devoreur2code
Create a blog with NextJS

Create a blog with NextJS

Fabien Schlegel's photo
Fabien Schlegel
·

10 min read

This year I decided to migrate my blog from Gatsby to NextJS. To be able to add features more easily.

I use ChakraUI as a design system, Google Analytics for metrics, Netlify to host it and Typescript as language.

Create the project

In the first version of the blog, I wrote posts in markdown. I must keep it. As a developer, it’s easier than maintaining a WordPress (or other) back-end.

NextJS

Creating a NextJS project, it’s straightforward, one command is enough. You add your project name, go to the project folder and start the dev environment.

npx create-next-app@latest --typescript

The project contains classic files of a typescript project. And some important files for NextJS too.

  • next.config.js: the file used for the configuration
  • the pages folder: contains an index file for the root page
  • the _app.tsx file: inside the pages folder, it contains the component used to initialize pages.
  • the public folder: used for all static files, like favicon and images.

In NextJS, each folder or file inside the pages folder is a route. You can follow all the routes by unfolding the tree.

Useful packages

When I code, I like to use valuable packages.

  • Prettier: to format my code. In Visual Studio Code, I use it with autosave.
  • Husky: Prettier formats all supported staged files before the commit.
  • Eslint: it gives a bunch of rules for linting your Typescript and Javascript files.
  • Stylelint: same as Eslint, but for CSS and SCSS files.
  • Markdownlint: another linter, for the markdown files.
  • Sass: to support SCSS files.

You may think that’s too much. Using linters and code formatting tools are time savers.

When your project grows in size, you will encounter bugs. Consuming mental load to check semicolons and code indentation is a waste of time. Forward these things to tools and keep your time to write new features and fix bugs.

All these packages need a configuration file.

These configurations come from my experience on different projects. They can be outdated if you read this post long after its release.

Structure of my app

For my React projects, I use a structure inside the src folder. It’s to dispatch files according to their purpose. Naming can change depending on the project.

  • a generic folder: for everything you can use in each project without changes.
  • a business folder: for all the features of the project. Each folder inside it is a feature.
  • a core folder: for the rest of it. Configuration files, specific or shared services.

This is my first NextJS project and not having an src folder is disorienting. But adapting is part of the developer's skill set. I define a new structure. As each first time, it can't be the best one, but I’ll improve it later.

  • the pages and the public folder: it’s used by the framework, so I must keep them.
  • the components folder: each folder inside it contains a component and all I need for it.
  • the hooks folder: contains all the custom hooks used by the components.
  • the core folder: all I need for the app, like API, utilities, the custom theme for Chakra, etc.
  • the posts folder: contains all the markdown files

If you have a few components, like me at the beginning of the project, don’t over-engineer your project. You don’t need plenty of empty folders and files.

Implement ChakraUI

To add ChakraUI to the project, I use this command, as it’s explained in the documentation.

yarn add @chakra-ui/react @emotion/react @emotion/styled framer-motion

To use ChakraUI, you must add a component called ChakraProvider. This component has a theme property, to override the ChakraUI default theme. You must add it inside the pages folder, in the _app.tsx file.

I create a custom theme file to extend colours and add fonts.

import { extendTheme } from "@chakra-ui/react";

const mainTheme = extendTheme({
  colors: {
    brand: {
      darkBlue: "#1f4f6f",
      blue: "#22577a",
      greenBlue: "#38a3a5",
      greener: "#57cc99",
      green: "#80ED99",
      lightGreen: "#c7f9cc",
    },
  },
  fonts: {
    heading: `'Raleway', sans-serif`,
    body: `'Libre Baskerville', sans-serif`,
  },
});

export default mainTheme;

It's my first bug. I use SCSS import to fetch the google fonts. NextJS V12 use SWC as a minifier tool. There is a bug with this version and import doesn't work in production.

To resolve this, I do a Google search and find the solution here.

I use the special file _document.tsx. The build updates the render of the html and body tags in this file.

