https://github.com/whitead/emoji-math

A complete waste of time
https://github.com/whitead/emoji-math

Last synced: 29 days ago
JSON representation

A complete waste of time

• Host: GitHub
• URL: https://github.com/whitead/emoji-math
• Owner: whitead
• Created: 2021-07-28T19:34:45.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2022-12-11T19:15:14.000Z (almost 2 years ago)
• Last Synced: 2024-08-06T21:20:49.428Z (3 months ago)
• Language: Python
• Homepage:
• Size: 24.4 KB
• Stars: 14
• Watchers: 3
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Emoji Math

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/whitead/emoji-math/blob/master/colab/EmojiMath.ipynb)

Because why not? I put a minimal effort into this project, so please have low expectations.

```sh

pip install emath@https://github.com/whitead/emoji-math/archive/refs/heads/main.zip

```

## Usage

`emoji-math` computes the given python expression and returns either the value or the nearest

5 emojis as measured by cosine similarity.

```sh

>emoji-math 👑 - 🚹 + 🚺

Best Matches:

  👑-🚹+🚺 = 👸

  👑-🚹+🚺 = 👑

  👑-🚹+🚺 = 🤴

```

```sh

>emoji-math 🚹 @ 🚺

🚹 @ 🚺 = 0.32784234338655205

```

```sh

>emoji-math np.sin(🏰)

Best Matches:

  np.sin(🏰) = 🏯

  np.sin(🏰) = 🏰

  np.sin(🏰) = 👸

```

```sh

>emoji-math 🚹 + 🚺

Best Matches:

  🚹+🚺 = 🚻

  🚹+🚺 = 🚺

  🚹+🚺 = 🚹

```

## Options

You can choose between using the Euclidean distance or cosine similarity for reporting results. Cosine works better for

multiplication/division because it only cares about direction. Default is Euclidean.

```sh

>emoji-math --cosine 🚹 + 🚺

Best Matches:

  🚹+🚺 = 👚

  🚹+🚺 = 🚺

  🚹+🚺 = 🚹

```

## Google Colab

Start using emoji math with [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/whitead/emoji-math/blob/master/colab/EmojiMath.ipynb)

## Credit

Made by [@andrewwhite01](https://twitter.com/andrewwhite01)

Vector embeddings from [emoji2vec](https://github.com/uclnlp/emoji2vec) as described in

```bibtex

@misc{eisner2016emoji2vec,

      title={emoji2vec: Learning Emoji Representations from their Description},

      author={Ben Eisner and Tim Rocktäschel and Isabelle Augenstein and Matko Bošnjak and Sebastian Riedel},

      year={2016},

      eprint={1609.08359},

      archivePrefix={arXiv},

      primaryClass={cs.CL}

}

```