hafnia 0.3.0__tar.gz → 0.4.1__tar.gz

{hafnia-0.3.0 → hafnia-0.4.1}/.github/workflows/build.yaml

@@ -29,7 +29,7 @@ jobs:
           echo "package_version=$VERSION" >> $GITHUB_OUTPUT
 
       - name: Install uv
-        uses: astral-sh/setup-uv@v6
+        uses: astral-sh/setup-uv@v7
         with:
           version: 0.6.8
 
@@ -45,7 +45,7 @@ jobs:
         run: uv build
 
       - name: Upload package artifact
-        uses: actions/upload-artifact@v4.6.2
+        uses: actions/upload-artifact@v5.0.0
         with:
           name: python-package
           path: dist/

{hafnia-0.3.0 → hafnia-0.4.1}/.github/workflows/check_release.yaml

@@ -20,7 +20,7 @@ jobs:
       make_release: ${{ steps.check_release.outputs.make_release }}
     steps:
       - name: Download package artifact
-        uses: actions/download-artifact@v5.0.0
+        uses: actions/download-artifact@v6.0.0
         with:
             name: python-package
             path: dist/

{hafnia-0.3.0 → hafnia-0.4.1}/.github/workflows/publish_docker.yaml

@@ -31,7 +31,7 @@ jobs:
           python-version-file: ${{ inputs.python-version-file }}
 
       - name: Download package artifact
-        uses: actions/download-artifact@v5.0.0
+        uses: actions/download-artifact@v6.0.0
         with:
           name: python-package
           path: dist/
@@ -47,7 +47,7 @@ jobs:
             echo "aws_region=${{ secrets.STAGE_AWS_REGION }}" >> $GITHUB_OUTPUT
           fi
       - name: Configure AWS credentials
-        uses: aws-actions/configure-aws-credentials@v5.0.0
+        uses: aws-actions/configure-aws-credentials@v5.1.0
         with:
           role-to-assume: arn:aws:iam::${{ steps.env-vars.outputs.aws_account_id }}:role/${{ secrets.AWS_ROLE_NAME }}
           aws-region: ${{ steps.env-vars.outputs.aws_region }} 

{hafnia-0.3.0 → hafnia-0.4.1}/.github/workflows/publish_pypi.yaml

@@ -17,7 +17,7 @@ jobs:
       contents: read
     steps:
       - name: Download package artifact
-        uses: actions/download-artifact@v5.0.0
+        uses: actions/download-artifact@v6.0.0
         with:
           name: python-package
           path: dist/

{hafnia-0.3.0 → hafnia-0.4.1}/.github/workflows/tests.yaml

@@ -11,6 +11,7 @@ jobs:
   test:
     runs-on: ${{ matrix.os }}
     strategy:
+      max-parallel: 1
       matrix:
         os: [ubuntu-latest, windows-latest]
     steps:
@@ -19,7 +20,7 @@ jobs:
         with:
           python-version-file: ${{ inputs.python-version-file }}
       - name: Install uv
-        uses: astral-sh/setup-uv@v6
+        uses: astral-sh/setup-uv@v7
         with:
           version: 0.6.8
       - name: Install the project

hafnia-0.4.1/.trivyignore

@@ -0,0 +1,3 @@
+# Ignore 'CVE-2024-37059' issue https://avd.aquasec.com/nvd/2024/cve-2024-37059/
+# The vulnerability does not apply to our platform as models are not loaded on our platform.
+CVE-2024-37059

