cloudlanguagetools 17.0.0__py3-none-any.whl

cloudlanguagetools/alibaba.py

@@ -0,0 +1,270 @@
+import json
+import requests
+import time
+import datetime
+import uuid
+import urllib
+import hmac
+import base64
+import logging
+import tempfile
+import pprint
+from typing import List
+
+import cloudlanguagetools.service
+import cloudlanguagetools.options
+import cloudlanguagetools.constants
+import cloudlanguagetools.languages
+import cloudlanguagetools.ttsvoice
+import cloudlanguagetools.constants
+import cloudlanguagetools.languages
+import cloudlanguagetools.ttsvoice
+import cloudlanguagetools.errors
+from cloudlanguagetools.options import AudioFormat
+
+logger = logging.getLogger(__name__)
+
+ALIBABA_VOICE_SPEED_DEFAULT = 0
+ALIBABA_VOICE_PITCH_DEFAULT = 0
+
+VOICE_OPTIONS = {
+    'speed': {'type': 'number_int', 'min': -500, 'max': 500, 'default': ALIBABA_VOICE_SPEED_DEFAULT},
+    'pitch': {'type': 'number_int', 'min': -500, 'max': 500, 'default': ALIBABA_VOICE_PITCH_DEFAULT},
+    'volume': {'type': 'number_int', 'min': 0, 'max': 100, 'default': 50},
+    cloudlanguagetools.options.AUDIO_FORMAT_PARAMETER: {
+        'type': cloudlanguagetools.options.ParameterType.list.name,
+        'values': ['pcm', 'wav', 'mp3'],
+        'default': 'mp3'
+    }
+}
+
+
+class AlibabaService(cloudlanguagetools.service.Service):
+    def __init__(self):
+        self.access_token = None
+        
+    def configure(self, config):
+        self.access_key_id = config['access_key_id']
+        self.access_key_secret = config['access_key_secret']
+        self.app_key = config['app_key']
+
+    def refresh_token(self):
+        logger.info("refreshing token")
+        params = {
+            "AccessKeyId": self.access_key_id,
+            "Action": "CreateToken",
+            "Version": "2019-07-17",
+            "Format": "JSON",
+            "RegionId": "ap-southeast-1",
+            "Timestamp": datetime.datetime.now(datetime.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
+            "SignatureMethod": "HMAC-SHA1",
+            "SignatureVersion": "1.0",
+            "SignatureNonce": str(uuid.uuid4())
+        }
+
+        # sort by keys alphabetically
+        params = dict(sorted(params.items()))
+        # timestamp needs to be double-quoted by the end
+        params["Timestamp"] = urllib.parse.quote(params["Timestamp"], safe='')
+        
+        # urlencode with noop lambda for no quoting - we will quote later
+        params_str = urllib.parse.urlencode(params, quote_via=lambda a, b, c, d: a)
+        
+        # this is always /, as we always hit the path / on the API
+        url_encoded = urllib.parse.quote("/", safe='')
+        str_to_sign = f"GET&{url_encoded}&{urllib.parse.quote(params_str, safe='')}"
+        str_to_sign = str_to_sign.encode("utf-8")
+        
+        key = self.access_key_secret + "&"
+        key = key.encode("utf-8")
+        
+        # calculate HMAC-SHA1 digest and convert to base64
+        dig = hmac.new(key, str_to_sign, "sha1").digest()
+        dig = base64.standard_b64encode(dig).decode("utf-8")
+        
+        # signature also needs to be quoted
+        signature = urllib.parse.quote(dig, safe='')
+        params_str = f"Signature={signature}&{params_str}"
+        
+        response = requests.get(f"http://nlsmeta.ap-southeast-1.aliyuncs.com/?{params_str}")
+        
+        if response.status_code != 200:
+            logger.warning(f"Token request failed: {response.text}")
+            raise cloudlanguagetools.errors.RequestError("Token request failed", None, response.text)
+            
+        data = response.json()
+        self.access_token = data["Token"]
+        logger.info(f"Got access token: {self.access_token}")
+
+    def get_tts_audio(self, text, voice_key, options):
+        logger.debug(f'get_tts_audio, text: {text} voice_key: {voice_key}')
+        if not self.access_token or self.access_token["ExpireTime"] <= int(time.time()):
+            self.refresh_token()
+
+        speed = int(options.get('speed', ALIBABA_VOICE_SPEED_DEFAULT))
+        pitch = int(options.get('pitch', ALIBABA_VOICE_PITCH_DEFAULT))
+        voice = voice_key['name']
+
+        params = {
+            "format": "mp3",
+            "appkey": self.app_key,
+            "speech_rate": speed,
+            "pitch_rate": pitch,
+            "text": text,
+            "token": self.access_token["Id"],
+            "voice": voice
+        }
+
+        response = requests.get(
+            "https://nls-gateway-ap-southeast-1.aliyuncs.com/stream/v1/tts",
+            params=params,
+            timeout=cloudlanguagetools.constants.RequestTimeout
+        )
+
