PBF

Library for serializing and deserializing PocketBase filter syntax.

Key Features

Supports all PocketBase filter operators and expressions
Supports both Node and browser
Provides a fluent and (almost) type-safe API for building filters
Can be used to serialize and deserialize filters
TypeScript-support

Installation

You can easily install PBF using npm or yarn:

npm install pbf

yarn add pbf

Usage

PBF makes it simple to work with PocketBase filter syntax. Here's a quick example of how to get started:

import * as pbf from "@nedpals/pbf";
import PocketBase from "pocketbase";

const pb = new PocketBase("<pocketbase url>");

const result = await pb.collection('example').getList(1, 20, {
    filter: pbf.stringify(pbf.and(
        pbf.eq('status', true),
        pbf.gt('created', new Date("2022-08-01"))
    )) // status = true && created > "2022-08-01 10:00:00.000Z"
});

Negating a filter

To negate a filter (eg. equals to not equal, and to or), you may use the not function. This not negate the value but only the operator used.

import * as pbf from "@nedpals/pbf";

pbf.stringify(pbf.not(pbf.eq("is_repost", false))) // is_repost = false

Conditional filter generation

In some instances you want to create a search filter with some of the filters conditionally enabled through short-circuiting. You can do this by adding the .maybe modifier before calling the operators. This will filter out any falsey values and output the appropriate filters.

import * as pbf from "@nedpals/pbf";

pbf.stringify(pbf.and.maybe(
    false && pbf.eq('f', 4),
    null,
    pbf.eq('d', 1),
    0,
    pbf.not(pbf.eq('e', 1)),
)); // d = 1 && e != 1

Comparing multiple values

Instead of repeating yourself writing multiple comparison filters of the same field, PBF provides an easy shortcut through the either modifier.

import * as pbf from "@nedpals/pbf";

// shortcut for pbf.or(pbf.eq("size", "L"), pbf.eq("size", "XL"), pbf.eq("size", "XXL"))
pbf.stringify(pbf.eq.either("size", ["L", "XL", "XXL"])); // (size = "L" || size = "XL") || size = "XXL"

Deserialize / parse raw filter strings

PBF also supports parsing raw filter strings into a proper PBF format. This is great when you want to parse from the URL search query or just want to build a PocketBase-like search experience:

import * as pbf from "@nedpals/pbf";

const result = pbf.parse("title = 'example'"); // equivalent to eq("title", "example");

// You can also inject/bind values to placeholders
const resultB = pbf.parse("title = {:title}", { title: "Foo bar" }) // equivalent of eq("title", "Foo bar")

Format

To make serializing/deserializing possible, PBF stores it as an object following a syntax tree format for distinguishing logical, comparison, and containerized/parenthesized filters.

// Taken and modified from the source code for brevity
type FilterValue = number | boolean | string | Date | null;
type Filter = ComparisonFilter | LogicalFilter | ContainerFilter;
type Metadata = Record<string, any>

// eg. a = 1
interface ComparisonFilter {
    field: string
    op: Operator
    value: FilterValue
    meta?: Metadata
}

// eg. a > 1 && b = 2
interface LogicalFilter {
    lhs: Filter
    op: Operator
    rhs: Filter
    meta?: Metadata
}

// eg. (c = 3)
interface ContainerFilter {
    op: Operator
    filter: Filter
    meta?: Metadata
}

This also makes it easier to craft filters by hand especially when building dynamic facet-like filters:

const filter: Filter = {
    op: "and",
    lhs: {
        op: "gte",
        field: "shoe_size",
        value: 20
    },
    rhs: {
        op: "eq",
        field: "color",
        value: "burgundy"
    }
}

pbf.stringify(filter) // shoe_size >= 20 && color = "burgundy"

Differences with PocketBase

Starting with PocketBase JS SDK 0.19.0, a new feature was added that allows filters to be built similarly to PBF. However, there are some key differences between the two approaches.

PocketBase only ensures that values are properly escaped and bound to the filter. The user is still responsible for constructing the filter syntax, which can be prone to mistakes. PBF, on the other hand, provides a more comprehensive solution by also providing an easy and extensive way to create complex search filters without worrying about the syntax.

PBF was also created as a one-off utility function before this feature was added to PocketBase.

License

pbf is licensed under the MIT License.

Contributing

Contributions are welcome! Please feel free to open issues or pull requests.

Submitting a pull request

Fork it (https://github.com/nedpals/pbf/fork)
Create your feature branch (git checkout -b my-new-feature)
Commit your changes (git commit -am 'Add some feature')
Push to the branch (git push origin my-new-feature)
Create a new Pull Request

Contributors

nedpals - creator and maintainer