docs.sheetjs.com/docz/docs/03-demos/04-static/05-vitejs.md
2023-10-22 03:32:00 -04:00

349 lines
9.9 KiB
Markdown

---
title: ViteJS Spreadsheet Plugins
sidebar_class_name: red
sidebar_label: ViteJS
description: Make static websites from spreadsheets using ViteJS. Seamlessly integrate data into your website using SheetJS. Empower non-technical people to write content from Excel.
pagination_prev: demos/net/index
pagination_next: demos/mobile/index
sidebar_custom_props:
type: bundler
---
import current from '/version.js';
import CodeBlock from '@theme/CodeBlock';
[ViteJS](https://vitejs.dev/) is a modern build tool for generating static sites.
It has a robust JavaScript-powered plugin system[^1]
[SheetJS](https://sheetjs.com) is a JavaScript library for reading and writing
data from spreadsheets.
This demo uses ViteJS and SheetJS to pull data from a spreadsheet and display
the content in an HTML table. We'll explore how to load SheetJS in a ViteJS
plugin and compare a few different data loading strategies.
The ["Complete Demo"](#complete-demo) section creates a complete website powered
by a XLSX spreadsheet.
:::info pass
This demo covers use cases where data is available at build time. This flow is
suitable for end of week or end of month (EOM) reports published in HTML tables.
For processing user-submitted files in the browser, the
[ViteJS "Bundlers" demo](/docs/demos/frontend/bundler/vitejs) shows client-side
bundling of the SheetJS library. The ["ReactJS" demo](/docs/demos/frontend/react)
shows example sites using ViteJS with the ReactJS starter.
:::
## Plugins
ViteJS supports static asset imports[^2], but the default raw loader interprets data
as UTF-8 strings. This corrupts binary formats like XLSX and XLS, but a custom
loader can override the default behavior.
For simple tables of data, ["Pure Data Plugin"](#pure-data-plugin) is strongly
recommended. The file processing is performed at build time and the generated
site only includes the raw data.
For more complex parsing or display logic, ["Base64 Plugin"](#base64-plugin) is
preferable. Since the raw parsing logic is performed in the page, the library
will be included in the final bundle.
### Pure Data Plugin
For a pure static site, a plugin can load data into an array of row objects. The
SheetJS work is performed in the plugin. The library is not loaded in the page!
The following diagram depicts the workbook waltz:
```mermaid
flowchart LR
file[(workbook\nfile)]
subgraph SheetJS operations
buffer(NodeJS\nBuffer)
aoo(array of\nobjects)
end
html{{HTML\nTABLE}}
file --> |vite.config.js\ncustom plugin| buffer
buffer --> |vite.config.js\ncustom plugin| aoo
aoo --> |main.js\nfrontend code| html
```
This ViteJS plugin will read spreadsheets using the SheetJS `read` method[^3]
and generate arrays of row objects with `sheet_to_json`[^4]:
```js title="vite.config.js"
import { readFileSync } from 'fs';
import { read, utils } from 'xlsx';
import { defineConfig } from 'vite';
export default defineConfig({
assetsInclude: ['**/*.xlsx'], // xlsx file should be treated as assets
plugins: [
{ // this plugin handles ?sheetjs tags
name: "vite-sheet",
transform(code, id) {
if(!id.match(/\?sheetjs$/)) return;
var wb = read(readFileSync(id.replace(/\?sheetjs$/, "")));
var data = utils.sheet_to_json(wb.Sheets[wb.SheetNames[0]]);
return `export default JSON.parse('${JSON.stringify(data)}')`;
}
}
]
});
```
In frontend code, the loader will look for all modules with a `?sheetjs`
query string. The default export is an array of row objects.
The following example script displays the data in a table:
```js title="main.js"
import data from './data/pres.xlsx?sheetjs';
document.querySelector('#app').innerHTML = `<table>
<thead><tr><th>Name</th><th>Index</th></tr></thead>
<tbody>
${data.map(row => `<tr>
<td>${row.Name}</td>
<td>${row.Index}</td>
</tr>`).join("\n")}
</tbody>
</table>`;
```
### Base64 Plugin
This plugin pulls in data as a Base64 string that can be read with `read`[^5].
While this approach works, it is not recommended since it loads the library in
the front-end site.
The following diagram depicts the workbook waltz:
```mermaid
flowchart LR
file[(workbook\nfile)]
subgraph SheetJS operations
base64(base64\nstring)
aoo(array of\nobjects)
end
html{{HTML\nTABLE}}
file --> |vite.config.js\ncustom plugin| base64
base64 --> |main.js\nfrontend code| aoo
aoo --> |main.js\nfrontend code| html
```
This ViteJS plugin will read spreadsheet files and export the data as a Base64
string. SheetJS is not imported in the plugin:
```js title="vite.config.js"
import { readFileSync } from 'fs';
import { defineConfig } from 'vite';
export default defineConfig({
assetsInclude: ['**/*.xlsx'], // mark that xlsx file should be treated as assets
plugins: [
{ // this plugin handles ?b64 tags
name: "vite-b64-plugin",
transform(code, id) {
if(!id.match(/\?b64$/)) return;
var path = id.replace(/\?b64/, "");
var data = readFileSync(path, "base64");
return `export default '${data}'`;
}
}
]
});
```
When importing using the `b64` query, the raw Base64 string will be exposed.