+        logger.debug(f'response status code: {response.status_code} headers: {pprint.pformat(response.headers)} length of response: {len(response.content)}')
+
+        if response.status_code != 200:
+            data = response.json()
+            error_message = f'could not retrieve audio using Alibaba: {response.content}'
+            logger.error(error_message)
+            raise cloudlanguagetools.errors.RequestError(text, voice, error_message)
+
+        if response.headers['Content-Type'] != 'audio/mpeg':
+            logger.error(f'Unexpected response type. Response as text: {response.text}')
+            raise cloudlanguagetools.errors.RequestError(
+                text, voice,
+                f'Got bad content type in response: {response.headers["Content-Type"]}'
+            )
+        
+        # Create a temporary file and write the audio content to it
+        output_temp_file = tempfile.NamedTemporaryFile(prefix=f'cloudlanguage_tools_{self.__class__.__name__}_audio', suffix='.mp3')
+        output_temp_filename = output_temp_file.name
+        with open(output_temp_filename, 'wb') as audio:
+            audio.write(response.content)        
+        logger.debug(f'result temporary file name: {output_temp_file.name}')
+        return output_temp_file
+
+    def get_tts_voice_list(self):
+        # returns list of TtsVoice
+        return []
+
+    def build_tts_voice_v3(self, name: str, voice_id: str, gender: cloudlanguagetools.constants.Gender, audio_languages: List[cloudlanguagetools.languages.AudioLanguage]):
+        return cloudlanguagetools.ttsvoice.TtsVoice_v3(
+            name=name,
+            voice_key={'name': voice_id},
+            options=VOICE_OPTIONS,
+            service=cloudlanguagetools.constants.Service.Alibaba,
+            gender=gender,
+            audio_languages=audio_languages,
+            service_fee=cloudlanguagetools.constants.ServiceFee.paid
+        )
+
+    def get_tts_voice_list_v3(self) -> List[cloudlanguagetools.ttsvoice.TtsVoice_v3]:
+        result = []
+
+        # Standard voices for all scenarios
+        result.extend([
+            # Standard Chinese voices
+            self.build_tts_voice_v3('Xiaoyun (Standard)', 'Xiaoyun', 
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Xiaogang (Standard)', 'Xiaogang',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Ruoxi (Gentle)', 'Ruoxi',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Siqi (Gentle)', 'Siqi',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Sijia (Standard)', 'Sijia',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Sicheng (Standard)', 'Sicheng',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aiqi (Gentle)', 'Aiqi',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aijia (Standard)', 'Aijia',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aicheng (Standard)', 'Aicheng',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aida (Standard)', 'Aida',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3("Ning'er", 'Ninger',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Ruilin', 'Ruilin',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+
+            # Customer service voices
+            self.build_tts_voice_v3('Siyue (Gentle)', 'Siyue',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aiya (Harsh)', 'Aiya',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aixia (Amiable)', 'Aixia',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aimei (Sweet)', 'Aimei',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aiyu (Natural)', 'Aiyu',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aiyue (Gentle)', 'Aiyue',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aijing (Harsh)', 'Aijing',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Xiaomei (Sweet)', 'Xiaomei',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+
+            # Regional accent voices
+            self.build_tts_voice_v3('Aina (Zhejiang)', 'Aina',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Yina (Zhejiang)', 'Yina',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Sijing (Harsh)', 'Sijing',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+
+            # Child voices
+            self.build_tts_voice_v3('Sitong (Child)', 'Sitong',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Xiaobei (Little Girl)', 'Xiaobei',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Aitong (Child)', 'Aitong',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Aiwei (Little Girl)', 'Aiwei',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Aibao (Little Girl)', 'Aibao',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+
+            # English voices
