Add ask_vlm_verify in ExperimentalApi for cloud VLM alert verification

generated/.openapi-generator/FILES

@@ -76,6 +76,11 @@ docs/TextModeConfiguration.md
 docs/TextRecognitionResult.md
 docs/UserApi.md
 docs/VerbEnum.md
+docs/VerdictEnum.md
+docs/VlmVerification.md
+docs/VlmVerificationCost.md
+docs/VlmVerificationResult.md
+docs/VlmVerificationsApi.md
 docs/WebhookAction.md
 docs/WebhookActionRequest.md
 git_push.sh
@@ -92,6 +97,7 @@ groundlight_openapi_client/api/month_to_date_account_info_api.py
 groundlight_openapi_client/api/notes_api.py
 groundlight_openapi_client/api/priming_groups_api.py
 groundlight_openapi_client/api/user_api.py
+groundlight_openapi_client/api/vlm_verifications_api.py
 groundlight_openapi_client/api_client.py
 groundlight_openapi_client/apis/__init__.py
 groundlight_openapi_client/configuration.py
@@ -162,6 +168,10 @@ groundlight_openapi_client/model/status_enum.py
 groundlight_openapi_client/model/text_mode_configuration.py
 groundlight_openapi_client/model/text_recognition_result.py
 groundlight_openapi_client/model/verb_enum.py
+groundlight_openapi_client/model/verdict_enum.py
+groundlight_openapi_client/model/vlm_verification.py
+groundlight_openapi_client/model/vlm_verification_cost.py
+groundlight_openapi_client/model/vlm_verification_result.py
 groundlight_openapi_client/model/webhook_action.py
 groundlight_openapi_client/model/webhook_action_request.py
 groundlight_openapi_client/model_utils.py

generated/README.md

@@ -146,6 +146,7 @@ Class | Method | HTTP request | Description
 *PrimingGroupsApi* | [**get_priming_group**](docs/PrimingGroupsApi.md#get_priming_group) | **GET** /v1/priming-groups/{id} | 
 *PrimingGroupsApi* | [**list_priming_groups**](docs/PrimingGroupsApi.md#list_priming_groups) | **GET** /v1/priming-groups | 
 *UserApi* | [**who_am_i**](docs/UserApi.md#who_am_i) | **GET** /v1/me | 
+*VlmVerificationsApi* | [**submit_vlm_verification**](docs/VlmVerificationsApi.md#submit_vlm_verification) | **POST** /v1/vlm-verifications | 
 
 
 ## Documentation For Models
@@ -215,6 +216,10 @@ Class | Method | HTTP request | Description
  - [TextModeConfiguration](docs/TextModeConfiguration.md)
  - [TextRecognitionResult](docs/TextRecognitionResult.md)
  - [VerbEnum](docs/VerbEnum.md)
+ - [VerdictEnum](docs/VerdictEnum.md)
+ - [VlmVerification](docs/VlmVerification.md)
+ - [VlmVerificationCost](docs/VlmVerificationCost.md)
+ - [VlmVerificationResult](docs/VlmVerificationResult.md)
  - [WebhookAction](docs/WebhookAction.md)
  - [WebhookActionRequest](docs/WebhookActionRequest.md)
 

generated/docs/VerdictEnum.md

@@ -0,0 +1,12 @@
+# VerdictEnum
+
+* `YES` - YES * `NO` - NO * `UNSURE` - UNSURE
+
+## Properties
+Name | Type | Description | Notes
+------------ | ------------- | ------------- | -------------
+**value** | **str** | * `YES` - YES * `NO` - NO * `UNSURE` - UNSURE |  must be one of ["YES", "NO", "UNSURE", ]
+
+[[Back to Model list]](../README.md#documentation-for-models) [[Back to API list]](../README.md#documentation-for-api-endpoints) [[Back to README]](../README.md)
+
+

generated/docs/VlmVerification.md

@@ -0,0 +1,19 @@
+# VlmVerification
+
+Response shape for POST /v1/vlm-verifications.
+
+## Properties
+Name | Type | Description | Notes
+------------ | ------------- | ------------- | -------------
+**id** | **str** |  | [readonly] 
+**type** | **str** |  | [readonly] 
+**created_at** | **datetime** |  | [readonly] 
+**query** | **str** |  | 
+**model_id** | **str** |  | 
+**result** | [**VlmVerificationResult**](VlmVerificationResult.md) |  | 
+**cost** | [**VlmVerificationCost**](VlmVerificationCost.md) |  | 
+**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)
+
+