{hafnia-0.3.0 → hafnia-0.4.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hafnia
-Version: 0.3.0
+Version: 0.4.1
 Summary: Python SDK for communication with Hafnia platform.
 Author-email: Milestone Systems <hafniaplatform@milestone.dk>
 License-File: LICENSE
@@ -9,7 +9,9 @@ Requires-Dist: boto3>=1.35.91
 Requires-Dist: click>=8.1.8
 Requires-Dist: emoji>=2.14.1
 Requires-Dist: flatten-dict>=0.4.2
-Requires-Dist: mlflow>=3.2.0
+Requires-Dist: keyring>=25.6.0
+Requires-Dist: mcp==1.16.0
+Requires-Dist: mlflow>=3.4.0
 Requires-Dist: more-itertools>=10.7.0
 Requires-Dist: opencv-python-headless>=4.11.0.86
 Requires-Dist: pathspec>=0.12.1
@@ -22,7 +24,6 @@ Requires-Dist: rich>=13.9.4
 Requires-Dist: s5cmd>=0.2.0
 Requires-Dist: sagemaker-mlflow>=0.1.0
 Requires-Dist: seedir>=0.5.0
-Requires-Dist: tqdm>=4.67.1
 Requires-Dist: xxhash>=3.5.0
 Description-Content-Type: text/markdown
 
@@ -80,6 +81,7 @@ Copy the key and save it for later use.
     Hafnia API Key:  # Pass your HAFNIA API key
     Hafnia Platform URL [https://api.mdi.milestonesys.com]:  # Press [Enter]
     ```
+
 1. Download `mnist` from terminal to verify that your configuration is working.  
 
     ```bash
@@ -91,7 +93,7 @@ With Hafnia configured on your local machine, it is now possible to download
 and explore the dataset sample with a python script:
 
 ```python
-from hafnia.data import load_dataset, get_dataset_path
+from hafnia.data import get_dataset_path
 from hafnia.dataset.hafnia_dataset import HafniaDataset
 
 # To download the sample dataset use:
@@ -156,7 +158,7 @@ and `dataset.samples` with annotations as a polars DataFrame
 print(dataset.samples.head(2))
 shape: (2, 14)
 ┌──────────────┬─────────────────────────────────┬────────┬───────┬───┬─────────────────────────────────┬──────────┬──────────┬─────────────────────────────────┐
-│ sample_index ┆ file_name                       ┆ height ┆ width ┆ … ┆ objects                         ┆ bitmasks ┆ polygons ┆ meta                            │
+│ sample_index ┆ file_name                       ┆ height ┆ width ┆ … ┆ bboxes                          ┆ bitmasks ┆ polygons ┆ meta                            │
 │ ---          ┆ ---                             ┆ ---    ┆ ---   ┆   ┆ ---                             ┆ ---      ┆ ---      ┆ ---                             │
 │ u32          ┆ str                             ┆ i64    ┆ i64   ┆   ┆ list[struct[11]]                ┆ null     ┆ null     ┆ struct[5]                       │
 ╞══════════════╪═════════════════════════════════╪════════╪═══════╪═══╪═════════════════════════════════╪══════════╪══════════╪═════════════════════════════════╡
@@ -216,7 +218,7 @@ sample_dict = dataset[0]
 
 for sample_dict in dataset:
     sample = Sample(**sample_dict)
-    print(sample.sample_id, sample.objects)
+    print(sample.sample_id, sample.bboxes)
     break
 ```
 Not that it is possible to create a `Sample` object from the sample dictionary.
@@ -360,7 +362,7 @@ logger.log_scalar("validation/loss", value=0.1, step=100)
 logger.log_metric("validation/accuracy", value=0.95, step=100)
 ```
 
-Similar to `load_dataset`, the tracker behaves differently when running locally or in the cloud. 
+The tracker behaves differently when running locally or in the cloud. 
 Locally, experiment data is stored in a local folder `.data/experiments/{DATE_TIME}`. 
 
 In the cloud, the experiment data will be available in the Hafnia platform under 
@@ -384,7 +386,7 @@ and datasets available in the data library.
 
