Metadata-Version: 2.1
Name: word-embedder
Version: 0.1.1
Summary: Word Embedder
Home-page: UNKNOWN
Author: Solumilken
Author-email: UNKNOWN
License: MIT
Platform: UNKNOWN
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3.6
Requires-Dist: mkdir-p (>=0.1.1)
Requires-Dist: numpy (>=1.15.1)
Requires-Dist: python-dotenv (==0.9.1)

# word-embedder

[![travis][travis-image]][travis-url]
[![pypi][pypi-image]][pypi-url]

[travis-image]: https://img.shields.io/travis/Yoctol/word-embedder.svg?style=flat
[travis-url]: https://travis-ci.org/Yoctol/word-embedder
[pypi-image]: https://img.shields.io/pypi/v/word-embedder.svg?style=flat
[pypi-url]: https://pypi.python.org/pypi/word-embedder

Get pretrained word embedding


## Installation

### Requirements
* Linux
* Python 3.6 and up

`$ pip install word-embedder`

## Usage

### Lookup all existed embedders
```python

from word_embedder import lib

lib.list_all_embedders()  # returns a list of embedder name

```

### Use an existed embedder

1. load an embedder called OHOH
```python

from word_embedder import lib

name = 'OHOH'  # embedder name
embedder = lib[name]

```

2. extract a word vector

- (1) given a word 'juice' (str)
    ```python

    word = 'juice'
    embedder[word]  # returns the corresponding word vector

    # Note: if 'juice' is not in the vocabulary, 
    # OOVError would be raised.

    ```

- (2) given an index 3 (int)
    ```python

    index = 3
    embedder[index]  # returns the corresponding word vector

    # Note: if the index is out of range of vocabulary size,
    # OOVError would be raised.

    ```


