attspeech

README

attspeech

Go client for the AT&T Speech API.

Installation

go get bitbucket.org/jsgoecke/attspeech

Documentation

http://godoc.org/github.com/jsgoecke/attspeech

Usage

Speech to Text Result

package main

import (
	"bytes"
	"fmt"
	"github.com/jsgoecke/attspeech"
	"io"
	"os"
)

func main() {
	client := attspeech.New(os.Getenv("ATT_APP_KEY"), os.Getenv("ATT_APP_SECRET"), "")
	client.SetAuthTokens()

	data := &bytes.Buffer{}
	file, err := os.Open("../test/test.wav")
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
	defer file.Close()
	_, err = io.Copy(data, file)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}

	apiRequest := client.NewAPIRequest(attspeech.STTResource)
	apiRequest.ContentType = "audio/x-wav"
	apiRequest.Data = data
	response, err := client.SpeechToText(apiRequest)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
	fmt.Println(response)
}

Speech to Text Custom Result

package main

import (
	"bytes"
	"fmt"
	"github.com/jsgoecke/attspeech"
	"io"
	"os"
)

func main() {
	client := attspeech.New(os.Getenv("ATT_APP_KEY"), os.Getenv("ATT_APP_SECRET"), "")
	client.SetAuthTokens()

	data := &bytes.Buffer{}
	file, err := os.Open("../test/test.wav")
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
	defer file.Close()
	_, err = io.Copy(data, file)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}

	apiRequest := client.NewAPIRequest(attspeech.STTCResource)
	apiRequest.ContentType = "audio/x-wav"
	apiRequest.Data = data
	apiRequest.Filename = "test.wav"
	response, err := client.SpeechToTextCustom(apiRequest, srgsXML(), plsXML())
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
	fmt.Println(response)
}

func plsXML() string {
	return `<?xml version="1.0" encoding="UTF-8"?> 
			<lexicon version="1.0" alphabet="sampa" xml:lang="en-US"> 
			   <lexeme> 
			       <grapheme>star</grapheme> 
			       <phoneme>tS { n</phoneme> 
			   </lexeme> 
			</lexicon>`
}

func srgsXML() string {
	return `<grammar root="top" xml:lang="en-US"> 
			  <rule id="CONTACT"> 
			      <one-of> 
			        <item>star</item> 
			        <item>key</item> 
			      </one-of> 
			  </rule> 
			  <rule id="top" scope="public"> 
			      <item> 
			          <one-of> 
			            <item>greeting</item> 
			            <item>the administration menu</item> 
			          </one-of> 
			      </item> 
			  <ruleref uri="#CONTACT"/> 
			  </rule> 
			</grammar>`
}

Text to Speech Result

package main

import (
	"fmt"
	"github.com/jsgoecke/attspeech"
	"io/ioutil"
	"os"
)

func main() {
	client := attspeech.New(os.Getenv("ATT_APP_KEY"), os.Getenv("ATT_APP_SECRET"), "")
	client.SetAuthTokens()
	apiRequest := client.NewAPIRequest(attspeech.TTSResource)
	apiRequest.ContentType = "text/plain"
	apiRequest.Accept = "audio/x-wav"
	apiRequest.Text = "I want to be an airborne ranger, I want to live the life of danger."
	data, err := client.TextToSpeech(apiRequest)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
	err = ioutil.WriteFile("/Users/jsgoecke/Desktop/tts_test.wav", data, 0644)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}
}

Testing

cd attspeech
go get github.com/smartystreets/goconvey/convey
go test

Test Coverage

http://gocover.io/github.com/jsgoecke/attspeech

Lint

http://go-lint.appspot.com/github.com/jsgoecke/attspeech

Documentation

Index

• Constants
• type APIError
• type APIRequest
• type Client
  • func New(id string, secret string, apiBase string) *Client
  • func (client *Client) NewAPIRequest(resource string) *APIRequest
  • func (client *Client) SetAuthTokens() error
  • func (client *Client) SpeechToText(apiRequest *APIRequest) (*Recognition, error)
  • func (client *Client) SpeechToTextCustom(apiRequest *APIRequest, grammar string, dictionary string) (*Recognition, error)
  • func (client *Client) TextToSpeech(apiRequest *APIRequest) ([]byte, error)
• type Recognition
• type Token

Constants

const (
	// APIBase is the base URL for the ATT Speech API
	APIBase = "https://api.att.com"
	// STTResource is the speech to text resource
	STTResource = "/speech/v3/speechToText"
	// STTCResource is the speech to text custom resource
	STTCResource = "/speech/v3/speechToTextCustom"
	// TTSResource is the text to speech resource
	TTSResource = "/speech/v3/textToSpeech"
	// OauthResource is the oauth resource
	OauthResource = "/oauth/access_token"
	// UserAgent is the user agent use for the HTTP client
	UserAgent = "GoATTSpeechLib"
	// Version is the version of the ATT Speech API
	Version = "0.1"
)

