feat(client): adapt openAPI "getStatistics"

Graviti-AI · Sep 29, 2021 · 69b868c · 69b868c
1 parent b65e505
commit 69b868c
Show file tree

Hide file tree

Showing 9 changed files with 265 additions and 28 deletions.
diff --git a/docs/source/api/client/client_module.rst b/docs/source/api/client/client_module.rst
@@ -16,3 +16,4 @@ tensorbay.client
    version
    diff
    profile
+   statistics
diff --git a/docs/source/api/client/statistics.rst b/docs/source/api/client/statistics.rst
@@ -0,0 +1,6 @@
+tensorbay.client.statistics
+===========================
+
+.. automodule:: tensorbay.client.statistics
+   :members:
+   :show-inheritance:
diff --git a/docs/source/examples/get_label_statistics.rst b/docs/source/examples/get_label_statistics.rst
@@ -0,0 +1,65 @@
+######################
+ Get Label Statistics
+######################
+
+This topic describes the get label statistics operation.
+
+Label statistics of dataset could be obtained via :func:`~tensorbay.client.dataset.DatasetClientBase.get_label_statistics`
+as follows:
+
+    >>> from tensorbay import GAS
+    >>> ACCESS_KEY = "Accesskey-*****"
+    >>> gas = GAS(ACCESS_KEY)
+    >>> dataset_client = gas.get_dataset("targetDataset")
+    >>> statistics = dataset_client.get_label_statistics()
+    >>> statistics
+    Statistics {
+        'BOX2D': {...},
+        'BOX3D': {...},
+        'KEYPOINTS2D': {...}
+    }
+
+The details of the statistics structure for the targetDataset are as follows:
+
+.. code-block:: json
+
+    {
+        "BOX2D": {
+            "quantity": 1508722,
+            "categories": [
+                {
+                    "name": "vehicle.bike",
+                    "quantity": 8425,
+                    "attributes": [
+                        {
+                            "name": "trafficLightColor",
+                            "enum": ["none", "red", "yellow"],
+                            "quantities": [8420, 3, 2]
+                        }
+                    ]
+                }
+            ],
+            "attributes": [
+                {
+                    "name": "trafficLightColor",
+                    "enum": ["none", "red", "yellow", "green"],
+                    "quantities": [1356224, 54481, 4107, 93910]
+                }
+            ]
+        },
+        "BOX3D": {
+            "quantity": 1234
+        },
+        "KEYPOINTS2D":{
+            "quantity": 43234,
+            "categories":[
+                {
+                    "name": "person.person",
+                    "quantity": 43234
+                }
+            ]
+        }
+    }
+
+.. note::
+   The method :func:`~tensorbay.client.statistics.Statistics.dumps` of :class:`~tensorbay.client.statistics.Statistics` can dump the statistics into a dict.
diff --git a/docs/source/features/dataset_management.rst b/docs/source/features/dataset_management.rst
@@ -10,6 +10,7 @@ This topic describes dataset management, including:
 - :ref:`features/dataset_management:Update Dataset`
 - :ref:`features/dataset_management:Move and Copy`
 - :ref:`features/dataset_management:Merge Datasets`
+- :ref:`features/dataset_management:Get Label Statistics`
 
 
 ******************
@@ -120,3 +121,11 @@ Please see :ref:`Move and copy<examples/move_and_copy:Move And Copy>` example fo
 Since TensorBay supports copy operation between different datasets, users can use it to merge datasets.
 
 Please see :ref:`examples/merge_datasets:Merge Datasets` example for more details.
+
+**********************
+ Get Label Statistics
+**********************
+
+TensorBay supports getting label statistics of dataset.
+
+Please see :ref:`examples/get_label_statistics:Get Label Statistics` example for more details.
diff --git a/docs/source/quick_start/examples.rst b/docs/source/quick_start/examples.rst
@@ -10,35 +10,36 @@ The following table lists a series of examples to help developers to use TensorB
    :align: center
    :widths: auto
 
