Message that represents an arbitrary HTTP body. It should only be used for payload formats that can’t be represented as JSON, such as raw binary or an HTML page. This message can be used both in streaming and non-streaming API methods in the request as well as the response. It can be used as a top-level request field, which is convenient if one wants to extract parameters from either the URL or HTTP template into the request fields and also want access to the raw HTTP body. Example: message GetResourceRequest { // A unique request id. string request_id = 1; // The raw HTTP body is bound to this field. google.api.HttpBody http_body = 2; } service ResourceService { rpc GetResource(GetResourceRequest) returns (google.api.HttpBody); rpc UpdateResource(google.api.HttpBody) returns (google.protobuf.Empty); } Example with streaming methods: service CaldavService { rpc GetCalendar(stream google.api.HttpBody) returns (stream google.api.HttpBody); rpc UpdateCalendar(stream google.api.HttpBody) returns (stream google.api.HttpBody); } Use of this type only changes how the request and response bodies are handled, all other features will continue to work unchanged.
The current facet key (i.e. attribute config) maps into the merged_facet_key. A facet key can have at most one child. The current facet key and the merged facet key need both to be textual custom attributes or both numerical custom attributes (same type).
Replaces a set of textual facet values by the same (possibly different) merged facet value. Each facet value should appear at most once as a value per CatalogAttribute. This feature is available only for textual custom attributes.
Options to rerank based on facet values engaged by the user for the current key. That key needs to be a custom textual key and facetable. To use this control, you also need to pass all the facet keys engaged by the user in the request using the field [SearchRequest.FacetSpec]. In particular, if you don’t pass the facet keys engaged that you want to rerank on, this control won’t be effective. Moreover, to obtain better results, the facet values that you want to rerank on should be close to English (ideally made of words, underscores, and spaces).
Metadata that is used to define a condition that triggers an action. A valid condition must specify at least one of ‘query_terms’ or ‘products_filter’. If multiple fields are specified, the condition is met if all the fields are satisfied e.g. if a set of query terms and product_filter are set, then only items matching the product_filter for requests with a query matching the query terms wil get boosted.
This field specifies the current user answer during the conversational filtering search. This can be either user selected from suggested answers or user input plain text.
Suggested answers to the follow-up question. If it’s numerical attribute, only ProductAttributeInterval will be set. If it’s textual attribute, only productAttributeValue will be set.
The proposed refined search for intent-refinement/bundled shopping conversation. When using CONVERSATIONAL_FILTER_ONLY mode, the refined_query from search response will be populated here.
Product image. Recommendations AI and Retail Search use product images to improve prediction and search results. Product images can be returned in results, and are shown in prediction or search previews in the console. Please try to provide correct product images and avoid using images with size too small.
Metadata that describes the training and serving parameters of a Model. A Model can be associated with a ServingConfig and then queried through the Predict API.
Product attribute which structured by an attribute name and value. This structure is used in conversational search filters and answers. For example, if we have name=color and value=red, this means that the color is red.
A rule is a condition-action pair * A condition defines when a rule is to be triggered. * An action specifies what occurs on that trigger. Currently rules only work for controls with SOLUTION_TYPE_SEARCH.
Rule Condition: - No Condition.query_terms provided is a global match. - 1 or more Condition.query_terms provided are combined with OR operator. * Action Input: The request query and filter that are applied to the retrieved products, in addition to any filters already provided with the SearchRequest. The AND operator is used to combine the query’s existing filters with the filter rule(s). NOTE: May result in 0 results when filters conflict. * Action Result: Filters the returned objects to be ONLY those that passed the filter.
Force returns an attribute/facet in the request around a certain position or above. * Rule Condition: Must specify non-empty Condition.query_terms (for search only) or Condition.page_categories (for browse only), but can’t specify both. * Action Inputs: attribute name, position * Action Result: Will force return a facet key around a certain position or above if the condition is satisfied. Example: Suppose the query is “shoes”, the Condition.query_terms is “shoes”, the ForceReturnFacetAction.FacetPositionAdjustment.attribute_name is “size” and the ForceReturnFacetAction.FacetPositionAdjustment.position is 8. Two cases: a) The facet key “size” is not already in the top 8 slots, then the facet “size” will appear at a position close to 8. b) The facet key “size” in among the top 8 positions in the request, then it will stay at its current rank.