 ```python
 # Load Hugging Face dataset
-dataset_splits = load_dataset("midwest-vehicle-detection")
+dataset_splits = HafniaDataset.from_name("midwest-vehicle-detection")
 
 # Define transforms
 train_transforms = v2.Compose(
@@ -419,7 +421,7 @@ pil_image.save("visualized_labels.png")
 
 # Create DataLoaders - using TorchVisionCollateFn
 collate_fn = torch_helpers.TorchVisionCollateFn(
-    skip_stacking=["objects.bbox", "objects.class_idx"]
+    skip_stacking=["bboxes.bbox", "bboxes.class_idx"]
 )
 train_loader = DataLoader(train_dataset, batch_size=8, shuffle=True, collate_fn=collate_fn)
 ```

{hafnia-0.3.0 → hafnia-0.4.1}/README.md

@@ -52,6 +52,7 @@ Copy the key and save it for later use.
     Hafnia API Key:  # Pass your HAFNIA API key
     Hafnia Platform URL [https://api.mdi.milestonesys.com]:  # Press [Enter]
     ```
+
 1. Download `mnist` from terminal to verify that your configuration is working.  
 
     ```bash
@@ -63,7 +64,7 @@ With Hafnia configured on your local machine, it is now possible to download
 and explore the dataset sample with a python script:
 
 ```python
-from hafnia.data import load_dataset, get_dataset_path
+from hafnia.data import get_dataset_path
 from hafnia.dataset.hafnia_dataset import HafniaDataset
 
 # To download the sample dataset use:
@@ -128,7 +129,7 @@ and `dataset.samples` with annotations as a polars DataFrame
 print(dataset.samples.head(2))
 shape: (2, 14)
 ┌──────────────┬─────────────────────────────────┬────────┬───────┬───┬─────────────────────────────────┬──────────┬──────────┬─────────────────────────────────┐
-│ sample_index ┆ file_name                       ┆ height ┆ width ┆ … ┆ objects                         ┆ bitmasks ┆ polygons ┆ meta                            │
+│ sample_index ┆ file_name                       ┆ height ┆ width ┆ … ┆ bboxes                          ┆ bitmasks ┆ polygons ┆ meta                            │
 │ ---          ┆ ---                             ┆ ---    ┆ ---   ┆   ┆ ---                             ┆ ---      ┆ ---      ┆ ---                             │
 │ u32          ┆ str                             ┆ i64    ┆ i64   ┆   ┆ list[struct[11]]                ┆ null     ┆ null     ┆ struct[5]                       │
 ╞══════════════╪═════════════════════════════════╪════════╪═══════╪═══╪═════════════════════════════════╪══════════╪══════════╪═════════════════════════════════╡
@@ -188,7 +189,7 @@ sample_dict = dataset[0]
 
 for sample_dict in dataset:
     sample = Sample(**sample_dict)
-    print(sample.sample_id, sample.objects)
+    print(sample.sample_id, sample.bboxes)
     break
 ```
 Not that it is possible to create a `Sample` object from the sample dictionary.
@@ -332,7 +333,7 @@ logger.log_scalar("validation/loss", value=0.1, step=100)
 logger.log_metric("validation/accuracy", value=0.95, step=100)
 ```
 
-Similar to `load_dataset`, the tracker behaves differently when running locally or in the cloud. 
+The tracker behaves differently when running locally or in the cloud. 
 Locally, experiment data is stored in a local folder `.data/experiments/{DATE_TIME}`. 
 
 In the cloud, the experiment data will be available in the Hafnia platform under 
@@ -356,7 +357,7 @@ and datasets available in the data library.
 
 ```python
 # Load Hugging Face dataset
-dataset_splits = load_dataset("midwest-vehicle-detection")
+dataset_splits = HafniaDataset.from_name("midwest-vehicle-detection")
 
 # Define transforms
 train_transforms = v2.Compose(
@@ -391,7 +392,7 @@ pil_image.save("visualized_labels.png")
 
 # Create DataLoaders - using TorchVisionCollateFn
 collate_fn = torch_helpers.TorchVisionCollateFn(
-    skip_stacking=["objects.bbox", "objects.class_idx"]
+    skip_stacking=["bboxes.bbox", "bboxes.class_idx"]
 )
 train_loader = DataLoader(train_dataset, batch_size=8, shuffle=True, collate_fn=collate_fn)
 ```

