You can access AT&T APIs by navigating to http://developer.att.com. Use a trial account or
full membership to create an 'app' and get the app details you will need to use the APIs. You
 need app key and app secret (config.py).

This code is offered without warranty and is provided for educational purposes. The Raspberry
Pi is an interesting target that has garnered considerable interest. This demo code provides
a working model for using audio on the Pi and a Pythonic abstraction for the AT&T Speech APIs.
The Linux ALSA audio system (alsaaudio) captures data in chunks. An alsaaudio input is configured. 
A queue is configured. An audio capture thread is started which puts that data into the queue. 
A speechToText API thread is started that pulls data from the queue and streams to the API.
The transcription is processed, action is taken, and text is converted to synthesized speech
with the textToSpeech API and then played. In this demo one of the Pi's onboard LEDs (OK) is
used for visual feedback controlled by voice command.
Audio files (as opposed to microphone) may also be converted.
    speech.py provides API methods
    speechapp.py provides the demo app

config.py: default configuration of APIs and a few demo app constants.
speech.py: python methods to call oauth and speech APIs
speechapp.py: demo app that captures audio and uses oauth, speechToText, and textToSpeech
utils.py: a few methods used by the other modules.
media files: several sample wav files are provided.

speech.py contains five classes and one method:
    AccessToken: Gets a new access token or restore a saved one.
    SpeechToText: Convert speech into text.
    TextToSpeech: Converts text into synthesized speech.
    APIException: Handles API exceptions.
    API: Base class for APIs.
    main(): Method that demonstrates:
        get access token
        play speech file
        convert that speech file into text
        convert that text into synthesized speech
        play the synthesized speech
    Command Line:
        python speech.py
            Using a default audio speech file, play it, convert to text,
            convert that text into synthesized speech and play that speech.
        python speech.py {audio file path}
            Provide an audio file, convert it to text, that text into synthesized
            speech and play that speech.
speechapp.py Uses the API classes in speech.py to implement a simple demonstration app.
    contains three classes and six methods:
    audio_env(): Display the cards and mixers in the audio subsystem.
    audio_output(): Configure and provide an audio output for playback.
    audio_play() : Play audio on the specified audio output device.
    audio_input(): Provides an audio input - microphone capture.
    AudioCapture: Capture audio from a stream (deque) or file.
    SpeechToTextThread: Execute speech to text call in a thread.
    LEDThread: Control the PI onboard OK LED.
    transcription_processor(): Analyze the transcription and take action.
    Speechapp does not explicity set audio mixers/controls; it uses whatever settings are
    made in the environment. This is a good future optimization target.
    Command Line:
        python speechapp.py
            must run as root (GPIO control)
            get access token
            capture real time audio from microphone and stream to speechToText API
            get the transcription and act on it - example: control the LED
            based on the transcription get a speech response using textToSpeech API and play it
        python speechapp.py {your file}
            The above but source is file not microphone.
config.py: API and program settings.
utils.py: A few common methods.
demo media files:
    static/updateStatus.wav
generated files:
    static/speechToText.wav
    static/textToSpeech.wav
    static/capture_debug.wav

This demo code has been tested on a Raspberry Pi Model B Board w/NOOBS on a 8GB SD.
Wi-Fi: WI-PI Setup:
    vi /etc/network/interfaces
        auto wlan0
        iface wlan0 inet dhcp
        wpa-ssid <name of your WiFi network> wpa-psk <password of your WiFi network>
    sudo /etc/init.d/networking restart
    ifconfig
LED Control:
    The demo app uses one of the on board LEDs to provide visual feedback of "control".
    This article tells you how to control the green OK LED:
        http://www.raspberrypi.org/phpBB3/viewtopic.php?p=136266
PI Specific Code:
    Hardware Specific:  Controlling the LEDs
    Removing the above and modifying audio system configuration should enable the code to run
     on most linux systems.
ISSUE: Audio Quality with USB Audio Devices
    This is an issue that results in less than stellar recognition. You can find additional
    information on this firmware issue at:
    http://www.raspyfi.com/anatomy-of-a-pi-usb-audio-quality-and-related-issues-on-pi/
    This is a big opportunity for improvement.

Use apt-get or pip to install any needed packages.
You may need to install:
    alsaaudio: a Python package for working with Linux ALSA audio subsystem.
        apt-get install python-alsaaudio
    RPi.GPIO: work with I/O - in this application the LEDs
    httplib2: used to make API requests
        pip install httplib2
    pycurl: used to make API requests
        apt-get python-pycurl

Audio commands:
    interactive set up of audio subsystem: alsamixer
    store alsamixer with: alsactl store 1
    list card contents: amixer --card 1 contents
    list your capture devices: arecord -l
    reset the alsa system with: alsactl init
    play a wav file: aplay ./static/updateStatus.wav
    what sound cards do you have:
        cat /proc/asound/cards
USB audio pi issue - read this!:
    ...
pip: If you don't have pip then:
    apt-get install python-pip
PI information:
    cat /proc/cpuinfo
    cat /proc/meminfo
    cat /proc/sys/kernel/hostname
    cat /proc/sys/kernel/ostype
    cat /proc/sys/kernel/osrelease
    cat /proc/sys/kernel/pid_max
    cat /proc/sys/kernel/poweroff_cmd
Display USB devices:
    lsusb -t
    lsusb --verbose | less
Find where your default packages are stored:
    python -c "from distutils.sysconfig import get_python_lib; print(get_python_lib())"

Running Your App at Startup:

    Add a command line to /etc/rc.local for example 'python speechapp.py'