The GitHub GraphQL API is very well documented, and even though the focus of this tutorial isn’t learning and using GraphQL, it’s a great opportunity to fetch Carbon-related data for this Carbon tutorial.
To do so, we’ll be using Apollo Client, the front-end component of the Apollo Platform. Apollo provides several open source tools for using GraphQL throughout your application’s stack. Apollo Client is a sophisticated GraphQL client that manages data and state in an application.
A preview of what you will build (see repositories page):
This tutorial has an accompanying GitHub repository called carbon-tutorial-angular that we’ll use as a starting point for each step. If you haven’t forked and cloned that repository yet, and haven’t added the upstream remote, go ahead and do so by following the step 1 instructions.
With your repository all set up, let’s check out the branch for this tutorial step’s starting point.
$ git fetch upstream$ git checkout -b angular-step-3 upstream/angular-step-3
Install the app’s dependencies:
$ npm install
Then, start the app:
$ npm run start
You should see something similar to where the previous step left off. Stop your app with CTRL-C and let’s get everything installed.
We’ll shortcut this using the Angular CLI, if you’d like more information head over to Angular Apollo Installation for details.
Note: If you have not yet installed the Angular CLI, you will need to install the Angular CLI before running the Angular Apollo Installation.
Install the following
apollo-client - package containing everything you need to set up Apollo Clientgraphql - parses your GraphQL queriesapollo-angular - Apollo integration for AngularUsing the command:
$ ng add apollo-angular$ ng lint --fix
Note: In case you receive ERROR in ../node_modules/graphql/subscription/subscribe.d.ts:44:3 - error TS2304: Cannot find name 'AsyncIterableIterator'., you’ll have to add the code below in tsconfig.app.json lib:
"lib": ["es2016","dom","esnext.asynciterable"],
Note: In case you receive typings.d.ts:2:13 - error TS2403: Subsequent variable declarations must have the same type. Variable 'module' must be of type 'NodeModule', but here has type '{ id: string; }'., you’ll have to comment out or remove the code below in typings.d.ts:
declare var module: {id: string,};
You’ll need a personal access token from your GitHub account in order to make requests to the GitHub API. Check out this guide to see how to get one.
When you get to the scope/permissions step, you can leave them all unchecked. We don’t need any special permissions, we just need access to the public API.
Once you have your token, we need to put it in a place where we can use it in our app. When your application is being built and developed, our app will parse environmental variables in any environment file and make them available.
export const environment = {production: false,githubPersonalAccessToken: 'xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx',};
Go ahead and start your app with npm run start, or, if your app is running, you’ll need to restart it to get access to this token.
We need to import the environment at the top of every file that we decide to use it in:
import { environment } from '../environments/environment';
Next, make your client by providing a URI for the GitHub GraphQL API as well as an authorization header using the environmental variable you just added to environment.ts.
import { HttpHeaders } from '@angular/common/http';const uri = 'https://api.github.com/graphql'; // <-- add the URL of the GraphQL server hereexport function createApollo(httpLink: HttpLink) {return {link: httpLink.create({uri,headers: new HttpHeaders({Authorization: `Bearer ${environment.githubPersonalAccessToken}`,}),}),cache: new InMemoryCache(),};}
Add the following import at the top of repo-table.component.ts:
import { Apollo } from 'apollo-angular';import gql from 'graphql-tag';
Next we’ll assemble our GraphQL query to fetch only the data we need from the GraphQL API. We’ll do this using the gql helper we just imported. The gql helper lets you write GraphQL queries using interpolated strings (backticks) in JavaScript.
You can use GitHub’s explorer tool to write and test your own queries. Try copying the query below and experiment with changing the properties. You can also click the “Docs” button in the top right of the explorer to view all of the available data and query parameters.
If you’d like some more information regarding writing queries, we recommend Apollo’s documentation on this topic.
First, we will add a private apollo parameter of type Apollo to the constructor to get access to the Apollo service.
constructor(private apollo: Apollo) { }
Next, we will fetch the data in ngOnInit(). Add the folowing code right below the model.header declaration:
this.apollo.watchQuery({query: gql`query REPO_QUERY {# Let's use carbon as our organizationorganization(login: "carbon-design-system") {# We'll grab all the repositories in one go. To load more resources# continuously, see the advanced topics.repositories(first: 75, orderBy: { field: UPDATED_AT, direction: DESC }) {totalCountnodes {urlhomepageUrlissues(filterBy: { states: OPEN }) {totalCount}stargazers {totalCount}releases(first: 1) {totalCountnodes {name}}nameupdatedAtcreatedAtdescriptionid}}}}`})
Our last column in the data table will be a comma-separated list of repository and home page links, so let’s create a custom cell template.
The column will have two values (url and homepageUrl) and will have the structure of an unordered list. If the repository does not have a home page URL, only render the repository link.
<ng-template #linkTemplate let-data="data"><ul style="display: flex"><li><a ibmLink [href]="data.github">Github</a></li><li *ngIf="data.homepage"><span> | </span><a ibmLink [href]="data.homepage">HomePage</a></li></ul></ng-template>
@ViewChild("linkTemplate", null)protected linkTemplate: TemplateRef<any>;
Note: Don’t forget to import ViewChild and TemplateRef.
