
+    size: int = None,
+) -> None:
+    """
+    Creates a global cache with the given size.
+    Args:
+        size (int, optional): The size of the global cache in bytes. If not specified, a default size of 1GB is used.
+    """
+
 def copy(
     src: str,
     dst: str,

python/deeplake/integrations/__init__.py

@@ -1,6 +1,9 @@
 from .labelbox import (
     create_labelbox_annotation_project,
+    create_dataset_from_image_annotation_project,
+    create_dataset_from_image_annotation_project_with_custom_data_filler,
     create_dataset_from_video_annotation_project,
     converter_for_video_project_with_id,
+    converter_for_image_project_with_id,
     load_blob_file_paths_from_azure,
 )

python/deeplake/integrations/labelbox/__init__.py

@@ -2,7 +2,10 @@
     create_labelbox_annotation_project,
     create_dataset_from_video_annotation_project,
     create_dataset_from_video_annotation_project_with_custom_data_filler,
+    create_dataset_from_image_annotation_project,
+    create_dataset_from_image_annotation_project_with_custom_data_filler,
     converter_for_video_project_with_id,
+    converter_for_image_project_with_id,
 )
 from .labelbox_azure_utils import (
     load_blob_file_paths_from_azure,

python/deeplake/integrations/labelbox/converters.py

@@ -55,6 +55,50 @@ def interpolator(start, end, progress):
 
     converter.registered_interpolators[obj.feature_schema_id] = interpolator
 
+def polygon_converter_(obj, converter, tensor_name, context, generate_labels):
+    ds = context["ds"]
+    try:
+        ds.create_tensor(tensor_name, **polygon_tensor_create_kwargs_())
+    except:
+        pass
+
+    if generate_labels:
+        print("polygon converter does not support generating labels")
+
+    converter.register_feature_id_for_kind("tool", "polygon", obj, tensor_name)
+
+    def polygon_converter(row, obj):
+        if tensor_name not in converter.values_cache:
+            converter.values_cache[tensor_name] = dict()
+        if row not in converter.values_cache[tensor_name]:
+            converter.values_cache[tensor_name][row] = []
+        polygon = obj["polygon"]
+        if len(polygon) != 0 and not isinstance(polygon[0], dict):
+            # if polygon is a list of points, convert it to a list of dicts
+            polygon = [{"x": float(p[0]), "y": float(p[1])} for p in polygon]
+        converter.values_cache[tensor_name][row].append(
+            np.array([[float(p["x"]), float(p["y"])] for p in polygon])
+        )
+
+    converter.regsistered_actions[obj.feature_schema_id] = polygon_converter
+
+    def interpolator(start, end, progress):
+        start_polygon = start["polygon"]
+        end_polygon = end["polygon"]
+        polygon = copy.deepcopy(start)
+        polygon["polygon"] = [
+            [
+                start_polygon[i]["x"]
+                + (end_polygon[i]["x"] - start_polygon[i]["x"]) * progress,
+                start_polygon[i]["y"]
+                + (end_polygon[i]["y"] - start_polygon[i]["y"]) * progress,
+            ]
+            for i in range(len(start_polygon))
+        ]
+
+        return polygon
+
+    converter.registered_interpolators[obj.feature_schema_id] = interpolator
 
 def radio_converter_(obj, converter, tensor_name, context, generate_labels):
     ds = context["ds"]
@@ -199,9 +243,12 @@ def polygon_converter(row, obj):
             converter.values_cache[tensor_name] = dict()
         if row not in converter.values_cache[tensor_name]:
             converter.values_cache[tensor_name][row] = []
-
+        line = obj["line"]
+        if len(line) != 0 and not isinstance(line[0], dict):
+            # if line is a list of points, convert it to a list of dicts
+            line = [{"x": int(l[0]), "y": int(l[1])} for l in line]
         converter.values_cache[tensor_name][row].append(
-            [[int(l["x"]), int(l["y"])] for l in obj["line"]]
+            [[int(l["x"]), int(l["y"])] for l in line]
         )
 
     converter.regsistered_actions[obj.feature_schema_id] = polygon_converter

python/deeplake/integrations/labelbox/deeplake_utils.py

@@ -33,7 +33,7 @@ def class_label_tensor_create_kwargs_(dtype="int32"):
     }
 
 
-def image_tensor_create_kwargs_(sample_compression="jpg"):
+def image_tensor_create_kwargs_(sample_compression="png"):
     if is_v3():
         raise ValueError("unexpected deeplake version 3")
     return {"dtype": deeplake.types.Image(sample_compression=sample_compression)}