Skip to content
Atlas Platform
GitHubDiscordYouTube

Using GraphQL Index API

Index API Documentation

NOTE: This API and its endpoints are in BETA. Beta Services as described by Section 2.e. of the terms of services located here

Overview

The public API includes mutations:

  • index mutation for creating or updating a document
  • delete mutation for deleting a document
  • deleteAll mutation for deleting all documents
  • config mutation for configuring semantic search

To use a mutation, you need to authenticate.

Authentication

Clients calling the API mutations are required to add an authentication header with the valid authentication token.

Authorization: Bearer {ACCESS_TOKEN}

Data needed to consume API can be found in your WP Admin area in the WP Engine Smart Search tab under Settings.

WP Engine Smart Search Settings

Creating or Updating a Document

To create or update a document in the index using the mutation, you need to:

  1. Use the index mutation.
  2. Provide a DocumentInput object as a parameter.
  3. In the DocumentInput object, provide the necessary document data as a JSON object.
  4. To ensure a document has a unique identifier in the DocumentInput object, follow these guidelines:
    • The id field of the DocumentInput object should be used to provide a unique identifier for the document. This field is required for updating a document.
    • If an id is provided in the DocumentInput object, the system checks whether a document with that id already exists. If it does, the existing document is updated with the new data instead of creating a new document.
    • In cases where no id is provided in the DocumentInput object, the system automatically generates a unique identifier for the document. This ensures that each document has a unique identifier, even if one is not explicitly provided. By following these guidelines, you can effectively handle document creation and updates, ensuring uniqueness of identifiers and appropriate handling based on the presence of an id in the DocumentInput object.
    • The data field is required and contains all the index data.
  5. Optionally, you can include metadata by using the meta field. This allows you to provide additional information that will be logged on the server. Metadata can be helpful for analyzing and understanding the logs more effectively. Meta data includes optional fields:
    • system - should be the name of the client “Atlas Search v0.2.24”
    • action - the name of the action performed e. g. “manual-index”
    • source - hostname from where action has been performed e. g. “atlasce.wpengine.com”

GraphQL example to create a document

mutation CreateIndexDocument($input: DocumentInput!) {
  index(input: $input) {
    success
    code
    message
    document {
      id
      data
    }
  }
}

GraphQL Variables

{
  "input": {
    "data": {
      "ID": 1,
      "post_title": "Post Title",
      "post_excerpt": "Post Excerpt",
      "post_content": "Post Content",
      "post_date": "11-06-2023T12:33:00",
      "post_type": "post",
      "post_status": "publish"
    },
    "meta": {
      "system": "Atlas Search v0.2.24",
      "action": "manual-index",
      "source": "cms.themessenger.com"
    }
  }
}

Mutation Response

{
  "data": {
    "index": {
      "success": true,
      "code": "200",
      "message": "Document was indexed successfully",
      "document": {
        "id": "yMI9vogB0FEetE6QV9fh",
        "data": {
          "ID": 1,
          "post_title": "Post Title",
          "post_excerpt": "Post Excerpt",
          "post_content": "Post Content",
          "post_date": "11-06-2023T12:33:00",
          "post_type": "post",
          "post_status": "publish"
        }
      }
    }
  }
}

In this example, we’re creating a new document with fields: ID, post_title, post_excerpt, post_content, post_date, post_type, post_status. The mutation returns a DocumentMutationResponse object that includes fields:

  • success - a boolean says if operation is successful or not
  • code - string with number of response code (200 - is success)
  • message - human readable operation status
  • document with id and data.

To create a new document with a provided id or update an existing document, you can use the index mutation. The behavior depends on whether a document with the provided id already exists:

  • If a document with the provided id doesn’t exist, a new document will be created with the specified id.
  • If a document with the provided id already exists, the existing document will be updated with the new data.

To create a new document with a provided id or update an existing document, you need to provide the id of the document you want to update in the DocumentInput object.

GraphQL example to update a document

mutation UpdateIndexDocument($input: DocumentInput!) {
  index(input: $input) {
    success
    code
    message
    document {
      id
      data
    }
  }
}

GraphQL Variables

{
  "input": {
    "id": "yMI9vogB0FEetE6QV9fh",
    "data": {
      "post_title": "Post with updated title"
    },
    "meta": {
      "system": "Atlas Search v0.2.24",
      "action": "manual-index",
      "source": "cms.themessenger.com"
    }
  }
}

Mutation Response

{
  "data": {
    "index": {
      "success": true,
      "code": "200",
      "message": "Document was indexed successfully",
      "document": {
        "id": "yMI9vogB0FEetE6QV9fh",
        "data": {
          "post_title": "Post with updated title"
        }
      }
    }
  }
}

In this example, we’re updating an existing document with the id of yMI9vogB0FEetE6QV9fh. We are mutating post_title field of the document. The mutation returns a DocumentMutationResponse object.

The same behavior would occur even if the document did not already exist. In that case, it would be created with the provided id and the DocumentMutationResponse would include this id.

IMPORTANT Rest of the fields not specified in the input will be removed.
In this case it will remove ID, post_excerpt, post_content, post_date, post_type, post_status and only post_title will be added.

