BelkaDev / GoogleTts

Unofficial Google text-to-speech with extra features like vocal translation and word filtering.

GoogleTts

Google Text-to-speech from the comfort of your command line. Wow very cool!

Requirements

jq
xclip (optional)

How to install?

git clone https://github.com/BelkaDev/GoogleTts.git ~/Tts && cd ~/Tts
chmod +x *
copy both scripts to your $PATH directory

Features

• Cli text to speech or textual translations
• Language selection and automatic detection
• Real-time translation and cross-language speech
• Patterns filtering
• Read current selection

How to use?

Text-to-speech

• -a, --auto for auto-detection mode.
  
  • -a "こにちわ" will read in Japanese
    
  • -a "hola" will read in Spanish
• -t, --translate <lang> to translate and read
  
  • -t "こにちわ" will translate to preferred language and read.
  • -t de "こにちわ" will translate to German and read.
• -l, --language <lang> to change the language
  
  • -l es "hi" reads input in Spanish voice

Notes
All options can be specified at runtime or set up
by default in the settings, accepted values are `true` , `false`
for boolean attributes, for language codes
use one of valid language tags supported by Google.

The complete list can be found here

Stack order follows this order: language>auto>translate
Meaning the highest will always neglect the others.
-l en -t こにちわ will not perform translation.

If no text input is given, current selection (highlighted text),
is processed, this requires an X environement.
When run on terminal, double quotes must wrap the input.

Translation
the translation script can be used independently, but not the other way around
most broad syntax is as follows:
translate [from <source> to <target> <text>]
Each parameter can be ommited, in that case automatic values are considered.
examples:
translate to ko "hello": detects language and translates to Korean
translate from fr "bonjour": translates from French to preferred language
translate: translate selection from detected language to preferred language

Output format contains synonyms and examples and is made easy to parse using pipes.

• parse translation: | grep -w "translation" | cut -f2 -d ":"
• parse source language: | grep -w "source_lang" | cut -f2 -d ":"

Extra

• A set of patterns can be configured to ignore undesirable/irrelevant words or replace specific sequences. As all options it can be changed at the top of the file, it follows this structure
  { "sequence" ; "alternative", ...}
  
  Setting up an empty string will ignore the corresponding sequence. (option is disabled by default.)

• Audio files can be stored in the directory of your choice.

To do:

Todo: feed files or web urls to the script

Known issues

All parameters should be interchangeable, but this can lead to unexpected results if they get mixed up.
Encoding erros: some characters may cause the script to fail.