Add detector metadata (#148)

* initial commit * Automatically reformatting code * add metadata field to the detector model * Automatically reformatting code * small clean up * make pylint happy * wip * add metadata field to detector serializer and api spec * address pr feedback --------- Co-authored-by: Auto-format Bot <[email protected]>
groundlight · Jan 18, 2024 · bb54236 · bb54236
1 parent c4d3f07
commit bb54236
Show file tree

Hide file tree

Showing 11 changed files with 119 additions and 5 deletions.
diff --git a/generated/README.md b/generated/README.md
@@ -81,6 +81,7 @@ with openapi_client.ApiClient(configuration) as api_client:
         group_name="group_name_example",
         confidence_threshold=0.9,
         pipeline_config="pipeline_config_example",
+        metadata="metadata_example",
     ) # DetectorCreationInput | 
 
     try:

diff --git a/generated/docs/Detector.md b/generated/docs/Detector.md
@@ -12,6 +12,7 @@ Name | Type | Description | Notes
 **query** | **str** | A question about the image. | [readonly] 
 **group_name** | **str** | Which group should this detector be part of? | [readonly] 
 **confidence_threshold** | **float** | If the detector&#39;s prediction is below this confidence threshold, send the image query for human review. | [optional]  if omitted the server will use the default value of 0.9
+**metadata** | **{str: (bool, date, datetime, dict, float, int, list, str, none_type)}, none_type** | A dictionary of custom key/value metadata to associate with the detector (limited to 1KB). | [optional] 
 **any string name** | **bool, date, datetime, dict, float, int, list, str, none_type** | any string name can be used but the value must be the correct type | [optional]
 
 [[Back to Model list]](../README.md#documentation-for-models) [[Back to API list]](../README.md#documentation-for-api-endpoints) [[Back to README]](../README.md)

diff --git a/generated/docs/DetectorCreationInput.md b/generated/docs/DetectorCreationInput.md
@@ -10,6 +10,7 @@ Name | Type | Description | Notes
 **group_name** | **str** | Which group should this detector be part of? | [optional] 
 **confidence_threshold** | **float** | If the detector&#39;s prediction is below this confidence threshold, send the image query for human review. | [optional]  if omitted the server will use the default value of 0.9
 **pipeline_config** | **str, none_type** | (Advanced usage) Configuration to instantiate a specific prediction pipeline. | [optional] 
+**metadata** | **str, none_type** | A dictionary of custom key/value metadata to associate with the detector (limited to 1KB). | [optional] 
 **any string name** | **bool, date, datetime, dict, float, int, list, str, none_type** | any string name can be used but the value must be the correct type | [optional]
 
 [[Back to Model list]](../README.md#documentation-for-models) [[Back to API list]](../README.md#documentation-for-api-endpoints) [[Back to README]](../README.md)

diff --git a/generated/docs/DetectorsApi.md b/generated/docs/DetectorsApi.md
@@ -54,6 +54,7 @@ with openapi_client.ApiClient(configuration) as api_client:
         group_name="group_name_example",
         confidence_threshold=0.9,
         pipeline_config="pipeline_config_example",
+        metadata="metadata_example",
     ) # DetectorCreationInput | 
 
     # example passing only required values which don't have defaults set

diff --git a/generated/model.py b/generated/model.py
@@ -1,6 +1,6 @@
 # generated by datamodel-codegen:
 #   filename:  public-api.yaml
-#   timestamp: 2023-11-09T05:00:29+00:00
+#   timestamp: 2024-01-16T18:09:48+00:00
 
 from __future__ import annotations
 
@@ -33,6 +33,9 @@ class DetectorCreationInput(BaseModel):
     pipeline_config: Optional[constr(max_length=8192)] = Field(
         None, description="(Advanced usage) Configuration to instantiate a specific prediction pipeline."
     )
+    metadata: Optional[str] = Field(
+        None, description="A dictionary of custom key/value metadata to associate with the detector (limited to 1KB)."
+    )
 
 
 class DetectorTypeEnum(Enum):
@@ -60,6 +63,9 @@ class Detector(BaseModel):
             "If the detector's prediction is below this confidence threshold, send the image query for human review."
         ),
     )
+    metadata: Optional[Dict[str, Any]] = Field(
+        None, description="A dictionary of custom key/value metadata to associate with the detector (limited to 1KB)."
+    )
 