+            self.build_tts_voice_v3('Harry (British)', 'Harry',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Abby (American)', 'Abby',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Andy (American)', 'Andy',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Eric (British)', 'Eric',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Emily (British)', 'Emily',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Luna (British)', 'Luna',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Luca (British)', 'Luca',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Wendy (British)', 'Wendy',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('William (British)', 'William',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+            self.build_tts_voice_v3('Olivia (British)', 'Olivia',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_GB]),
+
+            # Special dialect voices
+            self.build_tts_voice_v3('Shanshan (Cantonese)', 'Shanshan',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_HK]),
+            self.build_tts_voice_v3('Xiaoyue (Sichuan)', 'Xiaoyue',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Lydia (Bilingual)', 'Lydia',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Aishuo (Natural)', 'Aishuo',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN, cloudlanguagetools.languages.AudioLanguage.en_US]),
+            self.build_tts_voice_v3('Qingqing (Taiwanese)', 'Qingqing',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Cuijie (Northeastern)', 'Cuijie',
+                cloudlanguagetools.constants.Gender.Female, [cloudlanguagetools.languages.AudioLanguage.zh_CN]),
+            self.build_tts_voice_v3('Xiaoze (Hunan)', 'Xiaoze',
+                cloudlanguagetools.constants.Gender.Male, [cloudlanguagetools.languages.AudioLanguage.zh_CN])
+        ])
+        
+        return result

cloudlanguagetools/amazon.py

@@ -0,0 +1,286 @@
+import json
+import requests
+import tempfile
+import os
+import boto3
+import botocore.exceptions
+import contextlib
+import logging
+import pprint
+
+import cloudlanguagetools.service
+import cloudlanguagetools.constants
+import cloudlanguagetools.options
+import cloudlanguagetools.languages
+import cloudlanguagetools.ttsvoice
+import cloudlanguagetools.translationlanguage
+import cloudlanguagetools.transliterationlanguage
+import cloudlanguagetools.errors
+import cloudlanguagetools.audio_processing
+
+from cloudlanguagetools.options import AudioFormat
+
+logger = logging.getLogger(__name__)
+
+DEFAULT_VOICE_PITCH = 0
+DEFAULT_VOICE_RATE = 100
+
+def get_audio_language_enum(language_code):
+    language_map = {
+        'arb': 'ar_XA',
+        'cmn-CN': 'zh_CN',
+        'yue-CN': 'zh_HK'
+    }
+
+    language_enum_name = language_code.replace('-', '_')
+    if language_code in language_map:
+        language_enum_name = language_map[language_code]
+
+    return cloudlanguagetools.languages.AudioLanguage[language_enum_name]
+
+class AmazonVoice(cloudlanguagetools.ttsvoice.TtsVoice):
+    def __init__(self, voice_data, engine: str):
+        # print(voice_data)
+        # {'Gender': 'Female', 'Id': 'Lotte', 'LanguageCode': 'nl-NL', 'LanguageName': 'Dutch', 'Name': 'Lotte', 'SupportedEngines': ['standard']}
+        self.service = cloudlanguagetools.constants.Service.Amazon
+        self.service_fee = cloudlanguagetools.constants.ServiceFee.paid
+        self.gender = cloudlanguagetools.constants.Gender[voice_data['Gender']]
+        self.voice_id = voice_data['Id']
+        self.name = voice_data['Name']
+        self.audio_language = get_audio_language_enum(voice_data['LanguageCode'])
+        self.engine = engine
+
+    def get_voice_key(self):
+        return {
+            'voice_id': self.voice_id,
+            'engine': self.engine
+        }
+
+    def get_voice_shortname(self):
+        return f'{self.name} ({self.engine.capitalize()})'
+
+    def get_options(self):
+        return {
+            'rate' : {
+                'type': 'number',
+                'min': 20,
+                'max': 200,
+                'default': DEFAULT_VOICE_RATE
+            },
+            'pitch': {
+                'type': 'number',
+                'min': -50,
+                'max': 50,
+                'default': DEFAULT_VOICE_PITCH
+            },
+            cloudlanguagetools.options.AUDIO_FORMAT_PARAMETER: {
+                'type': cloudlanguagetools.options.ParameterType.list.name,
+                'values': [
+                    cloudlanguagetools.options.AudioFormat.mp3.name,
+                    cloudlanguagetools.options.AudioFormat.ogg_vorbis.name,
+                    cloudlanguagetools.options.AudioFormat.wav.name
+                ],
+                'default': cloudlanguagetools.options.AudioFormat.mp3.name
