A Shopify merchant needs an automated way to tag products being added to their Shopify store inventory. They source hundreds of products weekly from various dropshippers and upload the unstructured data to Shopify programmatically. Because the data is unstructured, Shopify is unable to power the merchant's storefront search. While the merchant can add tags inside the Shopify Admin, the experience of doing this on hundreds of products weekly is time-consuming.
To solve this, the merchant wants to build a custom Shopify app on Gadget that will run every new product description through an automated tagging script.
In this example, we'll show you how to build a custom product tagging app that listens to the product/create and product/update webhooks from Shopify, runs product descriptions through a tagging script, and sets tags back in Shopify.
To get the most out of this tutorial, you will need:
- a Shopify Partners account
- a development store
- at least one product in your store that has a product description
You can fork this Gadget project and try it out yourself.
You will still need to set up the Shopify Connection after forking. Continue reading if you want to learn how to connect Gadget to a Shopify store!
Our first step will be to set up a Gadget project and connect our backend to a Shopify store via the Shopify connection. Create a new Gadget application at gadget.new and select the Shopify app template.
Because we are adding an embedded frontend, we are going to build an app using the Partners connection.
Both the Shopify store Admin and the Shopify Partner Dashboard have an Apps section. Ensure that you are on the Shopify Partner Dashboard before continuing.
Now we get to select what Shopify scopes we give our application access to, while also picking what Shopify data models we want to import into our Gadget app.
Now we want to connect our Gadget app to our custom app in the Partners dashboard.
At this point, Gadget copies the selected Shopify models, their types, validations and associations into your Gadget backend. These models are ready to process webhooks as soon as you install the app on a Shopify store.
We opted to install this app on our development store, using the Test on Development Store link on our Shopify Partners dashboard. Gadget handles OAuth for us, all we need to do is grant permission to the API scopes and the connection is made.
The next step is to create a model that will store our list of vetted keywords that we can use to power our tagging script. These keywords can be different types of products or brands. Make sure to add keywords that will be found in your products' descriptions!
Gadget instantly creates and documents a GraphQL CRUD API for this model. Test it out in the API Playground!
Using the API Playground, we can make a Create call to our allowedTag model to store a new keyword.
We can run the same mutation again with a different keyword value to store additional keywords.
We can also check to make sure our tag keywords have been saved. Go to the Data page for the allowedTag model and you should be able to see an entry for any tags that have been added.
Gadget keeps your app and store in sync by generating a CRUD (Create, Read, Update, Delete) API around each of your cloned models and wiring up each of the API actions to their corresponding Shopify webhook. If the Shopify store fires a `products/create` webhook, Gadget will run your Create action on the Product model. By default, this action uses the incoming params from the webhook to create a record in your database. Similarly, if Shopify fires the `products/update` webhook, Gadget will run your Update action which updates the record with the incoming params.
What makes Actions special is that they can be completely customized. You can change what happens when the action runs by adding Code Effects that you define.
Now that we have keywords to check against, we can write our tagging script. Because we want this script to run every time a product record is created or updated, we'll add an Effect to the create and update actions on shopifyProduct:
That's not a lot of code!
This snippet will run on every incoming products/create webhook that is sent by Shopify, and determines if tags need to be added by cross-referencing the body of the incoming payload against the stored keyword records by making an internal API request to Gadget. Should any words match, they're sent back to Shopify as new tags for the product.
Gadget gives us a connections object as an argument to our effect function, which has an authenticated Shopify API client ready to go. We use this object to make API calls back to Shopify to update the tags and complete the process.
We also use Gadget's changed helper on our record to avoid entering an infinite loop. This looping can occur when a Shopify webhook triggers code that updates our Shopify store. Because we have added this change detection, we can use the same code for both the create and update actions.
Now this code will be run every time a product is created or updated in a Shopify store.
The record.changed helper is a special field that Gadget has included to help prevent an infinite loop when updating Shopify records.
When we call shopify.product.update(...) the product in our Shopify store will be updated. This update action will fire Shopify's products/update webhook. If we are using this webhook as a trigger for running custom code that updates a product, we will be stuck in an endless loop of updating our products and running our custom code.
We can use record.changed to determine if changes have been made to the key on this record and only run our code if changes have occurred.
For more info on change tracking in Gadget, refer to the documentation.