{hafnia-0.3.0 → hafnia-0.4.1}/examples/example_dataset_recipe.py

@@ -3,7 +3,7 @@ from pathlib import Path
 from rich import print as rprint
 
 from hafnia import utils
-from hafnia.data.factory import load_dataset
+from hafnia.dataset.dataset_names import OPS_REMOVE_CLASS
 from hafnia.dataset.dataset_recipe.dataset_recipe import DatasetRecipe
 from hafnia.dataset.dataset_recipe.recipe_transforms import (
     SelectSamples,
@@ -50,7 +50,7 @@ dataset_recipe.as_python_code()
 # executed in the TaaS platform. This is demonstrated below:
 if utils.is_hafnia_configured():  # First ensure you are connected to the hafnia platform
     # Upload the dataset recipe - this will make it available for TaaS and for users of your organization
-    dataset_recipe.as_platform_recipe(recipe_name="example-mnist-recipe")
+    dataset_recipe.as_platform_recipe(recipe_name="example-mnist-recipe", overwrite=True)
 
     # The recipe is now available in TaaS, for different environments and other users in your organization
     dataset_recipe_again = DatasetRecipe.from_recipe_name(name="example-mnist-recipe")
@@ -95,53 +95,128 @@ rprint(dataset_recipe)  # as a python object
 print(dataset_recipe.as_json_str())  # as a JSON string
 
