Search streaming video warehouse data using the command line

After you have created a Vision Warehouse, added it to an app, and deployed the app, you can search the data stored in the streaming video warehouse.

Search streaming video metadata

To search the data (assets) in your warehouse (corpus), populate the SearchAssetsRequest with the content that you would like to find. This content comes in a few different formats:

  1. criteria - Text, number, or date content provided by the user.
  2. facet_selections - Text content returned by the server, and selected by the user.
  3. content_time_ranges - Date ranges that all returned content must fall in.

In the following example, consider a Warehouse that contains security camera footage from different types of stores across the country. To retrieve all assets for the years 2018 or 2020 tagged with the annotation "state": "California", or the annotation "state":"Pennsylvania", send the following request:

REST

To search assets, send a POST request by using the projects.locations.corpora.searchAssets method.

In this sample body the criteria field use textArray values to provide two txt_values: "California" and "Pennsylvania". You can also provide search criteria for other data types. You can only specify one type of search criteria in each request.

Additional search criteria options

Integer ranges (inclusive)

    "int_range_array" : {
      "int_ranges": { "start": "5", "end": "10" }
      "int_ranges": { "start": "20", "end": "30" }
    }

Float ranges (inclusive)

    "float_range_array" : {
      "float_ranges": { "start": "2.6", "end": "14.3" }
      "float_ranges": { "start": "205.3", "end": "205.8" }
    }

Geolocations (coordinate and radius)

    "geo_location_array": {
      "circle_areas": {
        "latitude": "37.4221",
        "longitude": "122.0841",
        "radius_meter": "500"
      },
      "circle_areas": {
        "latitude": "12.46523",
        "longitude": "-95.2146",
        "radius_meter": "100"
      }
    }

Booleans

    "bool_value" : {
      "value": "true"
    }

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
  "page_size": "2",
  "content_time_ranges": {
    "date_time_ranges": {
      "start": {
        "year":"2018",
        "month":"1",
        "day":"1",
      },
      "end": {
        "year":"2019",
        "month":"1",
        "day":"1",
      }
    },
    "date_time_ranges": {
      "start": {
        "year":"2020",
        "month":"1",
        "day":"1",
      },
      "end": {
        "year":"2021",
        "month":"1",
        "day":"1",
      }
    }
  },
  "criteria": {
    "field": "state",
    "text_array": {
      "txt_values": "California",
      "txt_values": "Pennsylvania"
    }
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

To retrieve the next page of results, pass the original request parameters appended with the returned next_page_token.

The facet_results array shows content that matched the original query. The response above indicates that one of the security cameras is stationed at a sporting goods store, while the other one is stationed at a grocery store.

To restrict this query to only show the grocery store footage, pass back the same request with a facet selection.

Request JSON body with facet selection:

{
  "page_size": "2",
  "content_time_ranges": {
    "date_time_ranges": {
      "start": {
        "year":"2018",
        "month":"1",
        "day":"1",
      },
      "end": {
        "year":"2018",
        "month":"12",
        "day":"31",
      }
    },
    "date_time_ranges": {
      "start": {
        "year":"2020",
        "month":"1",
        "day":"1",
      },
      "end": {
        "year":"2020",
        "month":"12",
        "day":"31",
      }
    }
  },
  "criteria": {
    "field": "state",
    "text_array": {
      "txt_values": "California",
      "txt_values": "Pennsylvania"
    }
  },
  "facet_selections": {
    "facetId": "state",
    "displayName": "State",
    "buckets": {
        "value": {
         "stringValue": "California"
        }
    },
    "buckets": {
      "value": {
        "stringValue": "Pennsylvania"
      }
    },
    "bucketType": "FACET_BUCKET_TYPE_VALUE"
  },
  "facet_selections": {
    "facetId": "store-type",
    "displayName": "StoreType",
    "buckets": {
      "value": {
        "stringValue": "Sporting Goods"
      }
    },
    "buckets": {
      "value": {
        "stringValue": "Grocery"
      },
      "selected": "true"
    },
    "bucketType": "FACET_BUCKET_TYPE_VALUE"
  }
}

Because the Grocery facet is selected, any response will contain the annotation "store-type":"Grocery".

Return clip asset metadata when searching

The Vertex AI Vision API also allows users to specify extra clip metadata to return with the search result, using result_annotation_keys.