Maps a set of terms to a set of synonyms. Set of synonyms will be treated as synonyms of each query term only. query_terms will not be treated as synonyms of each other. Example: “sneakers” will use a synonym of “shoes”. “shoes” will not use a synonym of “sneakers”.
Pins one or more specified products to a specific position in the results. * Rule Condition: Must specify non-empty Condition.query_terms (for search only) or Condition.page_categories (for browse only), but can’t specify both. * Action Input: mapping of [pin_position, product_id] pairs (pin position uses 1-based indexing). * Action Result: Will pin products with matching ids to the position specified in the final result order. Example: Suppose the query is shoes, the Condition.query_terms is shoes and the pin_map has {1, "pid1"}, then product with pid1 will be pinned to the top position in the final results. If multiple PinActions are matched to a single request the actions will be processed from most to least recently updated. Pins to positions larger than the max allowed page size of 120 are not allowed.
Redirects a shopper to a specific page. * Rule Condition: Must specify Condition.query_terms. * Action Input: Request Query * Action Result: Redirects shopper to provided uri.
Removes an attribute/facet in the request if is present. * Rule Condition: Must specify non-empty Condition.query_terms (for search only) or Condition.page_categories (for browse only), but can’t specify both. * Action Input: attribute name * Action Result: Will remove the attribute (as a facet) from the request if it is present. Example: Suppose the query is “shoes”, the Condition.query_terms is “shoes” and the attribute name “size”, then facet key “size” will be removed from the request (if it is present).
Replaces a term in the query. Multiple replacement candidates can be specified. All query_terms will be replaced with the replacement term. Example: Replace “gShoe” with “google shoe”.
Creates a set of terms that will be treated as synonyms of each other. Example: synonyms of “sneakers” and “shoes”: * “sneakers” will use a synonym of “shoes”. * “shoes” will use a synonym of “sneakers”.
This field specifies the current user answer during the conversational search. This can be either user selected from suggested answers or user input plain text.
This field specifies the tile information including an attribute key, attribute value. More fields will be added in the future, eg: product id or product counts, etc.
A generic empty message that you can re-use to avoid defining duplicated empty messages in your APIs. A typical example is to use it as the request or the response type of an API method. For instance: service Foo { rpc Bar(google.protobuf.Empty) returns (google.protobuf.Empty); }
The Status type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by gRPC. Each Status message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the API Design Guide.
Represents a whole or partial calendar date, such as a birthday. The time of day and time zone are either specified elsewhere or are insignificant. The date is relative to the Gregorian Calendar. This can represent one of the following: * A full date, with non-zero year, month, and day values. * A month and day, with a zero year (for example, an anniversary). * A year on its own, with a zero month and a zero day. * A year and month, with a zero day (for example, a credit card expiration date). Related types: * google.type.TimeOfDay * google.type.DateTime * google.protobuf.Timestamp
Replaces the specified CatalogAttribute in the AttributesConfig by updating the catalog attribute with the same CatalogAttribute.key. If the CatalogAttribute to replace does not exist, a NOT_FOUND error is returned.
Gets the latest state of a long-running operation. Clients can use this method to poll the operation result at intervals as recommended by the API service.
We recommend that you use the ProductService.AddLocalInventories method instead of the ProductService.AddFulfillmentPlaces method. ProductService.AddLocalInventories achieves the same results but provides more fine-grained control over ingesting local inventory data. Incrementally adds place IDs to Product.fulfillment_info.place_ids. This process is asynchronous and does not require the Product to exist before updating fulfillment information. If the request is valid, the update will be enqueued and processed downstream. As a consequence, when a response is returned, the added place IDs are not immediately manifested in the Product queried by ProductService.GetProduct or ProductService.ListProducts. The returned Operations will be obsolete after 1 day, and GetOperation API will return NOT_FOUND afterwards. If conflicting updates are issued, the Operations associated with the stale updates will not be marked as done until being obsolete.
