https://github.com/parcellab/sift

mongodb queries in JS
https://github.com/parcellab/sift

team-backend

Last synced: 10 months ago
JSON representation

mongodb queries in JS

• Host: GitHub
• URL: https://github.com/parcellab/sift
• Owner: parcelLab
• Created: 2024-05-28T09:30:07.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-09-24T01:43:01.000Z (over 1 year ago)
• Last Synced: 2025-01-23T06:16:00.475Z (11 months ago)
• Topics: team-backend
• Language: TypeScript
• Homepage:
• Size: 2.85 MB
• Stars: 0
• Watchers: 9
• Forks: 0
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • Contributing: CONTRIBUTING.md
  • Codeowners: CODEOWNERS

Awesome Lists containing this project

README

          


[![codecov](https://codecov.parcellab.dev/gh/parcelLab/sift/graph/badge.svg?token=vvbtaY2PL7)](https://codecov.parcellab.dev/gh/parcelLab/sift)

# Sift

Run MongoDB queries in regular javascript.

(fork of https://github.com/crcn/sift.js)

## Table of Contents (Optional)

## About The Project

### Why fork https://github.com/crcn/sift.js?

The original project uses recursion and closures to construct a filter function based on the given MongoDB query.

- this makes debugging difficult, as you have to trace through deep function calls with little context of what the original filter was

- this makes it slow when V8 has not optimized it, since recursive and closure-heavy function calls make use of many stacks

- this makes it hard to change a behavior without affecting others

This fork intends to rewrite sift to a compiler like [ajv](https://github.com/ajv-validator/ajv).

- it intends to provide a verbose mode where each leaf comparison can provide a pass/fail status, and/or what the comparison function looks like

- it intends to be tiny and fast

- it intends to be safe, where $where function calls are constrained to the input object

- it intends to stick to the same behavior as MongoDB query operators as possible

## Installation

Supports node >= 18

`npm i @parcellab/sift`

## Usage

## Contributing

- You need a local copy of MongoDB (e.g. via docker, compose file provided) configured via TEST_MONGODB_URL which defaults to `mongo://localhost:27017/test`

- To update the benchmarks:

  1. run `npm run test:bench` which will update the human readable `output.txt`

  2. run `npm run test:bench:csv`

  3. open `output.xlsx` in Microsoft Excel

  4. click on Refresh Data Sources (this imports `output.csv` automatically)

  5. export the updated chart as `output.png`

[Contribution guidelines](CONTRIBUTING.md)