 
-# Example: Using the 'load_dataset' function
-merged_dataset: HafniaDataset = load_dataset(dataset_recipe)
-# You get a few extra things when using `load_dataset`.
-# 1) You get the dataset directly - you don't have to call `build()` on the recipe.
-# 2) The dataset is cached if it already exists, so you don't have to
-#    download or rebuild the dataset on the second run.
-# 3) You can use an implicit form of the recipe. One example of this is that you just specify
-#    the dataset name `load_dataset("mnist")` or path `load_dataset(Path(".data/datasets/mnist"))`
-
-
+### Real-world Example: Merge datasets to create a Person+Vehicle dataset ###
+# 1) The first step is to use the regular 'HafniaDataset' interface to investigate and understand the datasets
+
+# 1a) Explore 'coco-2017'
+coco = HafniaDataset.from_name("coco-2017")
+coco.print_stats()  # Print dataset statistics
+coco_class_names = coco.info.get_task_by_primitive("Bbox").class_names  # Get the class names for the bbox task
+# You will notice coco has 80 classes including 'person' and various vehicle classes such as 'car', 'bus', 'truck', etc.
+# but also many unrelated classes such as 'toaster', 'hair drier', etc.
+
+# 1b) Explore 'midwest-vehicle-detection'
+midwest = HafniaDataset.from_name("midwest-vehicle-detection")
+midwest.print_stats()  # Print dataset statistics
+midwest_class_names = midwest.info.get_task_by_primitive("Bbox").class_names
+# You will also notice midwest has similar classes, but they are named differently, e.g. 'Persons',
+# 'Vehicle.Car', 'Vehicle.Bicycle', etc.
+
+# 2) We will now use the 'HafniaDataset' interface to verify operations (class remapping, merging, filtering)
+
+# 2a) Remap class names to have the same class names across datasets
+mappings_coco = {
+    "person": "Person",
+    "bicycle": "Vehicle",
+    "car": "Vehicle",
+    "motorcycle": "Vehicle",
+    "bus": "Vehicle",
+    "train": "Vehicle",
+    "truck": "Vehicle",
+}
+mapping_midwest = {
+    "Person": "Person",
+    "Vehicle*": "Vehicle",  # Wildcard mapping. Selects class names starting with 'Vehicle.' e.g. 'Vehicle.Bicycle', "Vehicle.Car', etc.
+    "Vehicle.Trailer": OPS_REMOVE_CLASS,  # Use this to remove a class
+}
+coco_remapped = coco.class_mapper(class_mapping=mappings_coco, method="remove_undefined", task_name="object_detection")
+midwest_remapped = midwest.class_mapper(class_mapping=mapping_midwest, task_name="object_detection")
+
+# 2b) Merge datasets
+merged_dataset_all_images = HafniaDataset.from_merge(dataset0=coco_remapped, dataset1=midwest_remapped)
+
+# 2c) Remove images without 'Person' or 'Vehicle' annotations
+merged_dataset = merged_dataset_all_images.select_samples_by_class_name(
+    name=["Person", "Vehicle"], task_name="object_detection"
+)
+merged_dataset.print_stats()
+
+# 3) Once you have verified operations using the 'HafniaDataset' interface, you can convert
+# the operations to a single 'DatasetRecipe'
+merged_recipe = DatasetRecipe.from_merge(
+    recipe0=DatasetRecipe.from_name("coco-2017").class_mapper(
+        class_mapping=mappings_coco, method="remove_undefined", task_name="object_detection"
+    ),
+    recipe1=DatasetRecipe.from_name("midwest-vehicle-detection").class_mapper(
+        class_mapping=mapping_midwest, task_name="object_detection"
+    ),
+).select_samples_by_class_name(name=["Person", "Vehicle"], task_name="object_detection")
+
+# 3a) Verify again on the sample datasets, that the recipe works and can build as a dataset
+merged_dataset = merged_recipe.build()
+merged_dataset.print_stats()
+
+# 3b) Optionally: Save the recipe to file
+path_recipe = Path(".data/dataset_recipes/example-merged-person-vehicle-recipe.json")
+merged_recipe.as_json_file(path_recipe)
+if utils.is_hafnia_configured():
+    # 3c) Upload dataset recipe to Training-aaS platform
+    recipe_response = merged_recipe.as_platform_recipe(recipe_name="person-vehicle-detection", overwrite=True)
+    print(f"Recipe Name: '{recipe_response['name']}', Recipe id: '{recipe_response['id']}'")
+
+    # 4) The recipe is now available in TaaS for you and other users in your organization
+    # 4a) View recipes from your terminal with 'hafnia dataset-recipe ls'
+    # 4b) (Coming soon) Or go to 'Dataset Recipes' in the TaaS web platform:  https://hafnia.milestonesys.com/training-aas/dataset-recipes
+
+    # 5) Launch an experiment with the dataset:
+    # 5a) Using the CLI:
+    #   'hafnia experiment create --dataset-recipe person-vehicle-detection --trainer-path ../trainer-classification'
+    # 5b) (Coming soon) Or through the TaaS web platform:  https://hafnia.milestonesys.com/training-aas/experiments
+
+# 6) Monitor and manage your experiments
+# 6a) View experiments using the web platform https://staging02.mdi.milestonesys.com/training-aas/experiments
+# 6b) Or use the CLI: 'hafnia experiment ls'
 ### DatasetRecipe Implicit Form ###
 # Below we demonstrate the difference between implicit and explicit forms of dataset recipes.
 # Example: Get dataset by name with implicit and explicit forms
-dataset = load_dataset("mnist")  # Implicit form
-dataset = load_dataset(DatasetRecipe.from_name(name="mnist"))  # Explicit form
+recipe_implicit_form = "mnist"
+recipe_explicit_form = DatasetRecipe.from_name(name="mnist")
+
+# The implicit form can now be loaded and built as a dataset
+dataset_implicit = DatasetRecipe.from_implicit_form(recipe_implicit_form).build()
+# Or directly as a dataset
+dataset_implicit = HafniaDataset.from_recipe(recipe_implicit_form)
+
 