 
 class ImageQuery(BaseModel):
@@ -87,4 +93,4 @@ class PaginatedImageQueryList(BaseModel):
     count: Optional[int] = Field(None, example=123)
     next: Optional[AnyUrl] = Field(None, example="http://api.example.org/accounts/?page=4")
     previous: Optional[AnyUrl] = Field(None, example="http://api.example.org/accounts/?page=2")
-    results: Optional[List[ImageQuery]] = None
+    results: Optional[List[ImageQuery]] = None
diff --git a/generated/openapi_client/model/detector.py b/generated/openapi_client/model/detector.py
@@ -122,6 +122,10 @@ def openapi_types():
             "query": (str,),  # noqa: E501
             "group_name": (str,),  # noqa: E501
             "confidence_threshold": (float,),  # noqa: E501
+            "metadata": (
+                {str: (bool, date, datetime, dict, float, int, list, str, none_type)},
+                none_type,
+            ),  # noqa: E501
         }
 
     @cached_property
@@ -136,6 +140,7 @@ def discriminator():
         "query": "query",  # noqa: E501
         "group_name": "group_name",  # noqa: E501
         "confidence_threshold": "confidence_threshold",  # noqa: E501
+        "metadata": "metadata",  # noqa: E501
     }
 
     read_only_vars = {
@@ -193,6 +198,7 @@ def _from_openapi_data(cls, id, type, created_at, name, query, group_name, *args
                                 through its discriminator because we passed in
                                 _visited_composed_classes = (Animal,)
             confidence_threshold (float): If the detector's prediction is below this confidence threshold, send the image query for human review.. [optional] if omitted the server will use the default value of 0.9  # noqa: E501
+            metadata ({str: (bool, date, datetime, dict, float, int, list, str, none_type)}, none_type): A dictionary of custom key/value metadata to associate with the detector (limited to 1KB).. [optional]  # noqa: E501
         """
 
         _check_type = kwargs.pop("_check_type", True)
@@ -287,6 +293,7 @@ def __init__(self, name, *args, **kwargs):  # noqa: E501
                                 through its discriminator because we passed in
                                 _visited_composed_classes = (Animal,)
             confidence_threshold (float): If the detector's prediction is below this confidence threshold, send the image query for human review.. [optional] if omitted the server will use the default value of 0.9  # noqa: E501
+            metadata ({str: (bool, date, datetime, dict, float, int, list, str, none_type)}, none_type): A dictionary of custom key/value metadata to associate with the detector (limited to 1KB).. [optional]  # noqa: E501
         """
 
         _check_type = kwargs.pop("_check_type", True)

diff --git a/generated/openapi_client/model/detector_creation_input.py b/generated/openapi_client/model/detector_creation_input.py
@@ -114,6 +114,10 @@ def openapi_types():
                 str,
                 none_type,
             ),  # noqa: E501
+            "metadata": (
+                str,
+                none_type,
+            ),  # noqa: E501
         }
 
     @cached_property
@@ -126,6 +130,7 @@ def discriminator():
         "group_name": "group_name",  # noqa: E501
         "confidence_threshold": "confidence_threshold",  # noqa: E501
         "pipeline_config": "pipeline_config",  # noqa: E501
+        "metadata": "metadata",  # noqa: E501
     }
 
     read_only_vars = {}
@@ -175,6 +180,7 @@ def _from_openapi_data(cls, name, query, *args, **kwargs):  # noqa: E501
             group_name (str): Which group should this detector be part of?. [optional]  # noqa: E501
             confidence_threshold (float): If the detector's prediction is below this confidence threshold, send the image query for human review.. [optional] if omitted the server will use the default value of 0.9  # noqa: E501
             pipeline_config (str, none_type): (Advanced usage) Configuration to instantiate a specific prediction pipeline.. [optional]  # noqa: E501
+            metadata (str, none_type): A dictionary of custom key/value metadata to associate with the detector (limited to 1KB).. [optional]  # noqa: E501
         """
 
