TEA - Translation Engine Architect
Project description
TEA - Translation Engine Architect
A command line tool to create translation engine.
Install
First install pipx then:
pipx install pangeamt-tea
Usage
Step 1: Create a new project
tea new --customer customer --srcLang es --tgtLang en --flavor automotion --version 0.0.1
This command will create the project directory structure:
├── customer_es_en_automotion_0.0.1
│ ├── config.yml
│ └── data
Then enter in the directory
cd customer_es_en_automotion_0.0.1
Step 2: Configuration
Tokenizer
A tokenizer can be applied to source and target
tea tokenizer --src mecab --tgt moses
To list all available tokenizer:
tea tokenizer --list
Truecaser
tea truecaser --src --tgt
BPE
tea bpe -s -t
data['processors'], data['tokenizer'], data['truecaser'], data['bpe'], data['trainer'])
Step 3:
Copy some multilingual ressources (.tmx, bilingual files, .af ) into the 'data' directory
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
pangeamt-tea-0.2.5.tar.gz
(15.6 kB
view hashes)
Built Distribution
Close
Hashes for pangeamt_tea-0.2.5-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2ec84d6ac06a3d5e5148f580935b62fc7c8885ec4b57ea1797420d2fa4160588 |
|
MD5 | 9fc4c6403fcef2fb7b0bbc751a07c913 |
|
BLAKE2b-256 | 48e411c4ea4fd878a418bd6234987e64585824fadebf74c5ae5db73d05127b23 |