Updates local inventory information for a Product at a list of places, while respecting the last update timestamps of each inventory field. This process is asynchronous and does not require the Product to exist before updating inventory information. If the request is valid, the update will be enqueued and processed downstream. As a consequence, when a response is returned, updates are not immediately manifested in the Product queried by ProductService.GetProduct or ProductService.ListProducts. Local inventory information can only be modified using this method. ProductService.CreateProduct and ProductService.UpdateProduct has no effect on local inventories. The returned Operations will be obsolete after 1 day, and GetOperation API will return NOT_FOUND afterwards. If conflicting updates are issued, the Operations associated with the stale updates will not be marked as done until being obsolete.
Bulk import of multiple Products. Request processing may be synchronous. Non-existing items are created. Note that it is possible for a subset of the Products to be successfully updated.
Permanently deletes all selected Products under a branch. This process is asynchronous. If the request is valid, the removal will be enqueued and processed offline. Depending on the number of Products, this operation could take hours to complete. Before the operation completes, some Products may still be returned by ProductService.GetProduct or ProductService.ListProducts. Depending on the number of Products, this operation could take hours to complete. To get a sample of Products that would be deleted, set PurgeProductsRequest.force to false.
We recommend that you use the ProductService.RemoveLocalInventories method instead of the ProductService.RemoveFulfillmentPlaces method. ProductService.RemoveLocalInventories achieves the same results but provides more fine-grained control over ingesting local inventory data. Incrementally removes place IDs from a Product.fulfillment_info.place_ids. This process is asynchronous and does not require the Product to exist before updating fulfillment information. If the request is valid, the update will be enqueued and processed downstream. As a consequence, when a response is returned, the removed place IDs are not immediately manifested in the Product queried by ProductService.GetProduct or ProductService.ListProducts. The returned Operations will be obsolete after 1 day, and GetOperation API will return NOT_FOUND afterwards. If conflicting updates are issued, the Operations associated with the stale updates will not be marked as done until being obsolete.
Remove local inventory information for a Product at a list of places at a removal timestamp. This process is asynchronous. If the request is valid, the removal will be enqueued and processed downstream. As a consequence, when a response is returned, removals are not immediately manifested in the Product queried by ProductService.GetProduct or ProductService.ListProducts. Local inventory information can only be removed using this method. ProductService.CreateProduct and ProductService.UpdateProduct has no effect on local inventories. The returned Operations will be obsolete after 1 day, and GetOperation API will return NOT_FOUND afterwards. If conflicting updates are issued, the Operations associated with the stale updates will not be marked as done until being obsolete.
Updates inventory information for a Product while respecting the last update timestamps of each inventory field. This process is asynchronous and does not require the Product to exist before updating fulfillment information. If the request is valid, the update is enqueued and processed downstream. As a consequence, when a response is returned, updates are not immediately manifested in the Product queried by ProductService.GetProduct or ProductService.ListProducts. When inventory is updated with ProductService.CreateProduct and ProductService.UpdateProduct, the specified inventory field value(s) overwrite any existing value(s) while ignoring the last update time for this field. Furthermore, the last update times for the specified inventory fields are overwritten by the times of the ProductService.CreateProduct or ProductService.UpdateProduct request. If no inventory fields are set in CreateProductRequest.product, then any pre-existing inventory information for this product is used. If no inventory fields are set in SetInventoryRequest.set_mask, then any existing inventory information is preserved. Pre-existing inventory information can only be updated with ProductService.SetInventory, ProductService.AddFulfillmentPlaces, and ProductService.RemoveFulfillmentPlaces. The returned Operations is obsolete after one day, and the GetOperation API returns NOT_FOUND afterwards. If conflicting updates are issued, the Operations associated with the stale updates are not marked as done until they are obsolete.
Completes the specified prefix with keyword suggestions. This feature is only available for users who have Retail Search enabled. Enable Retail Search on Cloud Console before using this feature.