+            }            
+        }
+
+class AmazonTranslationLanguage(cloudlanguagetools.translationlanguage.TranslationLanguage):
+    def __init__(self, language, language_id):
+        self.service = cloudlanguagetools.constants.Service.Amazon
+        self.service_fee = cloudlanguagetools.constants.ServiceFee.paid
+        self.language = language
+        self.language_id = language_id
+
+    def get_language_id(self):
+        return self.language_id
+
+class AmazonService(cloudlanguagetools.service.Service):
+    def __init__(self):
+        pass
+
+    def configure(self, config):
+        aws_access_key_id = config['AWS_ACCESS_KEY_ID']
+        aws_secret_access_key = config['AWS_SECRET_ACCESS_KEY']
+        region_name = config['AWS_DEFAULT_REGION']
+
+        self.polly_client = boto3.client("polly", 
+            region_name=region_name,
+            aws_access_key_id=aws_access_key_id,
+            aws_secret_access_key=aws_secret_access_key,
+            config=botocore.config.Config(connect_timeout=cloudlanguagetools.constants.RequestTimeout, read_timeout=cloudlanguagetools.constants.RequestTimeout))
+        self.translate_client = boto3.client("translate",
+            region_name=region_name,
+            aws_access_key_id=aws_access_key_id,
+            aws_secret_access_key=aws_secret_access_key,        
+            config=botocore.config.Config(connect_timeout=cloudlanguagetools.constants.RequestTimeout, read_timeout=cloudlanguagetools.constants.RequestTimeout))
+
+    def get_translation(self, text, from_language_key, to_language_key):
+        result = self.translate_client.translate_text(Text=text, 
+                    SourceLanguageCode=from_language_key, TargetLanguageCode=to_language_key)
+        return result.get('TranslatedText')
+
+    def get_tts_audio(self, text, voice_key, options):
+        response_format_parameter, audio_format = self.get_request_audio_format({
+            AudioFormat.mp3: 'mp3',
+            AudioFormat.ogg_vorbis: 'ogg_vorbis',
+            AudioFormat.wav: 'pcm'
+        }, options, AudioFormat.mp3)
+
+        logger.info(f'generating audio with voice {voice_key}')
+
+        # wav, we need to convert as described here:
+        # https://aws.amazon.com/blogs/machine-learning/integrating-amazon-polly-with-legacy-ivr-systems-by-converting-output-to-wav-format/
+
+        output_temp_file = tempfile.NamedTemporaryFile()
+        output_temp_filename = output_temp_file.name
+
+        pitch = options.get('pitch', DEFAULT_VOICE_PITCH)
+        pitch_str = f'{pitch:+.0f}%'
+        rate = options.get('rate', DEFAULT_VOICE_RATE)
+        rate_str = f'{rate:0.0f}%'
+
+        prosody_tags = f'pitch="{pitch_str}" rate="{rate_str}"'
+        if voice_key['engine'] == 'neural':
+            # pitch not supported on neural voices
+            prosody_tags = f'rate="{rate_str}"'
+
+        ssml_str = f"""<speak>
+    <prosody {prosody_tags} >
+        {text}
+    </prosody>
+</speak>"""
+
+
+        try:
+            if audio_format == cloudlanguagetools.options.AudioFormat.wav:
+                response = self.polly_client.synthesize_speech(Text=ssml_str, 
+                    TextType="ssml", 
+                    OutputFormat=response_format_parameter,
+                    VoiceId=voice_key['voice_id'], 
+                    Engine=voice_key['engine'],
+                    SampleRate="16000")
+            else:
+                if voice_key['engine'] in ['generative', 'long-form']:
+                    logger.info(f'voice: {voice_key}, generating text format: {text}')
+                    response = self.polly_client.synthesize_speech(Text=text, TextType="text", OutputFormat=response_format_parameter, VoiceId=voice_key['voice_id'], Engine=voice_key['engine'])
+                else:
+                    logger.info(f'voice: {voice_key}, generating ssml format: {ssml_str}')
+                    response = self.polly_client.synthesize_speech(Text=ssml_str, TextType="ssml", OutputFormat=response_format_parameter, VoiceId=voice_key['voice_id'], Engine=voice_key['engine'])
+        except (botocore.exceptions.BotoCoreError, botocore.exceptions.ClientError) as error:
+            logger.warning(f'Amazon Polly exception: {type(error).__name__}: {error}')
+            raise cloudlanguagetools.errors.RequestError(str(error)) from error
+
+        if "AudioStream" in response:
+            # Note: Closing the stream is important because the service throttles on the
+            # number of parallel connections. Here we are using contextlib.closing to
+            # ensure the close method of the stream object will be called automatically
+            # at the end of the with statement's scope.
