https://github.com/stereobooster/pagefind-instantsearch
InstantSearch adapter for pagefind
https://github.com/stereobooster/pagefind-instantsearch
client-side faceted faceted-search facets full-text full-text-search instantsearch pagefind search
Last synced: 6 months ago
JSON representation
InstantSearch adapter for pagefind
- Host: GitHub
- URL: https://github.com/stereobooster/pagefind-instantsearch
- Owner: stereobooster
- Created: 2023-11-21T03:06:49.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-09-17T23:47:14.000Z (9 months ago)
- Last Synced: 2024-10-31T09:18:55.676Z (8 months ago)
- Topics: client-side, faceted, faceted-search, facets, full-text, full-text-search, instantsearch, pagefind, search
- Language: TypeScript
- Homepage: https://pagefind-instantsearch.netlify.app
- Size: 1.9 MB
- Stars: 3
- Watchers: 2
- Forks: 0
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# InstantSearch adapter for pagefind monorepo
It is the monorepo repository:
- [InstantSearch adapter for pagefind](/packages/pagefind-instantsearch/) 
- and [demo site](/packages/demo/) [](https://app.netlify.com/sites/pagefind-instantsearch/deploys)
## Idea
The idea is to create an InstantSearch adapter for [pagefind](https://pagefind.app/) and build the same demo site as I did for [facets](https://github.com/stereobooster/facets). This way we can compare "apples to apples".
To do so I wanted to use the same Schema as in **facets**:
- [to convert JSON file to pagefind index](/packages/demo/bin/pfis.js)
- [to convert pagefind response to InstantSearch Hit](/packages/pagefind-instantsearch/src/adaptResponse.ts)
I started implementation with hardcoded conversion. I managed to create a [working demo](https://pagefind-instantsearch.netlify.app/search/?query=chrome). However, I realized that pagefind can't be used as a general faceted search engine.
### Type conversion
pagefind supports only string values, so all values need to be converted to strings before indexing and back to the initial type on the client side.
Indexing:
```js
brand: [item.brand],
categories: item.categories,
price: [`${item.price}`],
rating: [`${item.rating}`],
free_shipping: [`${item.free_shipping}`],
```
Client-side:
```ts
{
objectID: item.id,
...data.meta,
categories: data.filters.categories,
price: parseFloat(data.filters.price[0]),
rating: parseFloat(data.filters.rating[0]),
}
```
### Sorting
See:
- https://github.com/CloudCannon/pagefind/issues/513
- https://github.com/CloudCannon/pagefind/issues/259
### Stats for numerical facets
Because there is no support for numbers to get stats we need to convert strings to numbers on the client side and calculate stats:
```ts
if (schema[field]?.type === "number") {
const values: number[] = [];
entries.forEach((a) => values.push(parseFloat(a[0])));
facetsStats[field] = {
min: Math.min(...values),
max: Math.max(...values),
};
}
```
### Filter for numeric fields
I didn't implement a filter for numeric fields, like `price >= 40 AND price <= 100`. The price slider doesn't work in the deme
## Status
After I realized that it was not practical to use a general faceted search engine I gave up and didn't finish code for schema transformation.
**This code is not recomended for prodcution use**.
## Development
```sh
pnpm i
pnpm run dev
```
## TODO
- CLI to covert JSON to pagefind index
- maybe cosmiconfig to pass configuration
- it can use the same schema as [`facets`](/packages/pagefind-instantsearch/src/Facets.ts)
- input
- output
- adapter for instantsearch
- filtering for numeric facets