REST

In this example the user provided annotation key "camera-location" is specified in the request body, and the key's value ("Sunnyvale") is provided in the response.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
   "page_size": "2",
   "criteria": {
     "field": "state",
     "text_array": {
       "txt_values": "California",
       "txt_values": "Pennsylvania"
     }
   },
   "result_annotation_keys": "camera-location"
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

Use criteria to return asset metadata in search

You can specify in search criteria whether to return the matched annotations for each search result item. This feature is supported for limited data schema types: INTEGER, FLOAT, BOOLEAN, STRING (EXACT_SEARCH only), and the annotation must be on partition-level.

Assume you create the following data schema in a warehouse corpus:

{
  "key": "image-classification",
  "schema_details": {
    "type":"STRING",
    "granularity":"GRANULARITY_PARTITION_LEVEL",
    "search_strategy": {
      "search_strategy_type":"EXACT_SEARCH"
    }
  }
}

Some annotations for "image-classification" are ingested to the corpus using streaming video ingestion or a CreateAnnotation request.

After there are annotations ingested, you can search for "image-classification" and get video results and their corresponding annotations:

REST

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
  "page_size": "5",
  "facet_selections": {
    "facet_id": "image-classification",
    "fetch_matched_annotations": "true",
    "bucket_type": "FACET_BUCKET_TYPE_VALUE",
      "buckets": {
        "value": {
          "string_value": "cat"
        },
        "selected" : "true"
      },
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

Global search provides a place for users to enter search queries, instead of specifying individual criteria. You can search against string-type criteria set to be searchable in its data schema. The matching results are retrieved and returned to you.

To use this feature, set the search_query field in the SearchAssetsRequest:

REST

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
    "page_size": "2",
    "search_query': "Pennsylvania"
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

Apply sort spec to order returned-asset metadata in search

You can use the sorting feature to sort search results by user provided annotation. This can be useful for sorting results with data schema types that can be ordered, such as string and numeric types.

To use this feature, specify schema_key_sorting_strategy, which requires at least a data schema key and ascending/descending order:

REST

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
    "page_size": "2",
    "schemaKeySortingStrategy":
    { "options": 
      { 
        "data_schema_key": "stream-display-name", 
        "sort_decreasing": true
      } 
    }

}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content

You should receive a successful status code (2xx) and an empty response.

Create search configurations

Vision Warehouse allows users to customize their search experience through search configuration. Search Configuration uses video data - such as user-provided annotations and insights generated by Google Cloud video understanding models - to provide additional search options to the user. For example, if you want to target clips with specific color vehicles from car video data in your warehouse, you can use a specific search configuration for your query.

You can use a SearchConfig to set more granular configuration options.

The following example shows you how to create a SearchConfig resource.

General guidelines

For all use cases your request must meet the following conditions to execute successfully:

  1. Request.search_configuration.name must not already exist.
  2. The mapped_fields array must not be empty, and must map to existing user-given annotation keys.
  3. All mapped_fields must be of the same type.
  4. All mapped_fields must share exact/smart match config.
  5. All mapped_fields must share the same granularity.

There are several use cases for creating a SearchConfig, each with distinct guidelines you must follow.

Create a search config with custom search criteria

This section describes how to map a custom operator to one or more user-given annotation keys. In this case you need to satisfy the general guidelines when building your request.

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

The user-given annotation keys in this example are "player", "coach", and "cheerleader".

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is person.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=person

Request JSON body: 

{
   "search_criteria_property": {
     "mapped_fields": "player",
     "mapped_fields": "coach",
     "mapped_fields": "cheerleader",
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=person"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=person" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

{
  "name": "projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/person",
  "searchCriteriaProperty":
    {
      "mappedFields": [
        "player",
        "coach",
        "cheerleader"
      ]
    }
}

Create a search config with 1:1 facet mapping

To create a facet for a single user-given annotation key, you must ensure that Request.search_configuration.facet_property.mapped_fields contains a single element. This element's value must be a user-given annotation key name.

The following example shows you how to create a facet mapping for the user-given annotation key "Location".

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

In this example, the request succeeds because the search_config_id (Location) in the request URL references an existing user-given annotation key, and mapped_fields contains exactly one element with a value equal to search_config_id (Location).

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is Location.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location

Request JSON body: 

{
   "facet_property": {
     "mapped_fields": "Location",
     "display_name": "Location",
     "result_size": "5",
     "bucket_type":"FACET_BUCKET_TYPE_VALUE"
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location" | Select-Object -Expand Content
 

The following requests fail due to not meeting the necessary requirements.