Bulk import of processed completion dataset. Request processing is asynchronous. Partial updating is not supported. The operation is successfully finished only after the imported suggestions are indexed successfully and ready for serving. The process takes hours. This feature is only available for users who have Retail Search enabled. Enable Retail Search on Cloud Console before using this feature.
Updates a Control. Control cannot be set to a different oneof field, if so an INVALID_ARGUMENT is returned. If the Control to update does not exist, a NOT_FOUND error is returned.
Update of model metadata. Only fields that currently can be updated are: filtering_option and periodic_tuning_state. If other values are provided, this API method ignores them.
Gets the latest state of a long-running operation. Clients can use this method to poll the operation result at intervals as recommended by the API service.
Performs a search. This feature is only available for users who have Retail Search enabled. Enable Retail Search on Cloud Console before using this feature.
Enables a Control on the specified ServingConfig. The control is added in the last position of the list of controls it belongs to (e.g. if it’s a facet spec control it will be applied in the last position of servingConfig.facetSpecIds) Returns a ALREADY_EXISTS error if the control has already been applied. Returns a FAILED_PRECONDITION error if the addition could exceed maximum number of control allowed for that type of control.
Disables a Control on the specified ServingConfig. The control is removed from the ServingConfig. Returns a NOT_FOUND error if the Control is not enabled for the ServingConfig.
Performs a search. This feature is only available for users who have Retail Search enabled. Enable Retail Search on Cloud Console before using this feature.
Set a specified branch id as default branch. API methods such as SearchService.Search, ProductService.GetProduct, ProductService.ListProducts will treat requests using “default_branch” to the actual branch id set as default. For example, if projects/*/locations/*/catalogs/*/branches/1 is set as default, setting SearchRequest.branch to projects/*/locations/*/catalogs/*/branches/default_branch is equivalent to setting SearchRequest.branch to projects/*/locations/*/catalogs/*/branches/1. Using multiple branches can be useful when developers would like to have a staging branch to test and verify for future usage. When it becomes ready, developers switch on the staging branch using this API while keeping using projects/*/locations/*/catalogs/*/branches/default_branch as SearchRequest.branch to route the traffic to this staging branch. CAUTION: If you have live predict/search traffic, switching the default branch could potentially cause outages if the ID space of the new branch is very different from the old one. More specifically: * PredictionService will only return product IDs from branch {newBranch}. * SearchService will only return product IDs from branch {newBranch} (if branch is not explicitly set). * UserEventService will only join events with products from branch {newBranch}.
Updates the AttributesConfig. The catalog attributes in the request will be updated in the catalog, or inserted if they do not exist. Existing catalog attributes not included in the request will remain unchanged. Attributes that are assigned to products, but do not exist at the catalog level, are always included in the response. The product attribute is assigned default values for missing catalog attribute fields, e.g., searchable and dynamic facetable options.
Writes a single user event from the browser. For larger user event payload over 16 KB, the POST method should be used instead, otherwise a 400 Bad Request error is returned. This method is used only by the Retail API JavaScript pixel and Google Tag Manager. Users should not call this method directly.
Bulk import of User events. Request processing might be synchronous. Events that already exist are skipped. Use this method for backfilling historical user events. Operation.response is of type ImportResponse. Note that it is possible for a subset of the items to be successfully inserted. Operation.metadata is of type ImportMetadata.
Deletes permanently all user events specified by the filter provided. Depending on the number of events specified by the filter, this operation could take hours or days to complete. To test a filter, use the list command first.
Starts a user-event rejoin operation with latest product catalog. Events are not annotated with detailed product information for products that are missing from the catalog when the user event is ingested. These events are stored as unjoined events with limited usage on training and serving. You can use this method to start a join operation on specified events with the latest version of product catalog. You can also use this method to correct events joined with the wrong product catalog. A rejoin operation can take hours or days to complete.
Gets the latest state of a long-running operation. Clients can use this method to poll the operation result at intervals as recommended by the API service.
Gets the latest state of a long-running operation. Clients can use this method to poll the operation result at intervals as recommended by the API service.