Now let’s create a function that transforms row data to our expected header keys. Notice how we’re using our new linkTemplate to generate the value of the links key in each row.
prepareData(data) {const newData = [];for (const datum of data) {newData.push([new TableItem({ data: datum.name, expandedData: datum.description }),new TableItem({ data: new Date(datum.createdAt).toLocaleDateString() }),new TableItem({ data: new Date(datum.updatedAt).toLocaleDateString() }),new TableItem({ data: datum.issues.totalCount }),new TableItem({ data: datum.stargazers.totalCount }),new TableItem({data: {github: datum.url,homepage: datum.homepageUrl},template: this.linkTemplate})]);}return newData;}
At this point, we should run our query and console.log() the results to verify that the request is working.
If there’s an issue, we’ll render the corresponding error message. We will also check when loading is true, but add the implementation in the following steps.
Finally, if neither of those are true, it means we have our data! One nice advantage of GraphQL is that as long as there are no errors, we can be certain the properties on the data we requested aren’t undefined.
At the end of the watchQuery in ngOnInit():
.valueChanges.subscribe((response: any) => {if (response.error) {const errorData = [];errorData.push([new TableItem({data: "error!" })]);this.model.data = errorData;} else if (response.loading) {// Add loading state} else {console.log(response);}});
The page will look the same as we’re still rendering our static example rows, but if you view your browser’s console (e.g. Chrome DevTools), you should see the response from GitHub!
Now that we have that data, let’s populate the data table. Replace console.log(response); with the following that destructures the organization object. Once we have the repositories, we can call our prepareData() to build the data table rows.
// If we're here, we've got our data!this.prepareData(response.data.organization.repositories.nodes);
Then, in ngOnInit() delete the static rows since we no longer need them.
At this point, the first time that you visit the repositories page, we’re querying the GitHub API and rendering the response through the DataTable component. We could stop here, but there’s more to be done! Let’s add the Table Skeleton.
To do so, we will need to modify repo-table.component.html and repo-table.component.ts:
<ibm-table[skeleton]="skeleton"[model]="skeleton ? skeletonModel : model"[showSelectionColumn]="false"[striped]="false"></ibm-table>
We need to tell the loading skeleton how many rows to render, so let’s use 10 skeleton rows to prepare for the next enhancement…
skeletonModel = Table.skeletonModel(10, 6);skeleton = true;
Note: Don’t forget to import Table.
Then replace the comment with:
else if (response.loading) {this.skeleton = true;}
and at the top of prepareData() add:
this.skeleton = false;
Pagination! Instead of rendering every repository, let’s add pagination to the data table to only render 10 at a time. Depending on your specific requirements, you may need to fetch new data each time that you interact with the pagination component, but for simplicity, we’re going to make one request to fetch all data, and then paginate the in-memory row data.
First we will create an array and populate it with 10 rows at a time.
data = [];
Next, replace the code where we call preapreData() with:
this.data = response.data.organization.repositories.nodes;this.model.pageLength = 10;this.model.totalDataLength = response.data.organization.repositories.totalCount;this.selectPage(1);
This initializes the page size to 10 as we also specified in our loading skeleton.
Next we will call selectPage(). This will initialize the offset and use prepareData() to only render the subset of rows for the current “page”.
selectPage(page) {const offset = this.model.pageLength * (page - 1);const pageRawData = this.data.slice(offset, offset + this.model.pageLength);this.model.data = this.prepareData(pageRawData);this.model.currentPage = page;}
Note: We only pass the rows that we want our table to display. We can do this by slicing our array of rows depending on the first item and the page size.
Finally, let’s import the LinkModule and PaginationModule to repositories.module.ts and repo-table.component.spec and add the template for pagination.
import { LinkModule, PaginationModule } from 'carbon-components-angular';
imports: [LinkModule, PaginationModule];
Immediately after the ibm-table closing tag (/>), add the ibm-pagination component.
<ibm-pagination[model]="model"(selectPage)="selectPage($event)"></ibm-pagination>
Note: Like the other Carbon Angular component, pagination component examples can be found in Storybook by browsing the story and knobs.
That does it! Your data table should fetch GitHub data on first render. You can expand each row to see the repository’s description. You can modify the pagination items per page and cycle through pages or jump to a specific page of repositories.
We’re going to submit a pull request to verify completion of this tutorial step.
Run the CI check to make sure we’re all set to submit a pull request.
$ npm run test
Note: Having issues running the CI check? Step 1 has troubleshooting notes that may help.
Before we can create a pull request, stage and commit all of your changes:
$ git add --all && git commit -m "feat(tutorial): complete step 3"
Then, push to your repository:
$ git push origin angular-step-3
Note: Having issues pushing your changes? Step 1 has troubleshooting notes that may help.
Finally, visit carbon-tutorial-angular to “Compare & pull request”. In doing so, make sure that you are comparing to angular-step-3 into base: angular-step-3.
Note: Expect your tutorial step PRs to be reviewed by the Carbon team but not merged. We’ll close your PR so we can keep the repository’s remote branches pristine and ready for the next person!