Failed requests

Failed request 1:

      curl -X POST \
      -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
      -H "Content-Type: application/json; charset=utf-8" \
      https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location \
      -d "{
         "facet_property": {
           "mapped_fields": "City", /* City is not equal to search_config_id. */
           "display_name": "City",
           "result_size": "5",
           "bucket_type":"FACET_BUCKET_TYPE_VALUE"
        }
      }"

Failed request 2:

      curl -X POST \
      -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
      -H "Content-Type: application/json; charset=utf-8" \
      https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=City \
      -d "{
         "facet_property": {
           "mapped_fields": "City", /* City doesn't map to an existing user-given annotation key. */
           "display_name": "City",
           "result_size": "5",
           "bucket_type":"FACET_BUCKET_TYPE_VALUE"
        }
      }"

Failed request 3:

      curl -X POST \
      -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
      -H "Content-Type: application/json; charset=utf-8" \
      https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location \
      -d "{
         "facet_property": {
           "mapped_fields": "Location",
           "mapped_fields": "City", /* mapped_fields contains more than 1 element. */
           "display_name": "Location",
           "result_size": "5",
           "bucket_type":"FACET_BUCKET_TYPE_VALUE"
        }
      }"

Create a search config with a custom 1:1 or more facet mapping

Clients that want to create a mapping between a custom facet value and one or more user-given annotation keys must ensure that:

  1. Request.search_configuration must contain a SearchCriteriaProperty such that Request.search_configuration.search_criteria_property.mapped_fields contains the same elements as Request.search_configuration.facet_property.mapped_fields.

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

The following example shows you how to create a facet mapping for the user-given annotation keys "City" and "State".

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is Location.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location

Request JSON body: 

{
  "search_criteria_property": {
     "mapped_fields": "City",
     "mapped_fields": "State",
     "mapped_fields": "Province",
  }
  "facet_property": {
     "mapped_fields": "City",
     "mapped_fields": "State",
     "display_name": "Province",
     "result_size": "5",
     "bucket_type":"FACET_BUCKET_TYPE_VALUE"
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location" | Select-Object -Expand Content
 

The following requests fail due to not meeting the necessary requirements.

Failed requests

Failed request 1:

      curl -X POST \
      -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
      -H "Content-Type: application/json; charset=utf-8" \
      https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location \
      -d "{
        "facet_property": { /* Request is missing a SearchCriteriaProperty object.*/
           "mapped_fields": "City",
           "mapped_fields": "State",
           "display_name": "Location",
           "result_size": "5",
           "bucket_type":"FACET_BUCKET_TYPE_VALUE"
        }
      }"

Failed request 2:

      curl -X POST \
      -H "Authorization: Bearer "$(gcloud auth application-default print-access-token) \
      -H "Content-Type: application/json; charset=utf-8" \
      https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=Location \
      -d "{
        "search_criteria_property": {
           "mapped_fields": "City",
           "mapped_fields": "State",
        }
        "facet_property": {
           "mapped_fields": "City",
           "mapped_fields": "State",
           "mapped_fields": "Province", /* Province is missing from search_criteria_property. */
           "display_name": "Location",
           "result_size": "5",
           "bucket_type":"FACET_BUCKET_TYPE_VALUE"
        }
      }"

Create a search config with range based facets

Range facets are similar to normal facets, but each facet bucket covers some continuous span. An additional configuration (range_facet_config) gives the system information about these facet bucket ranges.

Range facets are available for:

  1. Integers
  2. Dates

There are three types of range facets:

  1. Fixed range - Each bucket is the same size.
  2. Custom range - Programmable bucket sizes. For example, logarithmic.
  3. Date range - Fixed bucket granularities of DAY, MONTH, and YEAR. This only applies to date range facets.

The same conditions apply as singular facets, with some additional validation regarding the range specification.

Fixed range bucket specification

The following example creates a fixed range facet spec for the field inventory-count, and results in the buckets: [-inf, 0), [0, 10), [10, 20), [20, 30), [30, inf].

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

This request uses the FixedRangeBucketSpec to create multiple buckets with equal granularities.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is inventory-count.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=inventory-count

Request JSON body: 

{
  "facet_property": {
    "mapped_fields": "inventory-count",
    "display_name": "Inventory Count",
    "result_size": "5",
    "bucket_type":"FACET_BUCKET_TYPE_FIXED_RANGE",
    "fixed_range_bucket_spec": {
       "bucket_start": {
         "integer_value": 0
       },
       "bucket_granularity": {
         "integer_value": 10
       },
       "bucket_count": 5
    }
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=inventory-count"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=inventory-count" | Select-Object -Expand Content
 

Custom range bucket specification

The following example creates a fixed range facet spec for the field video-views, and results in the buckets: [inf, 0), [0, 10), [10, 100), [100, 1000), [1000, 10000), [10000, inf).