Right now the allowedTag model only has a single field, keyword. If you're building a public Shopify app, you also need to associate keywords with individual stores so that all the shops that install your app don't share keyword records.
It's also important to grant your embedded app permission to call the create and delete actions on the allowedTag model. Access to custom model APIs are always disabled by default for embedded app users. If you encounter a GGT_PERMISSION_DENIED error when building an embedded app, you probably need to go into the Roles and Permissions page and grant embedded app users access to your Gadget app API.
With this added Shop relationship, you will be able to track what keywords are used for individual shops. To automatically filter by the current shop when reading from your allowedTag model, you can add a Gelly snippet to enforce shop tenancy automatically.
This snippet selects all allowedTag records when the related shopId is equal to the shop id of the current session. The current session is managed for you when you connect to Shopify, and session records including the current session can be viewed on the session Data page in Gadget.
This handles shop tenancy for your read action, and will also be applied for update, delete, or custom model actions. For custom actions, it's advisable to leverage Gadget's preventCrossShopDataAccess helper which prevents the modification of the shop relationship on all model actions to which it is applied.
The relationship to the current shop is automatically set up by the preventCrossShopDataAccess helper in the run function. By using this helper, the shopId cannot be spoofed by malicious users.
The preventCrossShopDataAccess helper is also useful for other model actions, such as delete, to enforce data tenancy in public apps.
For delete actions (or update, or other custom actions), using preventCrossShopDataAccess in the run function verifies that the shopId associated with the current record matches the shopId for the current session. If these values do not match, an error is triggered, resulting in the failure of the action.
You also need to update how you read your allowedTag data in the shopifyProduct update and create actions. You must filter by the currentShopId to ensure that you are only reading the allowedTag records for the current shop, as Gelly tenancy is not applied when making API requests in your backend actions.
Now you have a multi-tenant backend. The final step is building an embedded frontend.
New Gadget apps include a frontend folder. When you set up a Shopify connection, Gadget automatically makes changes to this frontend folder by:
Additional packages have also been added to your package.json upon connecting to Shopify, including @gadgetinc/react which allows for the use of Gadget's handy React hooks for fetching data and calling your Gadget project's API, and @shopify/polaris which allows you to use Shopify's Polaris components out of the box when building embedded Shopify apps.
The entire tagger frontend code snippet is below. Additional details on some of Gadget's provided tooling are below the snippet.
If you go to your development app, you should now be able to test it out! Go back to the embedded frontend in your store admin and start adding custom tags. You'll be able to see the created tags, along with the related shop ID, in the allowedTag Data page in Gadget.
The above snippet has everything you need to build a frontend for your app. Let's take a closer look at how you read and write data using Gadget's React tooling.
Your Gadget API client is already set up for you in frontend/api.js! You can use this API client to fetch data from our models using the product tagger application's auto-generated API. You can also make use of some Gadget-provided React hooks that help to read (and write) using your app's API.
The useFindMany hook will run api.allowedTag.findMany() and return a response that has data, error, and fetching as properties. You can use fetching to display a Polaris Spinner while the data is being fetched, and the error property to display and handle any request errors.
Here is the code snippets from your tagger frontend that use the useFindMany hook:
Now you need to send a request to your Gadget app's backend API to create entered keywords in your app's database. The @gadgetinc/react package also has a useAction hook that will assist with making requests to any of your model actions. Similar to useFindMany, the useAction hook returns an object with data, fetching, and error properties. Additionally, useAction returns a function that needs to be called to actually run the action.
For the product tagger, you are using the useAction hook to call the api.allowedTag.create and api.allowedTag.delete actions:
Congrats! You have built a full-stack and fully functional embedded product tagger application! Now you can test it out.
Go back to the Connections page in Gadget and click Shop Installs and then Sync on the connected store if you set up a custom app through the Partners dashboard, or just click Sync if you used the store Admin to set up your app.
Gadget will fetch each of the records in Shopify and run them through your actions. Not only will this populate your Gadget backend with the store's inventory, but it will also run the effects we added, updating the tags for each synced product. Our tagging application will also run on products when they are added to the store, so any new products will also be tagged for us automatically.
Congratulations! In about 20 minutes you were able to build a custom app that updates tags in Shopify each time there is a match against the list of Allowed Tags.