generated/docs/VlmVerificationCost.md

@@ -0,0 +1,14 @@
+# VlmVerificationCost
+
+
+## Properties
+Name | Type | Description | Notes
+------------ | ------------- | ------------- | -------------
+**input_tokens** | **int, none_type** |  | 
+**output_tokens** | **int, none_type** |  | 
+**total_cost_usd** | **float, none_type** |  | 
+**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)
+
+

generated/docs/VlmVerificationResult.md

@@ -0,0 +1,14 @@
+# VlmVerificationResult
+
+
+## Properties
+Name | Type | Description | Notes
+------------ | ------------- | ------------- | -------------
+**verdict** | [**VerdictEnum**](VerdictEnum.md) |  | 
+**confidence** | **float** |  | 
+**reasoning** | **str** |  | 
+**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)
+
+

generated/docs/VlmVerificationsApi.md

@@ -0,0 +1,100 @@
+# groundlight_openapi_client.VlmVerificationsApi
+
+All URIs are relative to *https://api.groundlight.ai/device-api*
+
+Method | HTTP request | Description
+------------- | ------------- | -------------
+[**submit_vlm_verification**](VlmVerificationsApi.md#submit_vlm_verification) | **POST** /v1/vlm-verifications | 
+
+
+# **submit_vlm_verification**
+> VlmVerification submit_vlm_verification(media, query)
+
+
+
+ Submit one or more images for VLM-based alert verification.  Send everything as `multipart/form-data`: one to eight `media` parts, plus a `query` field and an optional `model_id` field.  The `query` describes what each image is and what to look for — the server makes no assumptions about the images' meaning. Images are presented to the model labeled `Image 1`, `Image 2`, ... in upload order, so the query can reference them (e.g. \"Image 1 is the full frame; image 2 is the cropped ROI ...\").  (Video parts are planned but not yet supported and are rejected.)  Requires `ENABLE_BEDROCK_VLM_ACCESS` (enabled for Standard_Internal and SciDuck accounts) and accepted terms of service.  ```bash curl https://api.groundlight.ai/device-api/v1/vlm-verifications \\     -F \"media=@full_frame.jpg;type=image/jpeg\" \\     -F \"media=@roi.jpg;type=image/jpeg\" \\     -F \"query=Image 1 is the full camera frame; image 2 is the cropped region a detector flagged. Is there really a fire?\" \\     -F \"model_id=gpt-5.4\" ``` 
+
+### Example
+
+* Api Key Authentication (ApiToken):
+
+```python
+import time
+import groundlight_openapi_client
+from groundlight_openapi_client.api import vlm_verifications_api
+from groundlight_openapi_client.model.vlm_verification import VlmVerification
+from pprint import pprint
+# Defining the host is optional and defaults to https://api.groundlight.ai/device-api
+# See configuration.py for a list of all supported configuration parameters.
+configuration = groundlight_openapi_client.Configuration(
+    host = "https://api.groundlight.ai/device-api"
+)
+
+# The client must configure the authentication and authorization parameters
+# in accordance with the API server security policy.
+# Examples for each auth method are provided below, use the example that
+# satisfies your auth use case.
+
+# Configure API key authorization: ApiToken
+configuration.api_key['ApiToken'] = 'YOUR_API_KEY'
+
+# Uncomment below to setup prefix (e.g. Bearer) for API key, if needed
+# configuration.api_key_prefix['ApiToken'] = 'Bearer'
+
+# Enter a context with an instance of the API client
+with groundlight_openapi_client.ApiClient(configuration) as api_client:
+    # Create an instance of the API class
+    api_instance = vlm_verifications_api.VlmVerificationsApi(api_client)
+    media = [
+        open('/path/to/file', 'rb'),
+    ] # [file_type] | One or more images (common formats: JPEG, PNG, WEBP). Video is not yet supported.
+    query = "query_example" # str | Natural-language prompt describing the media and what to verify.
+    model_id = "model_id_example" # str | Friendly model alias (e.g. 'gpt-5.4', 'claude-sonnet-4.5'). Defaults to the server default. (optional)
+
+    # example passing only required values which don't have defaults set
+    try:
+        api_response = api_instance.submit_vlm_verification(media, query)
+        pprint(api_response)
+    except groundlight_openapi_client.ApiException as e:
+        print("Exception when calling VlmVerificationsApi->submit_vlm_verification: %s\n" % e)
+
+    # example passing only required values which don't have defaults set
+    # and optional values
+    try:
+        api_response = api_instance.submit_vlm_verification(media, query, model_id=model_id)
+        pprint(api_response)
+    except groundlight_openapi_client.ApiException as e:
+        print("Exception when calling VlmVerificationsApi->submit_vlm_verification: %s\n" % e)
+```
+
+
+### Parameters
+
+Name | Type | Description  | Notes
+------------- | ------------- | ------------- | -------------
+ **media** | **[file_type]**| One or more images (common formats: JPEG, PNG, WEBP). Video is not yet supported. |
+ **query** | **str**| Natural-language prompt describing the media and what to verify. |
+ **model_id** | **str**| Friendly model alias (e.g. 'gpt-5.4', 'claude-sonnet-4.5'). Defaults to the server default. | [optional]
+
+### Return type
+
+[**VlmVerification**](VlmVerification.md)
+
+### Authorization
+
+[ApiToken](../README.md#ApiToken)
+
+### HTTP request headers
+
+ - **Content-Type**: multipart/form-data
+ - **Accept**: application/json
+
+
+### HTTP response details
+
+| Status code | Description | Response headers |
+|-------------|-------------|------------------|
+**201** |  |  -  |
+
+[[Back to top]](#) [[Back to API list]](../README.md#documentation-for-api-endpoints) [[Back to Model list]](../README.md#documentation-for-models) [[Back to README]](../README.md)
+

