python-polly 0.1.0

Wrapper for AWS Polly with segmentation, playback queue and client-server setup

python-polly

python-polly is a simple wrapper around AWS's text-to-speech service Polly.

Features are:

• Cache requests to AWS Polly locally
• Tokenize input text into sentences or words in order to maximize usage of cache
• Playback queue for audio output
• Client-server setup to support concurrent output by multiple applications

Setup

pip install --user python-polly[server]

Most of the dependencies are only required by the server so I put them in as an extra option. To get the dependencies for the server use python-polly[server] otherwise just python-polly.

Usage

Server

Start the server. Default port is 6000.

polly server

Client

CLI

polly say "hello world"

Python

import polly.client
polly.client.say('hello world')

Any other JSON-RPC client e.g. cURL

curl -d '{"id": "null", "jsonrpc": "2.0", "method": "say", "params": ["hello", "Joanna", "sentence", true]}' localhost:6000

Links

Playback

https://people.csail.mit.edu/hubert/pyaudio/docs/#example-callback-mode-audio-i-o
https://github.com/spatialaudio/python-sounddevice/blob/master/examples/play_long_file.py