Types

type APIError

type APIError struct {
	RequestError struct {
		ServiceException struct {
			MessageID string `json:"MessageId"`
			Text      string `json:"Text"`
			Variables string `json:"Variables"`
		} `json:"ServiceException"`
		PolicyException struct {
			MessageID string `json:"MessageId"`
			Text      string `json:"Text"`
			Variables string `json:"Variables"`
		} `json:"PolicyException"`
	} `json:"RequestError"`
}

APIError represents an error from the AT&T Speech API

type APIRequest

type APIRequest struct {
	Accept            string
	Authorization     string
	ContentLanguage   string
	ContentLength     string
	ContentType       string
	Data              io.Reader
	Filename          string
	Tempo             string
	Text              string
	TransferEncoding  string
	VoiceName         string
	Volume            string
	UserAgent         string
	XArg              string
	XSpeechContext    string
	XSpeechSubContext string
}

APIRequest represents the parameters for a Text to Speech request

type Client

type Client struct {
	APIBase       string
	STTResource   string
	STTCResource  string
	TTSResource   string
	OauthResource string
	ID            string
	Secret        string
	Tokens        map[string]*Token
	Scope         [3]string
}

Client is an ATT Speech API client

func New

func New(id string, secret string, apiBase string) *Client

New creates a new AttSpeechClient

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()

func (*Client) NewAPIRequest

func (client *Client) NewAPIRequest(resource string) *APIRequest

NewAPIRequest sets the common headers for TTS and STT

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()
apiRequest := client.NewAPIRequest(TTSResource)

Note, when setting apiRequest.XArg, always append with '+=', unless you specifically intend to overwrite the defaults for ClientApp, ClientVersion, DeviceType and DeviceOs

func (*Client) SetAuthTokens

func (client *Client) SetAuthTokens() error

SetAuthTokens sets the provided authorization tokens for the client

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()

func (*Client) SpeechToText

func (client *Client) SpeechToText(apiRequest *APIRequest) (*Recognition, error)

SpeechToText converts an audio file to text

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()
apiRequest := client.NewAPIRequest(STTResource)
apiRequest.Data = data // where data is audio content as io.Reader
apiRequest.ContentType = "audio/wav"
result, apiError, err := client.SpeechToText(apiRequest)

More details available here:

http://developer.att.com/apis/speech/docs#resources-speech-to-text

func (*Client) SpeechToTextCustom

func (client *Client) SpeechToTextCustom(apiRequest *APIRequest, grammar string, dictionary string) (*Recognition, error)

SpeechToTextCustom converts an audio file to text

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()
apiRequest := client.NewAPIRequest(STTResource)
apiRequest.Data = data // where data is audio content as io.Reader
apiRequest.ContentType = "audio/wav"
apiRequest.Filename = "test.wav"
result, apiError, err := client.SpeechToTextCustom(apiRequest, "<some srgs XML>", "<some pls XML>")

More details available here:

http://developer.att.com/apis/speech/docs#resources-speech-to-text-custom

func (*Client) TextToSpeech

func (client *Client) TextToSpeech(apiRequest *APIRequest) ([]byte, error)

TextToSpeech converts text to a speech file

client := attspeech.New("<id>", "<secret>", "")
client.SetAuthTokens()

apiRequest := client.NewAPIRequest(TTSResource)
apiRequest.Accept = "audio/x-wav",
apiRequest.VoiceName = "crystal",
apiRequest.Text = "I want to be an airborne ranger, I want to live the life of danger.",
data, err := client.TextToSpeech(apiRequest)

More details available here:

http://developer.att.com/apis/speech/docs#resources-text-to-speech

type Recognition

type Recognition struct {
	Recognition struct {
		Status     string `json:"Status"`
		ResponseID string `json:"ResponseId"`
		NBest      []struct {
			Hypothesis    string    `json:"Hypothesis"`
			LanguageID    string    `json:"LanguageId"`
			Confidence    float32   `json:"Confidence"`
			Grade         string    `json:"Grade"`
			ResultText    string    `json:"ResultText"`
			Words         []string  `json:"Words"`
			WordScores    []float32 `json:"WordScores"`
			NluHypothesis struct {
				OutComposite []struct {
					Grammar string `json:"Grammar"`
					Out     string `json:"Out"`
				} `json:"OutComposite"`
			} `json:"NluHypothesis"`
		} `json:"NBest"`
	} `json:"Recognition"`
}

Recognition represents at AT&T recognition response

type Token

type Token struct {
	AccessToken  string `json:"access_token"`
	TokenType    string `json:"token_type"`
	ExpiresIn    int    `json:"expires_in"`
	RefreshToken string `json:"refresh_token"`
}

Token represents the authorization tokens returned by the AT&T Speech API