         _check_type = kwargs.pop("_check_type", True)
@@ -270,6 +276,7 @@ def __init__(self, name, query, *args, **kwargs):  # noqa: E501
             group_name (str): Which group should this detector be part of?. [optional]  # noqa: E501
             confidence_threshold (float): If the detector's prediction is below this confidence threshold, send the image query for human review.. [optional] if omitted the server will use the default value of 0.9  # noqa: E501
             pipeline_config (str, none_type): (Advanced usage) Configuration to instantiate a specific prediction pipeline.. [optional]  # noqa: E501
+            metadata (str, none_type): A dictionary of custom key/value metadata to associate with the detector (limited to 1KB).. [optional]  # noqa: E501
         """
 
         _check_type = kwargs.pop("_check_type", True)

diff --git a/spec/public-api.yaml b/spec/public-api.yaml
@@ -265,6 +265,12 @@ components:
           description:
             If the detector's prediction is below this confidence threshold,
             send the image query for human review.
+        metadata:
+          type: object 
+          nullable: true
+          description:
+            A dictionary of custom key/value metadata to associate with the detector
+            (limited to 1KB). This is encoded as a URL-safe, base64-encoded JSON string.
       required:
         - created_at
         - group_name
@@ -303,6 +309,12 @@ components:
           nullable: true
           description: (Advanced usage) Configuration to instantiate a specific prediction pipeline.
           maxLength: 8192
+        metadata:
+          type: string 
+          nullable: true
+          description:
+            A dictionary of custom key/value metadata to associate with the detector
+            (limited to 1KB). This is encoded as a URL-safe, base64-encoded JSON string.
       required:
         # TODO: make name optional - that's how the web version is going.
         - name

diff --git a/src/groundlight/client.py b/src/groundlight/client.py
@@ -187,6 +187,7 @@ def create_detector(
         *,
         confidence_threshold: Optional[float] = None,
         pipeline_config: Optional[str] = None,
+        metadata: Union[dict, str, None] = None,
     ) -> Detector:
         """
         Create a new detector with a given name and query
@@ -199,13 +200,19 @@ def create_detector(
 
         :param pipeline_config: the pipeline config
 
+        :param metadata: A dictionary or JSON string of custom key/value metadata to associate with
+            the detector (limited to 1KB). You can retrieve this metadata later by calling
+            `get_detector()`.
+
         :return: Detector
         """
         detector_creation_input = DetectorCreationInput(name=name, query=query)
         if confidence_threshold is not None:
             detector_creation_input.confidence_threshold = confidence_threshold
         if pipeline_config is not None:
             detector_creation_input.pipeline_config = pipeline_config
+        if metadata is not None:
+            detector_creation_input.metadata = str(url_encode_dict(metadata, name="metadata", size_limit_bytes=1024))
         obj = self.detectors_api.create_detector(detector_creation_input)
         return Detector.parse_obj(obj.to_dict())
 
@@ -216,6 +223,7 @@ def get_or_create_detector(
         *,
         confidence_threshold: Optional[float] = None,
         pipeline_config: Optional[str] = None,
+        metadata: Union[dict, str, None] = None,
     ) -> Detector:
         """
         Tries to look up the detector by name.  If a detector with that name, query, and
@@ -230,6 +238,10 @@ def get_or_create_detector(
 
         :param pipeline_config: the pipeline config
 
+        :param metadata: A dictionary or JSON string of custom key/value metadata to associate with
+            the detector (limited to 1KB). You can retrieve this metadata later by calling
+            `get_detector()`.
+
         :return: Detector
         """
         try:
@@ -241,6 +253,7 @@ def get_or_create_detector(
                 query=query,
                 confidence_threshold=confidence_threshold,
                 pipeline_config=pipeline_config,
+                metadata=metadata,
             )
 
         # TODO: We may soon allow users to update the retrieved detector's fields.

diff --git a/src/groundlight/encodings.py b/src/groundlight/encodings.py
@@ -1,6 +1,5 @@
 import base64
 import json
-import sys
 from typing import Dict, Optional, Union
 
 
@@ -34,7 +33,7 @@ def url_encode_dict(maybe_dict: Union[Dict, str], name: str, size_limit_bytes: O
     data_json = json.dumps(maybe_dict)
 
     if size_limit_bytes is not None:
-        size_bytes = sys.getsizeof(data_json)
+        size_bytes = len(data_json)
         if size_bytes > size_limit_bytes:
             raise ValueError(f"`{name}` is too large: {size_bytes} bytes > {size_limit_bytes} bytes limit.")
 