 # Example: Get dataset from path with implicit and explicit forms:
-dataset = load_dataset(Path(".data/datasets/mnist"))  # Implicit form
-dataset = load_dataset(DatasetRecipe.from_path(path_folder=Path(".data/datasets/mnist")))  # Explicit form
+recipe_implicit_form = Path(".data/datasets/mnist")
+recipe_explicit_form = DatasetRecipe.from_path(path_folder=Path(".data/datasets/mnist"))
 
 # Example: Merge datasets with implicit and explicit forms
-dataset = load_dataset(("mnist", "mnist"))  # Implicit form
-dataset = load_dataset(  # Explicit form
-    DatasetRecipe.from_merger(
-        recipes=[
-            DatasetRecipe.from_name(name="mnist"),
-            DatasetRecipe.from_name(name="mnist"),
-        ]
-    )
+recipe_implicit_form = ("mnist", "mnist")
+recipe_explicit_form = DatasetRecipe.from_merger(
+    recipes=[
+        DatasetRecipe.from_name(name="mnist"),
+        DatasetRecipe.from_name(name="mnist"),
+    ]
 )
 
 # Example: Define a dataset with transformations using implicit and explicit forms
-dataset = load_dataset(["mnist", SelectSamples(n_samples=20), Shuffle()])  # Implicit form
-dataset = load_dataset(DatasetRecipe.from_name(name="mnist").select_samples(n_samples=20).shuffle())  # Explicit form
+recipe_implicit_form = ["mnist", SelectSamples(n_samples=20), Shuffle()]
+recipe_explicit_form = DatasetRecipe.from_name(name="mnist").select_samples(n_samples=20).shuffle()
 
 
 # Example: Complex nested example with implicit vs explicit forms
 # Implicit form of a complex dataset recipe
 split_ratio = {"train": 0.8, "val": 0.1, "test": 0.1}
-implicit_recipe = (
+recipe_implicit_complex = (
     ("mnist", "mnist"),
     [Path(".data/datasets/mnist"), SelectSamples(n_samples=30), SplitsByRatios(split_ratios=split_ratio)],
     ["mnist", SelectSamples(n_samples=20), Shuffle()],
 )
 
 # Explicit form of the same complex dataset recipe
-explicit_recipe = DatasetRecipe.from_merger(
+recipe_explicit_complex = DatasetRecipe.from_merger(
     recipes=[
         DatasetRecipe.from_merger(
             recipes=[
@@ -165,10 +240,10 @@ explicit_recipe = DatasetRecipe.from_merger(
 
 
 # To convert from implicit to explicit recipe form, you can use the `from_implicit_form` method.
-explicit_recipe_from_implicit = DatasetRecipe.from_implicit_form(implicit_recipe)
+explicit_recipe_from_implicit = DatasetRecipe.from_implicit_form(recipe_implicit_complex)
 rprint("Converted explicit recipe:")
 rprint(explicit_recipe_from_implicit)
 
 # Verify that the conversion produces the same result
-assert explicit_recipe_from_implicit == explicit_recipe
+assert explicit_recipe_from_implicit == recipe_explicit_complex
 rprint("Conversion successful - recipes are equivalent!")

{hafnia-0.3.0 → hafnia-0.4.1}/examples/example_hafnia_dataset.py

@@ -5,7 +5,6 @@ import numpy as np
 from PIL import Image
 from rich import print as rprint
 
-from hafnia.data import load_dataset
 from hafnia.dataset.dataset_names import SplitName
 from hafnia.dataset.hafnia_dataset import DatasetInfo, HafniaDataset, Sample, TaskInfo
 from hafnia.dataset.primitives.bbox import Bbox
@@ -19,7 +18,7 @@ from hafnia.dataset.primitives.polygon import Polygon
 # And configure it with your Hafnia account:
 #   hafnia configure
 
-# Load dataset
+# Load sample dataset
 dataset = HafniaDataset.from_name("mnist")
 