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

This request uses the CustomRangeBucketSpec to specify how values are bucketized.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is video-views.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=video-views

Request JSON body: 

{
  "facet_property": {
    "mapped_fields": "video-views",
    "display_name": "Video Views",
    "result_size": "6",
    "bucket_type":"FACET_BUCKET_TYPE_CUSTOM_RANGE",
    "custom_range_bucket_spec": {
       "endpoints": {
         "integer_value": 0
       },
       "endpoints": {
         "integer_value": 10
       },
       "endpoints": {
         "integer_value": 100
       },
       "endpoints": {
         "integer_value": 1000
       },
       "endpoints": {
         "integer_value": 10000
       }
    }
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=video-views"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=video-views" | Select-Object -Expand Content
 

Date / time range bucket specification

The following example creates a date range spec for the field film-date with DAY granularity.

REST

You must specify your new SearchConfig ID at the end of the request URL, not as a field in the request.

This request uses the DateTimeBucketSpec to specify how date values are bucketized.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG: The name of your target SearchConfig.
    • The SearchConfig in this example is film-date.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=film-date

Request JSON body: 

{
  "facet_property": {
    "mapped_fields": "film-date",
    "display_name": "Film Date",
    "result_size": "5",
    "bucket_type": "FACET_BUCKET_TYPE_DATETIME",
    "datetime_bucket_spec": {
       "granularity": "DAY"
    }
 }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=film-date"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs?search_config_id=film-date" | Select-Object -Expand Content
 

After you create these facet buckets you can use them to search the warehouse.

REST

This request uses facetSelections objects to specify a group of facet buckets.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT_NUMBER: Your Google Cloud project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.

HTTP method and URL: 

POST https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets

Request JSON body: 

{
   "page_size": "10",
   "facet_selections": {
     "facet_id": "inventory-count",
     "buckets": {
       "range": {
         "end" : {
           "integer_value": 0
         }
       }
     },
     "buckets": {
       "range": {
         "start" : {
           "integer_value": 20
         },
         "end" : {
           "integer_value": 30
         }
       }
     }
   },
   "facet_selections": {
     "facet_id": "video-views",
     "buckets": {
       "range": {
         "start" : {
           "integer_value": 100
         },
         "end" : {
           "integer_value": 1000
         }
       }
     }
   },
   "facet_selections": {
     "facet_id": "film-date",
     "buckets": {
       "range": {
         "start" : {
           "datetime_value": {
             "year": 2022,
             "month": 9,
             "day": 10
           }
         },
         "end" : {
           "datetime_value": {
             "year": 2022,
             "month": 9,
             "day": 11
           }
         }
       }
     }
   }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID:searchAssets" | Select-Object -Expand Content
 

Update a search configuration

To update the current SearchConfig, your request must fulfill the following requirements:

  1. Request.searchConfig.name must already exist.
  2. Request must contain at least one non-empty searchCriteriaProperty or facetProperty.
  3. The mappedFields array must not be empty, and must map to existing user-given annotation keys.
  4. All mappedFields must be of the same type.
  5. All mappedFields must share the same granularity.
  6. All mappedFields must share the same semantic SearchConfig match options.

REST & CMD LINE

The following code sample updates a warehouse search configuration resource using the projects.locations.corpora.searchConfigs.patch method.

Before using any of the request data, make the following replacements:

  • REGIONALIZED_ENDPOINT: Endpoint might include a prefix matching the LOCATION_ID such as europe-west4-. See more about regionalized endpoints.
  • PROJECT: Your Google Cloud project ID or project number.
  • LOCATION_ID: The region where you are using Vertex AI Vision. For example: us-central1, europe-west4. See available regions.
  • CORPUS_ID: The ID of your target corpus.
  • SEARCHCONFIG_ID: The ID of your target SearchConfig.
  • "mappedFields": One or more existing user-given annotation keys.

HTTP method and URL: 

PATCH https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/SEARCHCONFIG_ID

Request JSON body: 

{
  "name": "projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/SEARCHCONFIG_ID1",
  "searchCriteriaProperty": {
    "mappedFields": "dataschema2"
  }
}

To send your request, choose one of these options:

curl

Save the request body in a file named request.json, and execute the following command: 

curl -X PATCH \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/SEARCHCONFIG_ID"

PowerShell

Save the request body in a file named request.json, and execute the following command: 

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method PATCH `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://warehouse-visionai.googleapis.com/v1/projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/SEARCHCONFIG_ID" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

{
  "name": "projects/PROJECT_NUMBER/locations/LOCATION_ID/corpora/CORPUS_ID/searchConfigs/SEARCHCONFIG_ID1",
  "searchCriteriaProperty": {
    "mappedFields": [
      "dataschema2"
    ]
  }
}