diff --git a/test/integration/test_groundlight.py b/test/integration/test_groundlight.py
@@ -2,6 +2,8 @@
 # ruff: noqa: F403,F405
 # pylint: disable=wildcard-import,unused-wildcard-import,redefined-outer-name,import-outside-toplevel
 import json
+import random
+import string
 import time
 from datetime import datetime
 from typing import Any, Dict, Optional, Union
@@ -28,6 +30,27 @@ def is_valid_display_result(result: Any) -> bool:
     return True
 
 
+def generate_random_dict(target_size_bytes=1024, key_length=8, value_length=10) -> Dict[str, str]:
+    """
+    Generate a random dictionary with an approximate size in bytes.
+    """
+    key_chars = string.ascii_lowercase + string.digits
+    value_chars = string.ascii_letters + string.digits
+
+    random_dict: Dict[str, str] = {}
+    while len(json.dumps(random_dict).encode("utf-8")) < target_size_bytes:
+        key = "".join(random.choice(key_chars) for _ in range(key_length))
+        value = "".join(random.choice(value_chars) for _ in range(value_length))
+        random_dict[key] = value
+
+        # Check if adding another pair would likely exceed the size
+        # The 4 is for the quotes around the key and value, and the colon and comma
+        if len(json.dumps(random_dict).encode("utf-8")) + key_length + value_length + 4 > target_size_bytes:
+            break
+
+    return random_dict
+
+
 def is_valid_display_label(label: str) -> bool:
     """Is the image query result label valid to display to the user?."""
     # NOTE: For now, we strictly only show UPPERCASE labels to the user.
@@ -252,7 +275,50 @@ def test_submit_image_query_with_human_review_param(gl: Groundlight, detector: D
         assert is_valid_display_result(_image_query.result)
 
 
-@pytest.mark.skip_for_edge_endpoint(reason="The edge-endpoint does not support passing metadata.")
+@pytest.mark.skip_for_edge_endpoint(reason="The edge-endpoint does not support passing detector metadata.")
+def test_create_detector_with_metadata(gl: Groundlight):
+    name = f"Test {datetime.utcnow()}"  # Need a unique name
+    query = "Is there a dog?"
+    metadata = generate_random_dict(target_size_bytes=200)
+    detector = gl.create_detector(name=name, query=query, metadata=metadata)
+    assert detector.metadata == metadata
+
+    retrieved_detector = gl.get_detector(id=detector.id)
+    assert retrieved_detector.metadata == metadata
+
+
+@pytest.mark.skip_for_edge_endpoint(reason="The edge-endpoint does not support passing detector metadata.")
+def test_get_or_create_detector_with_metadata(gl: Groundlight):
+    unique_name = f"Unique name {datetime.utcnow()}"
+    query = "Is there a dog?"
+    metadata = generate_random_dict(target_size_bytes=200)
+    detector = gl.get_or_create_detector(name=unique_name, query=query, metadata=metadata)
+    assert detector.metadata == metadata
+
+    retrieved_detector = gl.get_or_create_detector(name=unique_name, query=query, metadata=metadata)
+    assert retrieved_detector.id == detector.id
+    assert retrieved_detector.metadata == metadata
+
+
+@pytest.mark.skip_for_edge_endpoint(reason="The edge-endpoint does not support passing detector metadata.")
+@pytest.mark.parametrize(
+    "metadata_list",
+    [
+        [generate_random_dict(target_size_bytes=3000)],
+        ["this is not valid JSON"],
+        [""],
+    ],
+)
+def test_create_detector_with_invalid_metadata(gl: Groundlight, metadata_list: Any):
+    name = f"Test {datetime.utcnow()}"  # Need a unique name
+    query = "Is there a dog?"
+
+    for metadata in metadata_list:
+        with pytest.raises((TypeError, ValueError)):
+            gl.create_detector(name=name, query=query, metadata=metadata)
+
+
+@pytest.mark.skip_for_edge_endpoint(reason="The edge-endpoint does not support passing image query metadata.")
 @pytest.mark.parametrize("metadata", [None, {}, {"a": 1}, '{"a": 1}'])
 def test_submit_image_query_with_metadata(
     gl: Groundlight, detector: Detector, image: str, metadata: Union[Dict, str, None]