+            with contextlib.closing(response["AudioStream"]) as stream:
+                with open(output_temp_filename, 'wb') as audio:
+                    audio.write(stream.read())
+
+                if audio_format == cloudlanguagetools.options.AudioFormat.wav:
+                    return cloudlanguagetools.audio_processing.wrap_pcm_data_wave(output_temp_file, 
+                        num_channels=1, 
+                        sample_width=2, 
+                        framerate=16000)
+                return output_temp_file
+
+        else:
+            # The response didn't contain audio data, exit gracefully
+            raise cloudlanguagetools.errors.RequestError('no audio stream')
+
+
+    def get_tts_voice_list(self):
+        result = []
+        response = self.polly_client.describe_voices()
+        # print(response['Voices'])
+        for voice in response['Voices']:
+            logger.debug(f'voice: {pprint.pformat(voice)}')
+            for engine in voice['SupportedEngines']:
+                result.append(AmazonVoice(voice, engine))
+        return result
+
+
+    def get_translation_language_list(self):
+        result = [
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.af, 'af'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sq, 'sq'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.am, 'am'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ar, 'ar'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.hy, 'hy'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.az, 'az'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.bn, 'bn'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.bs, 'bs'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.bg, 'bg'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ca, 'ca'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.zh_cn, 'zh'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.zh_tw, 'zh-TW'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.hr, 'hr'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.cs, 'cs'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.da, 'da'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.prs, 'fa-AF'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.nl, 'nl'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.en, 'en'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.et, 'et'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.fa, 'fa'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.tl, 'tl'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.fi, 'fi'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.fr, 'fr'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.fr_ca, 'fr-CA'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ka, 'ka'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.de, 'de'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.el, 'el'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.gu, 'gu'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ht, 'ht'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ha, 'ha'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.he, 'he'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.hi, 'hi'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.hu, 'hu'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.is_, 'is'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.id_, 'id'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.it, 'it'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ja, 'ja'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.kn, 'kn'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.kk, 'kk'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ko, 'ko'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.lv, 'lv'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.lt, 'lt'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.mk, 'mk'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ms, 'ms'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ml, 'ml'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.mt, 'mt'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.mn, 'mn'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.no, 'no'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.fa, 'fa'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ps, 'ps'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.pl, 'pl'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.pt_pt, 'pt'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ro, 'ro'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ru, 'ru'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sr_cyrl, 'sr'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.si, 'si'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sk, 'sk'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sl, 'sl'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.so, 'so'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.es, 'es'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.es_mx, 'es-MX'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sw, 'sw'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.sv, 'sv'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.tl, 'tl'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ta, 'ta'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.te, 'te'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.th, 'th'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.tr, 'tr'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.uk, 'uk'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.ur, 'ur'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.uz, 'uz'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.vi, 'vi'),
+            AmazonTranslationLanguage(cloudlanguagetools.languages.Language.cy, 'cy')
+        ]
+        return result
+
+    def get_transliteration_language_list(self):
+        result = []
+        return result
+   

cloudlanguagetools/audio_processing.py

@@ -0,0 +1,29 @@
+import wave
+import tempfile
+
+# audio utilities
+
+
+# take PCM audio and wrap it in a WAV container
+def wrap_pcm_data_wave(audio_temp_file: tempfile.NamedTemporaryFile,
+        num_channels,
+        sample_width,
+        framerate) -> tempfile.NamedTemporaryFile:
+    # read the audio data
+    f = open(audio_temp_file.name, 'rb')
+    data = f.read()
+    f.close()
+
+    wav_frames = []
+    wav_frames.append(data)
+
+    wav_audio_temp_file = tempfile.NamedTemporaryFile(prefix='clt_wav_audio_', suffix='.wav')
+
+    WAVEFORMAT = wave.open(wav_audio_temp_file.name,'wb')
+    WAVEFORMAT.setnchannels(num_channels) # one channel, mono
+    WAVEFORMAT.setsampwidth(sample_width) # Polly's output is a stream of 16-bits (2 bytes) samples
+    WAVEFORMAT.setframerate(framerate)
+    WAVEFORMAT.writeframes(b''.join(wav_frames))
+    WAVEFORMAT.close()    
+    
+    return wav_audio_temp_file