I add here all the scripts recommended by the Google fonts website. And it works.

import Document, { Html, Head, Main, NextScript } from "next/document";

export default class MyDocument extends Document {
  render(): JSX.Element {
    return (
      <Html>
        <Head>
          {/* Google fonts */}
          <link rel="preconnect" href="https://fonts.googleapis.com" />
          <link rel="preconnect" href="https://fonts.gstatic.com" />
          <link
            href="https://fonts.googleapis.com/css2?family=Libre+Baskerville&family=Raleway:wght@800&display=swap"
            rel="stylesheet"
          />
        </Head>
        <body>
          <Main />
          <NextScript />
        </body>
      </Html>
    );
  }
}

CSS in JS

ChakraUI uses CSS properties as props for components. In the beginning, I use SCSS first and props after.

It’s a mistake. I have chosen a design system to gain time. Even though I lost time during the learning curve I will retrieve it afterwards. So I must exploit it at 100%. I will delete most styling files and try to keep only the SCSS file for the posts.

CSS in JS is easy to use. Instead of using CSS classes, you add the property to the component. In CSS, I write properties in kebab-case (lowercase and separated with hyphens). A React property can’t contain a hyphen. Instead, they’re in camel-case. Chakra gives us some shortcuts too, like justify for justifyContent.

Sometimes we need media queries. It’s for responsive design like the flexbox direction property above. Chakra gives us two solutions: an array syntax and an object syntax. Using the object syntax is clearer. Keys are the different breakpoints.

<Flex
  width="100%"
  padding={{ base: "0.5em", lg: "6em" }}
  align="center"
  justify={"space-evenly"}
  backgroundColor="brand.darkBlue"
  minHeight="80vh"
  direction={{ base: "column-reverse", lg: "row" }}
>

From markdown to the post

The first version of the blog used Gatsby and a starter (like a template). At build, the engine uses GraphQL requests for posts. I haven’t had the time to understand how Gatsby uses GraphQL. So I left it out.

NextJS uses functions to read the posts and render them as props. To help, NextJS gives us a starter. I use it to understand how to proceed with markdown files.

First, we must execute this command to install the packages we need.

yarn add prismjs remark remark-html remark-prism gray-matter @types/remark-prism

It’s time to talk about getStaticProps and getStaticPath. It’s 2 special functions of NextJS.

getStaticProps

It’s used for static site generation. This function run during the build, to generate static pages. The unique parameter is context.

I use it to get the last part of the URL named slug. I give it to functions which return the current post and some data about the next and the previous post.

getStaticProps return an object used by the component to render the posts page.

export async function getStaticProps({ params }: Params) {
  const post = getPostBySlug(params.slug, [
    "title",
    "date",
    "slug",
    "author",
    "content",
    "coverImage",
    "ogImage",
    "tags",
  ]);
  const content = await markdownToHtml(post.content || "");

  const previous = getPreviousPost(params.slug);

  const next = getNextPost(params.slug);

  return {
    props: {
      post: {
        ...post,
        content,
      },
      previous,
      next,
    },
  };
}

getStaticPaths

It’s used for static site generation when you use dynamic routes. It builds all the static paths available during the build.

I use it to retrieve all the posts slug and get all the paths of my posts during the build.

export async function getStaticPaths() {
  const posts = getAllPostsByDate(["slug"]);

  return {
    paths: posts.map((post) => {
      return {
        params: {
          slug: post.slug,
        },
      };
    }),
    fallback: false,
  };
}

The posts API

This is the main part of the process. How do we get a JSON object with the post’s data from a markdown file?

The getPostBySlug function will take as a parameter a string and an array. It represents the name of the file without extension and a list of fields. It comes from the last part of the URL. The function reads the file from the posts folder. The gray-matter library divides the data between the content and the metadata. The function returns an object to getStaticProps. Fields and content are the keys of this object.

export function getPostBySlug(slug: string, fields: string[] = []) {
  const realSlug = slug.replace(/\.md$/, "");
  const fullPath = join(postsDirectory, `${realSlug}.md`);
  const fileContents = fs.readFileSync(fullPath, "utf8");
  const { data, content } = matter(fileContents);

  type Items = {
    [key: string]: string;
  };

  const items: Items = {};

  fields.forEach((field) => {
    if (field === "slug") {
      items[field] = realSlug;
    }
    if (field === "content") {
      items[field] = content;
    }

    if (typeof data[field] !== "undefined") {
      items[field] = data[field];
    }
  });

  return items;
}

The markdownToHtml function uses the remark library. It transforms the value of the content field into text with HTML tags.

