remote data -> static templates
Static is the best, but sometimes you need to fetch data from a remote source which makes things not so static. Spike Records is a little webpack plugin intended for use with spike which allows you to make data pulled from a file or url available in your view templates.
It can pull data from the following places:
- A javascript object
- A file containing a javascript object or JSON
- A URL that returns JSON
- A GraphQL endpoint
Install into your project with npm i spike-records -S
.
Then load it up as a plugin in app.js
like this:
const Records = require('spike-records')
const standard = require('reshape-standard')
const locals = {}
module.exports = {
reshape: standard({ locals: () => locals }),
plugins: [new Records({
addDataTo: locals,
test: { file: 'data.json' }
})]
}
The primary use case for spike-records is to inject local variables into your html templates, although technically it can be used for anything. In the example above, we use the reshape-standard plugin pack to add variables (among other functionality) to your html. Spike's default template also uses reshape-standard
.
In order to use the results from spike-records, you must pass it an object, which it will put the resolved data on, using the addDataTo
key. This plugin runs early in spike's compile process, so by the time templates are being compiled, the object will have all the data necessary on it. If you are using the data with other plugins, ensure that spike-records is the first plugin in the array.
I know this is an unusual pattern for a javascript library, but the way it works is quite effective in this particular system, and affords a lot of flexibility and power.
The records plugin accepts an object, and each key in the object (other than addDataTo
) should contain another object as it's value, with either a file
, url
, data
, graphql
or callback
property. For example:
const Records = require('spike-records')
const locals = {}
function myFunc () {
// call any arbitrary API or computation to produce data
return new Promise((resolve, reject) => {...})
}
module.exports = {
plugins: [new Records({
addDataTo: locals,
one: { file: 'data.json' },
two: { url: 'http://api.carrotcreative.com/staff' },
three: { data: { foo: 'bar' } },
four: {
graphql: {
url: 'http://localhost:1234',
query: 'query { allPosts { title } }',
variables: 'xxx', // optional
headers: { authorization: 'Bearer xxx' } // optional
}
},
five: { callback: myFunc }
})]
}
Whatever data source you provide, it will be resolved and added to your view templates as [key]
. So for example, if you were trying to access three
in your templates, you could access it with three.foo
, and it would return 'bar'
, as such:
p {{ three.foo }}
Now let's get into some more details for each of the data types.
file
accepts a file path, either absolute or relative to your spike project's root. So for the example above, it would resolve to /path/to/project/data.json
.
url
accepts either a string or an object. If provided with a string, it will make a request to the provided url and parse the result as JSON, then return it as a local. If you need to modify this behavior, you can pass in an object instead. The object is passed through directly to rest.js, their docs for acceptable values for this object can be found here. These options allow modification of the method, headers, params, request entity, etc. and should cover any additional needs.
The most straightforward of the options, this will just pass the data right through to the locals. Also if you provide a A+ compliant promise for a value, it will be resolved and passed in to the template.
Alongside any of the data sources above, there are a few additional options you can provide in order to further manipulate the output.
If you want to transform the data from your source in any way before injecting it as a local, you can use this option. For example:
const Records = require('spike-records')
const locals = {}
module.exports = {
plugins: [new Records({
addDataTo: locals,
blog: {
url: 'http://blog.com/api/posts',
transform: (data) => { return data.response.posts }
}
})]
}
If you want to transform the data from your source before it is processed by rest, for instance to remove cross site scripting protections, you can use this option. For example:
const Records = require('spike-records')
const locals = {}
module.exports = {
plugins: [new Records({
addDataTo: locals,
blog: {
transformRaw: (data) => { return data.replace('])}while(1);</x>', '') },
url: 'https://medium.com/glassboard-blog/?format=json'
}
})]
}
Using the template option allows you to write objects returned from records to single page templates. For example, if you are trying to render a blog as static, you might want each post
returned from the API to be rendered as a single page by itself.
The template
option is an object with path
and output
keys. path
is an absolute or relative path to a template to be used to render each item, and output
is a function with the currently iterated item as a parameter, which should return a string representing a path relative to the project root where the single view should be rendered. For example:
const Records = require('spike-records')
const locals = {}
module.exports = {
plugins: [new Records({
addDataTo: locals,
blog: {
url: 'http://blog.com/api/posts',
template: {
path: 'templates/single.sgr',
output: (post) => { return `posts/${post.slug}.html` }
}
}
})]
}
Note that for this feature to work correctly, the data returned from your data source must be an array. If it's not, the plugin will throw an error. If you need to transform the data before it is rendered into templates, you can do so using a transform
function, as such:
const Records = require('spike-records')
const locals = {}
module.exports = {
plugins: [new Records({
addDataTo: locals,
blog: {
url: 'http://blog.com/api/posts',
template: {
transform: (data) => { return data.response.posts },
path: 'templates/single.sml',
output: (post) => { return `posts/${post.slug}.html` }
}
}
})]
}
If you use a transform
function outside of the template
block, this will still work. The difference is that a transform
inside the template
block will only use the transformed data for rendering single templates, whereas the normal transform
option will alter that data that is injected into your view templates as locals, as well as the single templates.
Inside your template, a local called item
will be injected, which contains the contents of the item for which the template has been rendered. It will also contain all the other locals injected by spike-records and otherwise, fully transformed by any transform
functions provided.
- Details on the license can be found here
- Details on running tests and contributing can be found here