https://github.com/vjousse/speechtotext-api

Speechtotext API wrapper
https://github.com/vjousse/speechtotext-api

dramatiq fastapi speech-recognition speech-to-text tortoise-orm

Last synced: 6 days ago
JSON representation

Speechtotext API wrapper

• Host: GitHub
• URL: https://github.com/vjousse/speechtotext-api
• Owner: vjousse
• Created: 2022-01-28T17:06:22.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-03-07T14:08:44.000Z (almost 3 years ago)
• Last Synced: 2024-10-29T12:58:34.259Z (about 2 months ago)
• Topics: dramatiq, fastapi, speech-recognition, speech-to-text, tortoise-orm
• Language: Python
• Homepage:
• Size: 42 KB
• Stars: 0
• Watchers: 3
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Speech to text - API

This project provides a simple HTTP API as a frontend to any Sppech Recognition service like [Kaldi](https://kaldi-asr.org). It's even more generic: it provides a simple interface to any script taking a file as input and sending multiple files as results.

It's written in Python using [FastAPI](https://fastapi.tiangolo.com/) for the web framework, [Tortoise ORM](https://tortoise-orm.readthedocs.io/en/latest/) for the async ORM and [dramatiq](https://dramatiq.io/) for managing the background tasks. The stack is using Nginx, Redis and PostgreSQL. And of course, everything is available using Docker images.

## Client

A frontend client developped using Elm can be found [here](https://github.com/vjousse/speechtotext-api-client).

## Docker

### Web application

    docker-compose --env-file .env.docker up

### Kaldi worker

    docker-compose -f docker-compose-worker.yml --env-file .env.docker up --build

## Install

    pip install -r config/requirements.txt

## Run

Web API:

    uvicorn app.main:app --reload --reload-dir app

Task manager (dramatiq):

    dramatiq asr_worker.download

## Configure

Configuration can be found in `app/core/config.py`.

## Init database

    aerich init-db

## Testing

    pytest --disable-warnings -s app/tests

The `-s` flag is used to show captured stdout (print) even if the test is successful.

## Files

Upload a file:

        curl -L -v \

            -F file=@/home/vjousse/wifi.md \

            -H "Authorization: Bearer eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ1c2VyX2lkIjoiODAwZTk1NjQtNjgwNC00YWI1LWJjNTktYTA4ODE4MjIyN2JlIiwiYXVkIjpbImZhc3RhcGktdXNlcnM6YXV0aCJdLCJleHAiOjE2MzA1MDU5ODN9.W6eq5DZMteTSHMZ1bxlqGaBsQTyqpCSYPSTdJcK3C04" \

            http://localhost:8000/api/file/upload

## Users

Using [FastAPIUsers](https://fastapi-users.github.io/fastapi-users/usage/flow/)

### Registering

    curl \

        -H "Content-Type: application/json" \

        -X POST \

        -d "{\"email\": \"[email protected]\",\"password\": \"phil\"}" \

        http://localhost:8000/auth/register

or

    http POST http://localhost:8000/auth/register [email protected] password=phil --session=session

Returns:

```json

{

    "id":"800e9564-6804-4ab5-bc59-a088182227be",

    "email":"[email protected]",

    "is_active":true,

    "is_superuser":false,

    "is_verified":false

}

```

### Login

    curl \

        -H "Content-Type: multipart/form-data" \

        -X POST \

        -F "[email protected]" \

        -F "password=phil" \

        http://localhost:8000/auth/jwt/login

Returns:

```json

{

    "access_token":"eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ1c2VyX2lkIjoiODAwZTk1NjQtNjgwNC00YWI1LWJjNTktYTA4ODE4MjIyN2JlIiwiYXVkIjpbImZhc3RhcGktdXNlcnM6YXV0aCJdLCJleHAiOjE2MzAxMzk5OTJ9.w-ZWpm51fyybFivmKjun3qbXuqwXCgYyxGbPD1yhIr4",

    "token_type":"bearer"

}

```