generated/groundlight_openapi_client/api/vlm_verifications_api.py

@@ -0,0 +1,153 @@
+"""
+    Groundlight API
+
+    Groundlight makes it simple to understand images. You can easily create computer vision detectors just by describing what you want to know using natural language.  # noqa: E501
+
+    The version of the OpenAPI document: 0.18.2
+    Contact: support@groundlight.ai
+    Generated by: https://openapi-generator.tech
+"""
+
+import re  # noqa: F401
+import sys  # noqa: F401
+
+from groundlight_openapi_client.api_client import ApiClient, Endpoint as _Endpoint
+from groundlight_openapi_client.model_utils import (  # noqa: F401
+    check_allowed_values,
+    check_validations,
+    date,
+    datetime,
+    file_type,
+    none_type,
+    validate_and_convert_types,
+)
+from groundlight_openapi_client.model.vlm_verification import VlmVerification
+
+
+class VlmVerificationsApi(object):
+    """NOTE: This class is auto generated by OpenAPI Generator
+    Ref: https://openapi-generator.tech
+
+    Do not edit the class manually.
+    """
+
+    def __init__(self, api_client=None):
+        if api_client is None:
+            api_client = ApiClient()
+        self.api_client = api_client
+        self.submit_vlm_verification_endpoint = _Endpoint(
+            settings={
+                "response_type": (VlmVerification,),
+                "auth": ["ApiToken"],
+                "endpoint_path": "/v1/vlm-verifications",
+                "operation_id": "submit_vlm_verification",
+                "http_method": "POST",
+                "servers": None,
+            },
+            params_map={
+                "all": [
+                    "media",
+                    "query",
+                    "model_id",
+                ],
+                "required": [
+                    "media",
+                    "query",
+                ],
+                "nullable": [],
+                "enum": [],
+                "validation": [
+                    "media",
+                ],
+            },
+            root_map={
+                "validations": {
+                    ("media",): {
+                        "max_items": 8,
+                        "min_items": 1,
+                    },
+                },
+                "allowed_values": {},
+                "openapi_types": {
+                    "media": ([file_type],),
+                    "query": (str,),
+                    "model_id": (str,),
+                },
+                "attribute_map": {
+                    "media": "media",
+                    "query": "query",
+                    "model_id": "model_id",
+                },
+                "location_map": {
+                    "media": "form",
+                    "query": "form",
+                    "model_id": "form",
+                },
+                "collection_format_map": {
+                    "media": "csv",
+                },
+            },
+            headers_map={"accept": ["application/json"], "content_type": ["multipart/form-data"]},
+            api_client=api_client,
+        )
+
+    def submit_vlm_verification(self, media, query, **kwargs):
+        """submit_vlm_verification  # noqa: E501
+
+         Submit one or more images for VLM-based alert verification.  Send everything as `multipart/form-data`: one to eight `media` parts, plus a `query` field and an optional `model_id` field.  The `query` describes what each image is and what to look for — the server makes no assumptions about the images' meaning. Images are presented to the model labeled `Image 1`, `Image 2`, ... in upload order, so the query can reference them (e.g. \"Image 1 is the full frame; image 2 is the cropped ROI ...\").  (Video parts are planned but not yet supported and are rejected.)  Requires `ENABLE_BEDROCK_VLM_ACCESS` (enabled for Standard_Internal and SciDuck accounts) and accepted terms of service.  ```bash curl https://api.groundlight.ai/device-api/v1/vlm-verifications \\     -F \"media=@full_frame.jpg;type=image/jpeg\" \\     -F \"media=@roi.jpg;type=image/jpeg\" \\     -F \"query=Image 1 is the full camera frame; image 2 is the cropped region a detector flagged. Is there really a fire?\" \\     -F \"model_id=gpt-5.4\" ```   # noqa: E501