`read` will process the Base64 string using the `base64` input type[^6]:
```js title="main.js"
import { read, utils } from "xlsx";
/* import workbook data */
import b64 from './data.xlsx?b64';
/* parse workbook and pull data from the first worksheet */
const wb = read(b64, { type: "base64" });
const wsname = wb.SheetNames[0];
const data = utils.sheet_to_json(wb.Sheets[wsname]);
document.querySelector('#app').innerHTML = `<table>
<thead><tr><th>Name</th><th>Index</th></tr></thead>
<tbody>
${data.map(row => `<tr>
<td>${row.Name}</td>
<td>${row.Index}</td>
</tr>`).join("\n")}
</tbody>
</table>`;
```
## Complete Demo
:::note
This demo was tested on 2023 September 07 with ViteJS version `4.4.9`.
:::
The demo walks through the process of creating a new ViteJS website from scratch.
A Git repository with the completed site can be cloned[^7].
### Initial Setup
1) Create a new site with the `vue-ts` template and install the SheetJS package:
<CodeBlock language="bash">{`\
npm create vite@latest sheetjs-vite -- --template vue-ts
cd sheetjs-vite
npm i
npm i --save https://cdn.sheetjs.com/xlsx-${current}/xlsx-${current}.tgz`}
</CodeBlock>
2) Download and replace [`vite.config.ts`](pathname:///vitejs/vite.config.ts)
```bash
curl -O https://docs.sheetjs.com/vitejs/vite.config.ts
```
3) Make a `data` folder and download <https://sheetjs.com/pres.xlsx> :
```bash
mkdir -p data
curl -L -o data/pres.xlsx https://sheetjs.com/pres.xlsx
```
### Pure Data Test
4) Run the dev server:
```bash
npm run dev
```
Open a browser window to the displayed URL (typically `http://localhost:5173` )
5) Replace the component `src/components/HelloWorld.vue` with:
```html title="src/components/HelloWorld.vue"
<script setup lang="ts">
// @ts-ignore
import data from '../../data/pres.xlsx?sheetjs';
</script>
<template>
<table>
<tr><th>Name</th><th>Index</th></tr>
<tr v-for="(row,R) in data" v-bind:key="R">
<td>{{row.Name}}</td>
<td>{{row.Index}}</td>
</tr>
</table>
</template>
```
Save and refresh the page. A data table should be displayed
6) Stop the dev server and build the site
```bash
npm run build
npx http-server dist/
```
The terminal will display a URL, typically `http://127.0.0.1:8080` . Access
that page with a web browser.
7) To confirm that only the raw data is present in the page, view the page
source. The code will reference some script like `/assets/index-HASH.js`.
Open that script.
Searching for `Bill Clinton` reveals the following:
```js
{"Name":"Bill Clinton","Index":42}
```
Searching for `BESSELJ` should reveal no results. The SheetJS scripts are not
included in the final site!
:::info pass
ViteJS also supports "Server-Side Rendering". In SSR, only the HTML table
would be added to the final page. Details are covered in the ViteJS docs[^8].
:::
### Base64 Test
8) Run the dev server:
```bash
npm run dev
```
Open a browser window to the displayed URL.
9) Replace the component `src/components/HelloWorld.vue` with:
```html title="src/components/HelloWorld.vue"
<script setup lang="ts">
// @ts-ignore
import b64 from '../../data/pres.xlsx?b64';
import { read, utils } from "xlsx";
/* parse workbook and convert first sheet to row array */
const wb = read(b64);
const ws = wb.Sheets[wb.SheetNames[0]];
interface IPresident { Name: string; Index: number; };
const data = utils.sheet_to_json<IPresident>(ws);
</script>
<template>
<table>
<tr><th>Name</th><th>Index</th></tr>
<tr v-for="(row,R) in data" v-bind:key="R">
<td>{{row.Name}}</td>
<td>{{row.Index}}</td>
</tr>
</table>
</template>
```
10) Stop the dev server and build the site
```bash
npm run build
npx http-server dist/
```
The terminal will display a URL ( `http://127.0.0.1:8080` ). Access that page
with a web browser.
11) To confirm that the object data is not present in the page, view the page
source. The code will reference some script like `/assets/index-HASH.js` with
a different hash from the previous test. Open that script.
Searching for `BESSELJ` should match the code:
```
425:"BESSELJ"
```
Searching for `Bill Clinton` should yield no results. The SheetJS library is
embedded in the final site and the data is parsed when the page is loaded.
[^1]: See ["Using Plugins"](https://vitejs.dev/guide/using-plugins.html) in the ViteJS documentation.
[^2]: See ["Static Asset Handling"](https://vitejs.dev/guide/assets.html) in the ViteJS documentation.
[^3]: See [`read` in "Reading Files"](/docs/api/parse-options)
[^4]: See [`sheet_to_json` in "Utilities"](/docs/api/utilities/array#array-output)
[^5]: See [`read` in "Reading Files"](/docs/api/parse-options)
[^6]: See [the "base64" type in "Reading Files"](/docs/api/parse-options#input-type)
[^7]: See [`SheetJS/sheetjs-vite`](https://git.sheetjs.com/sheetjs/sheetjs-vite/) on the SheetJS git server.
[^8]: See ["Server-Side Rendering"](https://vitejs.dev/guide/ssr.html) in the ViteJS documentation.