Indexing third party data

It is reccomended to have a data structure to tag third party data so that it can be filtered later if needed. Otherwise, by default, the search query considers every document in the index.

To index third party data, provision a WP environment with smart search. Then, by using the url and access_token found on the WP Smart Search settings page, configure the third party data client with these credentials and then calling the index API.

Once the application is running the headless app, use the find api to query for documents held in the index.

The only way to obtain a license at the moment is to purchase a plan that has the smart-search entitlement.

Code sample

The following is an example written in javascript which queries a “space” endpoint and indexes the first record via the WP Engine Smart Search API.

const spaceUrl = 
  "https://www.space.com/api/product_search?first_edition=true&ship_to_country=USA&minimum_price=50";
const spaceHeaders = {
  "X-API-KEY": "YOUR_SPACE_API_KEY",
};
const spacePayload = {
  method: "GET",
  headers: spaceHeaders,
};
const smartSearchEndpoint = "YOUR_SMART_SEARCH_URL";
const smartSearchHeaders = {
  "Content-Type": "application/json",
  Authorization: "Bearer YOUR_SMART_SEARCH_TOKEN",
};

// First lets fetch data from the space endpoint
fetch(spaceUrl, spacePayload)
.then((response) => response.json())
.then((data) => {
  // Here lets get the first document returned in the response
  const firstDocument = data.data[0];
  // lets image the data structure is the following
  // { "id": 12345, "title": "hello" }
  // lets create a custom id by prepending 'space:'
  const documentId = `space:${firstDocument.id}`;
  // assign our variables for the index request

  // add space as the root of the document
  const variables = {
  input: {
    id: documentId,
    data: {
      space: firstDocument
    },
  }
  };
  // This is the graphql mutation for indexing  documents
  // a link to the API documentation can be found at the top
  // of this document
  const mutation = `
    mutation CreateIndexDocument($input: DocumentInput!) {
      index(input: $input) {
        success
        code
        message
        document {
          id
          data
        }
      }
    }
  `;
  const payload = {
    method: "POST",
    headers: smartSearchHeaders,
    body: JSON.stringify({
    query: mutation,
    variables: variables,
    }),
  };
  // Send the GraphQL request which will index  the first document
  return fetch(smartSearchEndpoint, payload);
})
.then((response) => response.json())
.then((data) => console.log(data))
.catch((error) => console.error("Error:", error));
{
   "post_title": "Star Wars",
   "post_contents": "Star Wars conent",
   "post_type": "metadata"
}

Third Party Data Diagram

Sample query filtering for just metadata content

query FindSimpleQuery {
  find(
    query: "Star Wars"
    filter: "post_type:metadata"
    ) {
    total
    documents {
      id
      data
    }
  }
}

These query appendages could happen under the hood or just considering all content

query FindSimpleQuery {
  find(
    query: "Star Wars"
    filter: "post_type:post,metadata"
    ) {
    total
    documents {
      id
      data
    }
  }
}

Deleting a Document

To delete a document from the index, you need to:

  1. Use the delete mutation.
  2. Provide the id of the document you want to delete.
  3. Optionally provide metadata by using the meta field. This allows you to provide additional information that will be logged on the server. Including metadata can be helpful for analyzing and understanding the logs more effectively. Meta data includes optional fields:
    • system - should be the name of the client “Atlas Search v0.2.24”
    • action - the name of the action performed e. g. “manual-index”
    • source - hostname from where action has been performed e. g. “atlasce.wpengine.com”

GraphQL example to delete a document

mutation deleteDocument($id: ID!, $meta: MetaInput) {
  delete(id: $id, meta: $meta) {
    code
    success
    message
    document {
      id
      data
    }
  }
}

GraphQL Variables

{
  "id": "yMI9vogB0FEetE6QV9fh",
  "meta": {
    "system": "Atlas Search v0.2.24",
    "action": "manual-delete",
    "source": "cms.themessenger.com"
  }
}

GraphQL Response

{
  "data": {
    "delete": {
      "code": "200",
      "success": true,
      "message": "Document was deleted successfully",
      "document": {
        "id": "yMI9vogB0FEetE6QV9fh",
        "data": {}
      }
    }
  }
}

We are deleting a document with the id of yMI9vogB0FEetE6QV9fh.
The mutation returns a DocumentMutationResponse object that includes the id and data fields of the deleted document.

Deleting All Documents

To delete all documents from the index you need to:

  1. Use deleteAll mutation.
  2. Optionally provide metadata by using the meta field. This allows you to provide additional information that will be logged on the server. Including metadata can be helpful for analyzing and understanding the logs more effectively.

GraphQL Example to Delete All Documents

mutation deleteAllDocuments($meta: MetaInput) {
  deleteAll(meta: $meta) {
    code
    success
    message
  }
}

GraphQL Variables

{
  "meta": {
    "system": "Atlas Search v0.2.24",
    "action": "reset-data",
    "source": "cms.themessenger.com"
  }
}

GraphQL Response

{
  "data": {
    "deleteAll": {
      "code": "200",
      "success": true,
      "message": "Delete all documents was successful"
    }
  }
}