+        This method makes a synchronous HTTP request by default. To make an
+        asynchronous HTTP request, please pass async_req=True
+
+        >>> thread = api.submit_vlm_verification(media, query, async_req=True)
+        >>> result = thread.get()
+
+        Args:
+            media ([file_type]): One or more images (common formats: JPEG, PNG, WEBP). Video is not yet supported.
+            query (str): Natural-language prompt describing the media and what to verify.
+
+        Keyword Args:
+            model_id (str): Friendly model alias (e.g. 'gpt-5.4', 'claude-sonnet-4.5'). Defaults to the server default.. [optional]
+            _return_http_data_only (bool): response data without head status
+                code and headers. Default is True.
+            _preload_content (bool): if False, the urllib3.HTTPResponse object
+                will be returned without reading/decoding response data.
+                Default is True.
+            _request_timeout (int/float/tuple): timeout setting for this request. If
+                one number provided, it will be total request timeout. It can also
+                be a pair (tuple) of (connection, read) timeouts.
+                Default is None.
+            _check_input_type (bool): specifies if type checking
+                should be done one the data sent to the server.
+                Default is True.
+            _check_return_type (bool): specifies if type checking
+                should be done one the data received from the server.
+                Default is True.
+            _spec_property_naming (bool): True if the variable names in the input data
+                are serialized names, as specified in the OpenAPI document.
+                False if the variable names in the input data
+                are pythonic names, e.g. snake case (default)
+            _content_type (str/None): force body content-type.
+                Default is None and content-type will be predicted by allowed
+                content-types and body.
+            _host_index (int/None): specifies the index of the server
+                that we want to use.
+                Default is read from the configuration.
+            async_req (bool): execute request asynchronously
+
+        Returns:
+            VlmVerification
+                If the method is called asynchronously, returns the request
+                thread.
+        """
+        kwargs["async_req"] = kwargs.get("async_req", False)
+        kwargs["_return_http_data_only"] = kwargs.get("_return_http_data_only", True)
+        kwargs["_preload_content"] = kwargs.get("_preload_content", True)
+        kwargs["_request_timeout"] = kwargs.get("_request_timeout", None)
+        kwargs["_check_input_type"] = kwargs.get("_check_input_type", True)
+        kwargs["_check_return_type"] = kwargs.get("_check_return_type", True)
+        kwargs["_spec_property_naming"] = kwargs.get("_spec_property_naming", False)
+        kwargs["_content_type"] = kwargs.get("_content_type")
+        kwargs["_host_index"] = kwargs.get("_host_index")
+        kwargs["media"] = media
+        kwargs["query"] = query
+        return self.submit_vlm_verification_endpoint.call_with_http_info(**kwargs)

generated/groundlight_openapi_client/apis/__init__.py

@@ -24,3 +24,4 @@
 from groundlight_openapi_client.api.notes_api import NotesApi
 from groundlight_openapi_client.api.priming_groups_api import PrimingGroupsApi
 from groundlight_openapi_client.api.user_api import UserApi
+from groundlight_openapi_client.api.vlm_verifications_api import VlmVerificationsApi