-   ======================================================== =========================================================================
-   Examples                                                 Description
-   ======================================================== =========================================================================
-   :ref:`examples/DogsVsCats:Dogs vs Cats`                  | Topic: Dataset Management
-                                                            | Data Type: Image
-                                                            | Label Type: :ref:`reference/label_format/Classification:Classification`
-   :ref:`examples/Newsgroups20:20 Newsgroups`               | Topic: Dataset Management
-                                                            | Data Type: Text
-                                                            | Label Type: :ref:`reference/label_format/Classification:Classification`
-   :ref:`examples/BSTLD:BSTLD`                              | Topic: Dataset Management
-                                                            | Data Type: Image
-                                                            | Label Type: :ref:`reference/label_format/Box2D:Box2D`
-   :ref:`examples/NeolixOD:Neolix OD`                       | Topic: Dataset Management
-                                                            | Data Type: Point Cloud
-                                                            | Label Type: :ref:`reference/label_format/Box3D:Box3D`
-   :ref:`examples/LeedsSportsPose:Leeds Sports Pose`        | Topic: Dataset Management
-                                                            | Data Type: Image
-                                                            | Label Type: :ref:`reference/label_format/Keypoints2D:Keypoints2D`
-   :ref:`examples/THCHS30:THCHS-30`                         | Topic: Dataset Management
-                                                            | Data Type: Audio
-                                                            | Label Type: :ref:`reference/label_format/Sentence:Sentence`
-   :ref:`examples/VOC2012Segmentation:VOC2012 Segmentation` | Topic: Dataset Management
-                                                            | Data Type: Image
-                                                            | Label Types: :ref:`reference/label_format/SemanticMask:SemanticMask`,
-                                                              :ref:`reference/label_format/InstanceMask:InstanceMask`
-   :ref:`examples/update_dataset:Update Dataset`            | Topic: Update Dataset
-   :ref:`examples/move_and_copy:Move And Copy`              | Topic: Move And Copy
+   ========================================================= =========================================================================
+   Examples                                                  Description
+   ========================================================= =========================================================================
+   :ref:`examples/DogsVsCats:Dogs vs Cats`                   | Topic: Dataset Management
+                                                             | Data Type: Image
+                                                             | Label Type: :ref:`reference/label_format/Classification:Classification`
+   :ref:`examples/Newsgroups20:20 Newsgroups`                | Topic: Dataset Management
+                                                             | Data Type: Text
+                                                             | Label Type: :ref:`reference/label_format/Classification:Classification`
+   :ref:`examples/BSTLD:BSTLD`                               | Topic: Dataset Management
+                                                             | Data Type: Image
+                                                             | Label Type: :ref:`reference/label_format/Box2D:Box2D`
+   :ref:`examples/NeolixOD:Neolix OD`                        | Topic: Dataset Management
+                                                             | Data Type: Point Cloud
+                                                             | Label Type: :ref:`reference/label_format/Box3D:Box3D`
+   :ref:`examples/LeedsSportsPose:Leeds Sports Pose`         | Topic: Dataset Management
+                                                             | Data Type: Image
+                                                             | Label Type: :ref:`reference/label_format/Keypoints2D:Keypoints2D`
+   :ref:`examples/THCHS30:THCHS-30`                          | Topic: Dataset Management
+                                                             | Data Type: Audio
+                                                             | Label Type: :ref:`reference/label_format/Sentence:Sentence`
+   :ref:`examples/VOC2012Segmentation:VOC2012 Segmentation`  | Topic: Dataset Management
+                                                             | Data Type: Image
+                                                             | Label Types: :ref:`reference/label_format/SemanticMask:SemanticMask`,
+                                                               :ref:`reference/label_format/InstanceMask:InstanceMask`
+   :ref:`examples/update_dataset:Update Dataset`             | Topic: Update Dataset
+   :ref:`examples/move_and_copy:Move And Copy`               | Topic: Move And Copy
    :ref:`examples/merge_datasets:Merge Datasets`             | Topic: Merge Datasets
-   ======================================================== =========================================================================
+   :ref:`examples/get_label_statistics:Get Label Statistics` | Topic: Get Label Statistics
+   ========================================================= =========================================================================
 
 .. toctree::
    :hidden:
@@ -54,3 +55,4 @@ The following table lists a series of examples to help developers to use TensorB
    ../examples/update_dataset
    ../examples/move_and_copy
    ../examples/merge_datasets
+   ../examples/get_label_statistics
diff --git a/tensorbay/client/dataset.py b/tensorbay/client/dataset.py
@@ -35,6 +35,7 @@
 from .log import UPLOAD_SEGMENT_RESUME_TEMPLATE
 from .requests import Tqdm, multithread_upload
 from .segment import _STRATEGIES, FusionSegmentClient, SegmentClient
+from .statistics import Statistics
 from .status import Status
 from .version import VersionControlClient
 
@@ -278,6 +279,20 @@ def delete_segment(self, name: str) -> None:
 
         self._client.open_api_do("DELETE", "segments", self._dataset_id, json=delete_data)
 
+    def get_label_statistics(self) -> Statistics:
+        """Get label statistics of the dataset.
+
+        Returns:
+            Required :class:`~tensorbay.client.dataset.Statistics`.
+
+        """
+        params: Dict[str, Any] = self._status.get_status_info()
+        return Statistics(
+            self._client.open_api_do(
+                "GET", "labels/statistics", self._dataset_id, params=params
+            ).json()["labelStatistics"]
+        )
+
 
 class DatasetClient(DatasetClientBase):
     """This class defines :class:`DatasetClient`.