The last library I use there is PrismJS. This library highlight code parts of the post for better understanding.

export default async function markdownToHtml(markdown: string) {
  const result = await remark()
    .use(html, { sanitize: false })
    .use(prism)
    .process(markdown);
  return result.toString();
}

Create a list of posts from tags

Each of my posts contains tags. You click on a tag and you’re redirected to a dedicated page with filtered posts.

This is a 2 steps feature. Remember the getStaticPaths, it will create all the paths from the list of tags. We use it to get the list of all tags and create all the needed paths during the build. All is in this function.

export function getAllTags(): Array<string> {
  const allPosts = getAllPosts(["slug", "tags"]);

  const flattenTags = allPosts.map((post) => post?.tags).flat();

  const allTags = flattenTags.filter(
    (item, pos) => flattenTags.indexOf(item) == pos
  );
  return allTags;
}

Let me explain. Each post contains an array of tags. I use a map function to retrieve them. The flat function concatenate my array of arrays. So in flattenTags, I have an array of tags with duplicates. In the next function, I use a filter to remove all duplicates.

export function getPostsByTag(tag: string, fields: string[] = []) {
  return getAllPostsByDate(fields).filter((post) => post.tags.includes(tag));
}

The second part retrieves the list of posts sorted by date. It keeps the posts if the array of tags includes the requested tag.

This API needs some improvements, like lowercase tags.

Add Metrics

I like to have metrics on sites. I can know which content is most popular, the best size of the post, etc.

I use Google Analytics on the first version of the site with a Gatsby plugin. With NextJS, I’m unable to keep the same tool. After some Google research, I find what I’m looking for.

{
  isProduction && (
    <>
      <script
        async
        src={`https://www.googletagmanager.com/gtag/js?id=${GA_TRACKING_ID}`}
      />
      <script
        // eslint-disable-next-line react/no-danger
        dangerouslySetInnerHTML={{
          __html: `
            window.dataLayer = window.dataLayer || [];
            function gtag(){dataLayer.push(arguments);}
            gtag('js', new Date());
            gtag('config', '${GA_TRACKING_ID}', {
              page_path: window.location.pathname,
            });
          `,
        }}
      />
    </>
  );
}

This snippet of code inside the _document.tsx file loads the script with my tracking ID. The isProduction boolean enable it if NODE_ENV variable is set to production.

Add Sitemaps

Each website must have sitemaps and robots.txt files.

It helps search engines to index your site. I add the next-sitemap package to perform this. It runs after the build to collect all routes and create dedicated files. I add the package command the postbuild command in the package.json file and a configuration file.

After the deployment, sitemaps and robots.txt files will be available.

/** @type {import('next-sitemap').IConfig} */
module.exports = {
  siteUrl: process.env.NEXT_PUBLIC_SITE_URL || "https://devoreur2code.com",
  generateRobotsTxt: true,
};

Deployment

The previous version is already in production. I want to deploy the new version without creating duplicate content.

It is easy with Netlify. I create a new app in the dashboard and link it to my repository in 2 clicks. Netlify detects my app as a NextJS project and set everything for me I remove the domain in the first version. I add it to my new site in domain settings.

I deploy the new version in less than 10 minutes.

Conclusion

I learn a lot of things by doing this rewriting of my website. I did some things wrong with my lack of experience with NextJS. But working as a developer is so interesting. You can learn by practice and progress.

Adding new features will help me to find bugs and improvements. I decide to open source the repository. It’s available on Gitlab and Github.

You can follow the account of the site and mine on Twitter. I’ll post amazing news about next features.

See you soon!

ReactTypeScriptNext.jsBlogging