 # Dataset information is stored in 'dataset.info'
@@ -34,8 +33,8 @@ dataset.print_class_distribution()
 dataset.print_stats()  # Print verbose dataset statistics
 
 # Get dataset stats
-dataset.class_counts_all()  # Get class counts for all tasks
-dataset.class_counts_for_task(primitive=Classification)  # Get class counts for a specific task
+dataset.calculate_class_counts()  # Get class counts for all tasks
+dataset.calculate_task_class_counts(primitive=Classification)  # Get class counts for a specific task
 
 # Create a dataset split for training
 dataset_train = dataset.create_split_dataset("train")
@@ -53,6 +52,12 @@ new_dataset_splits = dataset.splits_by_ratios(split_ratios)
 # Get only samples with specific class names
 dataset_ones = dataset.select_samples_by_class_name(name="1 - one", primitive=Classification)
 
+# Get access to a few full and public dataset through Hafnia (no login required)
+# Available datasets: "mnist", "caltech-101", "caltech-256", "cifar10", "cifar100"
+public_dataset = HafniaDataset.from_name_public_dataset("mnist", n_samples=100)
+public_dataset.print_stats()
+
+
 # Rename class names with mapping
 class_mapping_strict = {
     "0 - zero": "even",  # "0 - zero" will be renamed to "even". "even" appear first and get class index 0
@@ -70,7 +75,7 @@ dataset_mapped = dataset.class_mapper(class_mapping=class_mapping_strict)
 dataset_mapped.print_class_distribution()
 
 # Support Chaining Operations (load, shuffle, select samples)
-dataset = load_dataset("midwest-vehicle-detection").shuffle(seed=42).select_samples(n_samples=10)
+dataset = HafniaDataset.from_name("midwest-vehicle-detection").shuffle(seed=42).select_samples(n_samples=10)
 
 
 # Write dataset to disk
@@ -81,9 +86,19 @@ dataset.write(path_dataset)
 # Load dataset from disk
 dataset_again = HafniaDataset.from_path(path_dataset)
 
+## Dataset importers and exporters ##
+dataset_coco = HafniaDataset.from_name("coco-2017").select_samples(n_samples=5, seed=42)
+path_yolo_format = Path(".data/tmp/yolo_dataset")
+
+# Export dataset to YOLO format
+dataset_coco.to_yolo_format(path_export_yolo_dataset=path_yolo_format)
+
+# Import dataset from YOLO format
+dataset_coco_imported = HafniaDataset.from_yolo_format(path_yolo_format)
 
+## Custom dataset operations and statistics ##
 # Want custom dataset transformations or statistics? Use the polars table (dataset.samples) directly
-n_objects = dataset.samples["objects"].list.len().sum()
+n_objects = dataset.samples["bboxes"].list.len().sum()
 n_objects = dataset.samples[Bbox.column_name()].list.len().sum()  # Use Bbox.column_name() to avoid magic variables
 n_classifications = dataset.samples[Classification.column_name()].list.len().sum()
 
@@ -101,7 +116,7 @@ for sample_dict in dataset_train:
 # Unpack dict into a Sample-object! Important for data validation, useability, IDE completion and mypy hints
 sample: Sample = Sample(**sample_dict)
 
-objects: List[Bbox] = sample.objects  # Use 'sample.objects' access bounding boxes as a list of Bbox objects
+bboxes: List[Bbox] = sample.bboxes  # Use 'sample.bboxes' access bounding boxes as a list of Bbox objects
 bitmasks: List[Bitmask] = sample.bitmasks  # Use 'sample.bitmasks' to access bitmasks as a list of Bitmask objects
 polygons: List[Polygon] = sample.polygons  # Use 'sample.polygons' to access polygons as a list of Polygon objects
 classifications: List[Classification] = sample.classifications  # As a list of Classification objects
@@ -124,12 +139,12 @@ for i_fake_sample in range(5):
     bboxes = [Bbox(top_left_x=0.1, top_left_y=0.20, width=0.1, height=0.2, class_name="car")]
     classifications = [Classification(class_name="vehicle", class_idx=0)]
     sample = Sample(
-        file_name=f"path/to/image_{i_fake_sample:05}.jpg",
+        file_path=f"path/to/image_{i_fake_sample:05}.jpg",
         height=480,
         width=640,
         split="train",
         tags=["sample"],
-        objects=bboxes,
+        bboxes=bboxes,
         classifications=classifications,
     )
     fake_samples.append(sample)