diff --git a/tensorbay/client/statistics.py b/tensorbay/client/statistics.py
@@ -0,0 +1,68 @@
+#!/usr/bin/env python3
+#
+# Copyright 2021 Graviti. Licensed under MIT License.
+#
+
+"""Class Statistics.
+
+:class:`Statistics` defines the basic structure of the label statistics obtained by
+:meth:`DatasetClientBase.get_label_statistics`.
+
+"""
+from typing import Any, Dict
+
+from ..utility import UserMapping
+
+
+class Statistics(UserMapping[str, Any]):  # pylint: disable=too-many-ancestors
+    """This class defines the basic structure of the label statistics.
+
+    Arguments:
+        data: The dict containing label statistics.
+
+    """
+
+    def __init__(self, data: Dict[str, Any]) -> None:
+        self._data: Dict[str, Any] = data
+
+    def dumps(self) -> Dict[str, Any]:
+        """Dumps the label statistics into a dict.
+
+        Returns:
+            A dict containing all the information of the label statistics.
+
+        Examples:
+            >>> label_statistics = Statistics(
+            ...     {
+            ...         'BOX3D': {
+            ...             'quantity': 1234
+            ...         },
+            ...         'KEYPOINTS2D': {
+            ...             'quantity': 43234,
+            ...             'categories': [
+            ...                 {
+            ...                     'name': 'person.person',
+            ...                     'quantity': 43234
+            ...                 }
+            ...             ]
+            ...         }
+            ...     }
+            ... )
+            >>> label_statistics.dumps()
+            ... {
+            ...    'BOX3D': {
+            ...        'quantity': 1234
+            ...     },
+            ...    'KEYPOINTS2D': {
+            ...         'quantity': 43234,
+            ...         'categories': [
+            ...             {
+            ...                 'name': 'person.person',
+            ...                 'quantity': 43234
+            ...             }
+            ...         ]
+            ...     }
+            ... }
+
+        """
+        return self._data
diff --git a/tensorbay/client/tests/test_dataset.py b/tensorbay/client/tests/test_dataset.py
@@ -16,6 +16,7 @@
 from ..lazy import ReturnGenerator
 from ..requests import Tqdm
 from ..segment import FusionSegmentClient, SegmentClient
+from ..statistics import Statistics
 from ..status import Status
 from ..struct import ROOT_COMMIT_ID
 from .utility import mock_response
@@ -247,6 +248,38 @@ def test_delete_segment(self, mocker):
             "DELETE", "segments", self.dataset_client._dataset_id, json=delete_data
         )
 
+    def test_get_label_statistics(self, mocker):
+        params = self.dataset_client._status.get_status_info()
+        response_data = {
+            "labelStatistics": {
+                "BOX2D": {
+                    "quantity": 10,
+                    "categories": [
+                        {
+                            "name": "vehicles.bike",
+                            "quantity": 10,
+                            "attributes": [
+                                {
+                                    "name": "trafficLightColor",
+                                    "enum": ["none", "red", "yellow"],
+                                    "quantities": [5, 3, 2],
+                                }
+                            ],
+                        }
+                    ],
+                }
+            }
+        }
+        open_api_do = mocker.patch(
+            f"{gas.__name__}.Client.open_api_do",
+            return_value=mock_response(data=response_data),
+        )
+        statistics1 = self.dataset_client.get_label_statistics()
+        open_api_do.assert_called_once_with(
+            "GET", "labels/statistics", self.dataset_client.dataset_id, params=params
+        )
+        assert statistics1 == Statistics(response_data["labelStatistics"])
+
 
 class TestDatasetClient(TestDatasetClientBase):
     def test__generate_segments(self, mocker):

diff --git a/tests/test_label.py b/tests/test_label.py
@@ -7,6 +7,7 @@
 import pytest
 
 from tensorbay import GAS
+from tensorbay.client.statistics import Statistics
 from tensorbay.dataset import Data, Dataset
 from tensorbay.label import Catalog, InstanceMask, Label, PanopticMask, SemanticMask
 from tensorbay.label.label_mask import RemoteInstanceMask, RemotePanopticMask, RemoteSemanticMask
@@ -108,6 +109,41 @@
     "MULTI_POLYGON": MULTI_POLYGON_LABEL,
     "RLE": RLE_LABEL,
 }
+STATISTICS = {
+    "MULTI_POLYGON": {
+        "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+        "categories": [
+            {
+                "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+                "name": "cat",
+                "quantity": 1,
+            }
+        ],
+        "quantity": 1,
+    },
+    "MULTI_POLYLINE2D": {
+        "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+        "categories": [
+            {
+                "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+                "name": "cat",
+                "quantity": 1,
+            }
+        ],
+        "quantity": 1,
+    },
+    "RLE": {
+        "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+        "categories": [
+            {
+                "attributes": [{"enum": ["male"], "name": "gender", "quantities": [1]}],
+                "name": "cat",
+                "quantity": 1,
+            }
+        ],
+        "quantity": 1,
+    },
+}
 
 
 @pytest.fixture
@@ -143,6 +179,8 @@ def test_upload_dataset_with_label(self, accesskey, url, tmp_path):
         assert dataset.catalog == Catalog.loads(CATALOG_CONTENTS)
         assert dataset[0][0].label == Label.loads(LABEL)
 
+        statistics1 = dataset_client.get_label_statistics()
+        assert statistics1 == Statistics(STATISTICS)
         gas_client.delete_dataset(dataset_name)
 
     def test_upload_dataset_with_mask(self, accesskey, url, tmp_path, mask_file):
-Original file line number
+Diff line change
@@ Expand Up / @@ -16,3 +16,4 @@ tensorbay.client @@
        version
        diff
        profile
+       statistics