Gatsby Lifecycle APIs
Gatsby provides a rich set of lifecycle APIs to hook into its bootstrap, build, and client runtime operations.
Gatsby’s design principles include:
- Conventions > code, but use low-level primitives to build conventions with code.
- Extracting logic and configuration into plugins should be trivial and encouraged.
- Plugins are easy to open source and reuse. They’re just npm packages.
High level Overview
The following model gives a conceptual overview of how data is sourced and transformed in the process of building a Gatsby site:
site:
title: Home
description: Gatsby tips
Content is often organized in systems like databases, content management systems, files, or external APIs.
Any source of data can be connected to Gatsby through plugins or using Gatsby’s APIs.
Bootstrap sequence
During the main bootstrap sequence, Gatsby (in this order):
- reads and validates
gatsby-config.js
to load in your list of plugins (it doesn’t run them yet). - deletes HTML and CSS files from previous builds (public folder)
- initializes its cache (stored in
/.cache
) and checks if any plugins have been updated since the last run, if so it deletes the cache - sets up
gatsby-browser
andgatsby-ssr
for plugins that have them - starts main bootstrap process
- runs onPreBootstrap. e.g. implemented by
gatsby-plugin-typography
- runs onPreBootstrap. e.g. implemented by
- runs sourceNodes e.g. implemented by
gatsby-source-wikipedia
- within this,
createNode
can be called multiple times, which then triggers onCreateNode
- within this,
- creates initial GraphQL schema
- runs resolvableExtensions which lets plugins register file types or extensions e.g.
gatsby-plugin-typescript
- runs createPages from the gatsby-node.js in the root directory of the project e.g. implemented by
page-hot-reloader
- within this,
createPage
can be called any number of times, which then triggers onCreatePage
- within this,
- runs createPagesStatefully
- runs source nodes again and updates the GraphQL schema to include pages this time
- runs onPreExtractQueries e.g. implemented by
gatsby-transformer-sharp
andgatsby-source-contentful
, and extracts queries from pages and components (StaticQuery
) - compiles GraphQL queries and creates the Abstract Syntax Tree (AST)
- runs query validation based on schema
- executes queries and stores their respective results
- writes page redirects (if any) to
.cache/redirects.json
- the onPostBootstrap lifecycle is executed
In development this is a running process powered by webpack and react-refresh
), so changes to any files get re-run through the sequence again, with smart cache invalidation. For example, gatsby-source-filesystem
watches files for changes, and each change triggers re-running queries. Other plugins may also perform this service. Queries are also watched, so if you modify a query, your development app is hot reloaded.
The core of the bootstrap process is the “api-runner”, which helps to execute APIs in sequence, with state managed in Redux. Gatsby exposes a number of lifecycle APIs which can either be implemented by you (or any of your configured plugins) in gatsby-node.js
, gatsby-browser.js
or gatsby-ssr.js
.
Build sequence
(to be written)
Client sequence
(to be written)
Please see the links along the left under “REFERENCE” for the full API documentation.