{hafnia-0.3.0 → hafnia-0.4.1}/examples/example_torchvision_dataloader.py

@@ -7,12 +7,12 @@ from torch.utils.data import DataLoader
 from torchvision.transforms import v2
 
 from hafnia import torch_helpers
-from hafnia.data import load_dataset
+from hafnia.dataset.hafnia_dataset import HafniaDataset
 
 if __name__ == "__main__":
     torch.manual_seed(1)
     # Load Hugging Face dataset
-    dataset = load_dataset("midwest-vehicle-detection")
+    dataset = HafniaDataset.from_name("midwest-vehicle-detection")
 
     # Define transforms
     train_transforms = v2.Compose(

{hafnia-0.3.0 → hafnia-0.4.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "hafnia"
-version = "0.3.0"
+version = "0.4.1"
 description = "Python SDK for communication with Hafnia platform."
 readme = "README.md"
 authors = [
@@ -13,7 +13,7 @@ dependencies = [
     "click>=8.1.8",
     "emoji>=2.14.1",
     "flatten-dict>=0.4.2",
-    "mlflow>=3.2.0",
+    "keyring>=25.6.0",
     "more-itertools>=10.7.0",
     "opencv-python-headless>=4.11.0.86",
     "pathspec>=0.12.1",
@@ -24,10 +24,11 @@ dependencies = [
     "pydantic>=2.10.4",
     "rich>=13.9.4",
     "s5cmd>=0.2.0",
-    "sagemaker-mlflow>=0.1.0",
     "seedir>=0.5.0",
-    "tqdm>=4.67.1",
     "xxhash>=3.5.0",
+    "mlflow>=3.4.0",
+    "sagemaker-mlflow>=0.1.0",
+    "mcp==1.16.0",
 ]
 
 [dependency-groups]
@@ -37,7 +38,7 @@ dev = [
     "torch>=2.6.0",
     "torchvision>=0.21.0",
     "flatten-dict>=0.4.2",
-    "pytest-cov>=7.0.0",    
+    "pytest-cov>=7.0.0",
 ]
 
 test = ["pytest>=8.3.4", "pre-commit>=4.2.0", "ruff>=0.9.1"]
@@ -71,5 +72,9 @@ disallow_incomplete_defs = false
 disallow_untyped_calls = false
 warn_unused_ignores = false
 
+[[tool.mypy.overrides]]
+module = "yaml"
+ignore_missing_imports = true
+
 [tool.pytest.ini_options]
 markers = ["slow: marks tests as slow (deselect with '-m \"not slow\"')"]

{hafnia-0.3.0 → hafnia-0.4.1}/src/cli/__main__.py

@@ -37,7 +37,9 @@ def configure(cfg: Config) -> None:
 
     platform_url = click.prompt("Hafnia Platform URL", type=str, default=consts.DEFAULT_API_URL)
 
-    cfg_profile = ConfigSchema(api_key=api_key, platform_url=platform_url)
+    use_keychain = click.confirm("Store API key in system keychain?", default=False)
+
+    cfg_profile = ConfigSchema(platform_url=platform_url, api_key=api_key, use_keychain=use_keychain)
     cfg.add_profile(profile_name, cfg_profile, set_active=True)
     cfg.save_config()
     profile_cmds.profile_show(cfg)